Home

ALICE DATE User's Guide - Compass

Contents

1. 1 update bookkeeping information 2 if LOGLEVEL gt 1 33 sprintf bookLog Run d DAQCONTROL gt runNumber 4 LOGBOOK bookLog SY sprintf bookLog Readout S summary on s READOUT_VERSION getenv DATE_HOSTNAME G2 LOGBOOK bookLog fs sprintf bookLog Data sob and eob count d DAQCONTROL gt eventCount os LOGBOOK bookLog 9 s sprintf bookLog Physics data trigger count d DAQCONTROL gt triggerCount LOS LOGBOOK bookLog Tis sprintf bookLog No of records inserted in the monitoring buffer d DAQCONTROL gt eventsMonitored 12 LOGBOOK bookLog To LOGBOOK_MARKER 14 5 4 Customization of the front end software 5 4 1 The Start of Run and End of Run scripts The user may customize the front end software for the experiment specific needs through common to all front end processors or specific to each front end processor Unix shell scripts which are called at every start of run and at every end of run called respectively SOR commands and EOR commands The common scripts are executed first and must reside in the DATE_SITE CONFIG directory while the specific ones reside in the DATE_SITE DATE_HOSTNAME directory These source scripts may contain the execution of other user written stand alone programs which do not have to be linked to the DAQ processes ALICE DATE V2 User s Guide Customization of t
2. Click to refresh the RUN with the latest information available on disk lt lt lt Disable Auto follow Refresh Search gt gt gt EJ Figure 2 14 Example of statsBrowser window The window of the statsBrowser can be divided in four sections As seen top to bottom they are 1 the top menu used for control over the tool and the font used for the display 2 the status line where the description of the tool and of the stats collection mechanism is shown 3 the run descriptors area where the information concerning the selected run is shown 4 the selection control used to control which type of information is shown on the display As the browser is written in Java it can run on any host were the standard DATE set is installed The display runs via a standard X11 link and therefore it is vital to have the X11 display setup and running At this purpose it is necessary to setup the DISPLAY shell variable either via the shell or opening a new X terminal and authorize the client on the server prior to run the infoBrowser ALICE DATE V3 User s Guide Data recording and data format This chapter describes how the data can be recorded in the LDCs and in the GDC It explains also the conventions concerning the filenames It then describes the format of the data produced in the LDCs the different event types used in DATE and the format of the full event built in the GDC 3 1 Data
3. date_mon 6601 tcp DATE Monitoring server 3 To make the modifications effective the following command will be invoked On AIX gt refresh s inetd On Solaris gt kill HUP inetdPID where inetdPID is the process ID of the inetd process it can be determined via the command gt ps e lgrep inetd ALICE DATE V3 User s Guide Installation of a new release 129 11 4 Installaton ofa new release The procedure of installing a new release of DATE consists of replacing all the files in the DATE_ROOT area 1 Import the distribution kit as described in 11 2 2 Expand the distribution kit as described in 11 3 1 point 6 All the site specific files are left untouched by the procedure It is though important to check the release notes to see whether the new packages require modification or additions to the configuration files 11 5 Run contol configuration 11 5 1 Run control windows configuration The run control presents to the operator a few windows from which it is possible to set up and modify the hardware configuration by selecting the machines involved in the run and their relationship and to establish the parameters that define the run conditions The content of the run control windows is itself parametrized Items to appear on the display windows must previously be defined in the file DATE_SITE_CONFIG runControl config This file must be prepared with an editor An example is sh
4. FORTRAN Synopsis Description Table 4 2 Table 4 3 INTEGER MONITOR_DECLARE_TABLE CHARACTI CI Re A table describing the desired monitoring policy is declared within the monitoring scheme Each monitoring program can declare a monitoring table at any time This table will be used for all subsequent calls to monitorSetDataSource and will be kept valid in case monitorLogout is called It is possible to declare a table in the middle of a monitoring stream this will force a flush of all events eventually available in the monitoring buffer and in the monitoring channel The input parameter should have the following C syntax char table nEntries event type monitoring type 0 where the nEntries is the actual number of entries belonging to the table and the field event type and monitoring type can assume one of the following values and aliases Event types Event type Single word alias Short alias All events All_events ALL Start of run Start_of_run SOR Start of run files Start_of_run_files SORE Start of burst Start_of_burst SOB Calibration event Calibration_event CAL Physics event Physics_event PHY Event format error Event_format_error FERR End of burst End_of_burst SEOB End of run
5. Triggering with the CORBO 7 1 7 2 7 3 7 4 7 4 1 7 4 2 7 4 3 7 4 4 7 4 5 The readout equipment types The trigger equipment types The detector parameters The detectors Using the CORBO as LDC trigger n module Start of run initialization End of run tidy up Trigger processing Reading the CORBO counters 7 5 Using the CORBO to control the trigger Chapter 8 Guide to use the infoLogger system 33 33 40 43 44 44 45 46 47 ol lt 52 52 53 54 54 54 55 57 58 59 61 62 63 63 64 64 65 65 66 66 66 68 68 69 70 70 71 72 74 75 76 76 77 79 80 80 81 83 ALICE DATE V3 User s Guide Contents vii 8 1 Introduction 8 2 The infoBrowser 8 2 1 The infoBrowser operator mode 8 3 Browsing HTML help pages 8 4 The information repository 8 5 Extracting portions of the log files 8 6 Injection of messages 8 6 1 The messaging system l 8 6 2 Setting the facility name a ah i 8 6 3 The infoLogger callable interface 8 6 4 The Java LogChannel Class constructors fields and methods 8 6 5 Interactive injection of messages Chapter 9 Guide to use the bookkeeping system 9 1 The DATE bookkeeping package 9 2 ThestatsCollector 9 2 1 The statsCollector configuration file 9 3 The statsBrowser 9 4 Browsing HTML help pages 9 5 The bookkeeping repository 9 5 1 How to re create the DATE_SITE_STATS rep
6. Monitoring programs have the choice to stall if no data is available or to continue with their execution knowing that no data has been received This allows the implementation of event driven processes such as X11 clients that should not be blocked in absence of data Another feature of the monitoring library is to let a monitoring program discard all data eventually stored in the monitoring buffer This is useful to access only future events at any given point in time Some experimental setups might hide their Data Acquisition hosts behind routers or firewalls making remote monitoring difficult or impossible To solve this problem the DATE monitoring library allows a mechanism called relayed monitoring where the monitoring channel travels through a dedicated relay host visible from the offline host and with access to the hidden online host The scheme is described in Figure 4 3 It is possible to filter the access through the relay host only to a restricted set of clients according to the type of monitoring requested Relayed monitoring performs worse then direct monitoring and should be used only whenever absolutely unavoidable ONLINE HOST LDC GDC or PDS attached host monitoring buffer RELAY HOST FIREWALL OFFLINE HOST remote monitoring program Figure 4 3 The DATE relayed monitoring 4 2 Monitoring and Analysis in C FORTRAN A monitoring program should acc
7. PHYSICS_EVENT are assembled into full events For all the sub events of the other types the event builder adds a header to the sub event coming from one LDC and record them The event builder does not attempt to assemble these sub events into complete events ALICE DATE V3 User s Guide 24 Data recording and data format Listing 3 2 Event types 1 Possible Values for type in eventHeaderStruct Zs 3 Mask to separate th vent type from the event flags 4 define EVENT_TYPE_MASK unsigned long32 OxOO00FFFF 53 define EVENT_FLAGS_MASK unsigned long32 EVENT_TYPE_MASK 6 TS Event error flags 8 define EVENT_ERROR unsigned long32 0x80000000 Qs define EVENT_DATA_TRUNCATED unsigned long32 0x40000000 10 11 Event flags ea define EVENT_EQUIPMENT unsigned long32 0x00010000 13 14 Event types 15 3 define START_OF_RUN unsigned long32 1 16 define END_OF_RUN unsigned long32 2 Let define START_OF_RUN_FILES unsigned long32 3 18 define END_OF_RUN_FILES unsigned long32 4 19 define START_OF_BURS unsigned long32 5 20s define END_OF_BURST unsigned long32 6 21 define PHYSICS_EVENT unsigned long32 7 22 define CALIBRATION_EVENT unsigned long32 8
8. gpar gt namePtr printf with min d and max d n randPar gt eventMinSizePtr randPar gt eventMaxSizePtr void DisArmRand char parPtr int ReadEventRand char parPtr struct eventHeaderStruct header_ptr struct equipmentHeaderStruct eq_header_ptr int data_ptr equipmentDataTableType equipmentDataTable equipmentDataType Rand int nbDtEgps 1 equipmentNoDataTableType equipmentNoDataTable int nbNoDtEqps 0 triggerDataTableType triggerDataTable int nbDtTrgs 0 triggerNoDataTableType triggerNoDataTable triggerNoDataType Corbo bi int nbNoDtTrgs 1 ALICE DATE V3 User s Guide 68 The generic readList 6 6 The detectors configuration fie 6 6 1 The detector configuration file DATE_SITE_CONFIG detectors config contains the description of the readout system in terms of equipments This file describes all the detectors and is unique in the experiment Each detector is composed of equipments and triggers that may either produce data or not Each detector and each equipment has a set of parameters associated with it Each equipment may be activated and de activated The configuration file is made of four sections appearing in the order indicated below and identified by their keywords The first three sections are declarations of types and of list of parameters 1 gt EQTYPES followed by the d
9. 23 define END_OF_LINK unsigned long32 9 24 define EVENT_FORMAT_ERROR unsigned long32 10 25 26 define EVENT_TYPE_MI 1 27 define EVENT_TYPE_MAX 10 The list of possible record types of the field type in the event header is given in Table 3 2 Table 3 2 List of record types Type of record Set by START_OF_RUN END_OF_RUN START_OF_RUN_FILES END_OF_RUN_FILES START_OF_BURST END_OF_BURST PHYSICS_EVENT CALIBRATION_EVENT the readout process the readout process the readout process the readout process the user routine ReadEvent the user routine ReadEvent the readout process not used in the present version 3 5 The full event format The data format structure described before applies to sub events and to full events Each event will include a header and a data block In the cases of a full event assembled by the event builder the data block is itself subdivided into sub events Each subevent will include a header and a data block The event builder assembles ALICE DATE V3 User s Guide The full event format 25 the subevents pertaining to the same event and adds one header relative to the complete event at the beginning of the event as shown in Figure 3 1 The sub event refers here to the data read out by one LDC and assembled later on by one GDC The full event refers here to the collection of data collected by one DATE system Therefore there are two types o
10. ALICE DATE V3 User s Guide ALICE DATE architecture Figure 1 6 1 6 Figure 1 7 Information logger architecture To log a message a process calls a library routine which sends the message onto a socket A machine or workstation is designated to receive these messages Internet dzemons called infoDaemon handle each socket and save the messages on disk files A program called infoBrowser may be interactively invoked to browse these files and apply various selection criteria to the messages to be displayed Run bookkeeping It is possible to record on disk files information concerning each run Any program may generate summary information and log it This feature uses the information logging system to write the summary file A process called statsCollector will select the summaries from the file and save them files dedicated to each run figure 1 7 A program called statsBrowser may be interactively invoked to browse these files Run bookkeeping ALICE DATE V3 User s Guide Guide to operate the system This chapter describes the person machine interface of DATE There is one part concerning the configuration and the control of the data acquisition another one on the information provided to the operator and a last one on the run bookkeeping 2 1 The control console 0 2 000000 8 4G 8 2 2 TheinfoBrowser console 00 0004 15 2 3 ThestatsBrowser console 0 0004 17 A
11. Access to the VME bus 73 The register access is obtained by calculating the pointer to it as the sum of the device base address and the register offset Read and writes from to the register are performed as virtual memory read and writes Listing 7 2 Example of an application accessing a VME device 7 anApplication c 8 Application controlling the device aDevice OS 10 include aDevice h 11 12 main 13 14 Set the offset and the size of the VME window IRo unsigned long vmeWinOffset vmeWinSize 16 vmeWinOffset 0 Start from the bottom Ls vmeWinSize 0xF00000 Map part of the 24 bit the VME space 18 19 Map the VME window onto a memory window 20 A pointer to the first 32 bit word is returned Zits unsigned long vmeWinAddr 22 3 vmeWinAddr unsigned long MapVME vmeWinOffset vmeWinSize 23 24 Set the vme base address of the device 20 8 It may depend on a switch setting on the board 26 unsigned int aDeviceVmeBaseAddr 2h aDeviceVmeBaseAddr 0x700000 28 29 Calculate the memory base address of the device 30 unsigned int aDeviceBaseAddr 3s aDeviceBaseAddr unsigned int vmeWinAddr 32 aDeviceVmeBaseAddr unsigned int vmeWinOffset 33 34 Calculate the pointers to the 16 bit registers 35s register unsigned short aDeviceCsr aDevice InputReg aDevi
12. 91 8 6 where the directory is where the result file is in our example tmp logs The infoBrowser will allow browsing of all files contained in the given directory as long as their format agrees with the infoLogger conventions Injection of messages Any process running on any DATE host can make use of the infoLogger to transfer debug information and error logs This can be done from executable images main support is given for programs written in C using the C callable interface interpreted environments Java via the LogChannel class Tcl using the shell interface and interactive batch processes via the shell interface The first choice to be made by any DATE developer is where the log messages generated by his her package should go If the message is of common interest to the whole Data Acquisition system then it should belong to the standard DATE_SITE_LOGS runLog stream this is where all processes send their general interest information If the developer needs a dedicated stream this can be established and created at run time without prior configuration needed There are no limits on the number of streams to be created and there is no convention outside the common stream rule A running image can alternate between different streams according to the situation or to the type of message to exchange It is also possible to ship a message both the generic log stream and to a package specific stream with a single call to the
13. 14 triggerNb regPtrShort T5 J 7 5 Using the CORBO to contol the tigger The DATE run control software may optionally launch at start of run a script called EnableTrigger and conversely launch at end of run a script called DisableTrigger The purpose of this is to generate via software the general trigger enable signal described in 7 2 The CORBO module may be used as a simple signal generator The internal flip flop is set by software and the output signal is latched to the flip flop status We suppose here to use one of the CORBOs already in use for triggering a LDC where channels 1 and 2 have a defined task Therefore we will use in the example a spare channel namely the channel 4 The program EnableTrigger performs the following operations Map to the VME space in the same way shown in Listing 7 5 Initialize the CORBO channel 4 as shown in Listing 7 12 1 2 3 Clear the busy output as shown in Listing 7 13 4 Unmap the VME space as shown in Listing 7 9 ALICE DATE V3 User s Guide VME access and trigger system Listing 7 12 CORBO channel 4 initialization 1 Channel 4 initialization 2 Input not used output enable signal 3 regPtrShort unsigned short corboBaseAddr CORBO_CSR4 4 regPtrShort INPUTINTERNAL Enable internal trigger 5 regPtrChar unsigned char corboBaseAddr CORBO_BIM1 OFST_CR3 6 regPtrChar 0 D
14. Control status display While in the event loop the readout process continuously checks for the arrival of the end of run command The readout exits the event loop if one of the following two conditions is met the maximum number of events to be taken has been reached or somebody asked to stop the run All the records are always inserted in the buffer named RECORDINGBUFFER including SOR and EOR records SOR files and EOR files records while they are injected in the buffer reserved for monitoring only if the following conditions are met 1 the monitor enable flag is set to 1 in the Run Control panel 2 amonitor program requesting this type of events is running 3 there is enough space in the monitoring buffer At end of run the readout process performs the following sequence of operations in the order described below e execute the common end of run scripts if present e execute the specific end of run scripts if present e call the user routine DisArmHw e build the header of the EOR record and fill the data of the EOR record with the ALICE DATE V2 User s Guide 46 Guide to prepare a readout program run conditions copying the Run Control shared segment e prepare the common EOR files one record per file e prepare the specific EOR files one record per file e update the bookkeeping information with the physics events count the SOB records count the EOB records count the trigger count and the number of reco
15. files and EOR files The common files are recorded first and must reside in the DATE_SITE_CONFIG directory while the specific ones reside in the DATE_SITE DATE_HOSTNAME directory Each file in the list of filenames in SOR files or EOR files creates one separate record on the output medium and is written after the execution of the SOR commands or EOR commands scripts so that the latest version produced by those scripts is always obtained Listing 5 6 and Listing 5 7 are an example on how to set high voltages and at the beginning of each run produce a file with these values run a calibration program and add a run comment to the tape Listing 5 6 SOR commands bin sh SOR commands date setup sh high_voltage_program gt CALIBRATION_DIR HV dat sleep 5 wait for HV to settle ADC_calibration_program gt CALIBRATION_DIR ADC dat log File HV DAT contains the current high_voltages DANA HAWNEH ALICE DATE V2 User s Guide Guide to prepare a readout program Listing 5 7 5 5 5 5 1 5 5 2 SOR files 1 CALIBRATION_DIR HV dat 2 CALIBRATION_DIR ADC dat 3 run comment The programs high_voltage_program and ADC_calibration_program are totally independent from the DAQ system and they have just to produce result files The files can have any format but standard ASCII formatted text files are strongly recommended since it is the only format which can be read on any monitorin
16. x DATE_SITE directory where directory is the name of the new directory 6 Login with a user account 7 Expand the distribution kit previously retrieved in tmp All the files will be automatically placed in the right location gt ed date gt rm rf gt tar xf tmp distribution_kit tar Removing all the files before invoking tar is useful to make sure that files no longer necessary are deleted Be careful not to delete either the directory commonScripts or the files in it If you get the compressed version of the kit suffixed with tar gz you should first de compress it gt gunzip distribution_kit tar gz ALICE DATE V3 User s Guide 126 DATE installation guide 10 11 12 13 If disk space is a problem decompression and unpacking can be done in a single step With date as your working directory do gt gunzip c distribution_kit tar gz tar xf Setup all the DATE symbols In Bourne shell and in BASH do the following gt r date setup sh amp amp date setup sh In C she11 do the following gt source date setup csh The operations described in the points 5 and 8 above are necessary each time you want to use DATE Therefore you may want to add them to your login file e g in zshenv profileor bashrc Create under DATE_SITE the following sub directories logFiles configurationFiles runControl eventBuilder eb_host where eb_host is the name of the machin
17. 1 DATE features The processors may be of any type such as embedded VME processors or fully equipped workstations Some of the activities of the data acquisition such as run ALICE DATE V3 User s Guide Readout and data flow 3 control and monitoring may be more conveniently located on one or more workstations which may or may not share the file system of the readout processors The DATE architecture may be concentrated onto one single processor In this case the GDC function is missing but all the others functions are supported The DATE system besides the data flow function provides a number of other features depicted in figure 1 1 namely e Parametrization of the hardware configuration and interactive setting up e Run control e Display of the run status e Event monitoring e Information reporting e Run bookkeeping 1 2 Readout and data flow The readout and data flow architecture is shown in figure 1 2 The readout is performed in each LDC by the process readout which waits for a trigger and then reads the front end electronics and fills a circular buffer Another process called recorder off loads the buffer and sends the events to whatever device has been indicated in the configuration It may be either a disk file this is the case of a single LDC when there is no GDC or more usually it is the IP address of the GDC Figure 1 2 Dataflow architecture In the GDC an Internet d mon called gdc
18. 128 128 130 139 140 152 List of Listings Listing 11 6 Creation of configuration files 2 2 2 2 142 Listing 11 7 Remote monitoring test procedure WM AS A Re Ae MA Listing 11 8 Example of failed remote monitoring installation et ADEs Soe 41a Se eee ALICE DATE V3 User s Guide Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table 3 1 3 2 3 3 3 4 4 1 4 2 4 3 4 4 5 1 8 1 8 2 11 1 11 2 11 3 11 4 List of Tables Event header fields List of record types Usage of the detector identification mask header field Usage of the header len and of the size Monitor source parameter syntax Event types Monitoring types Bytes swapping control LOGLEVEL definitions for the redont and the acohder processes Selection fields Some examples of regular expressions Hardware and software platforms Keywords of the configuration file Run control parameters i Monitoring configuration parameters ALICE DATE V3 User s Guide 22 24 25 25 34 35 739 39 52 89 90 124 131 132 142
19. 94 Guide to use the infoLogger system Java Synopsis Description C Synopsis Java Synopsis Description C Synopsis Java Synopsis Description LOG_ERROR LOG_FATAL void logTo String stream char severity String message severity can be one of DB severityInfo DB severityError DB severityFatal The given message is sent to the infoLogger host to the given information stream with the given severity LOG_ALL logAIl include infoLogger h void LOG_ALL char stream char severity char message severitycan be one of LOG_INFO LOG_ERROR LOG_FATAL void logAll String stream char severity String message severitycan be one of DB severityInfo DB severityError DB severityFatal The given message is sent to the infoLogger host both to the default runLog and to the given information st ream with the given severity INFO info include infoLogger h void INFO char message j void info String message The given message is sent to the infoLogger host standard information stream with Info severity ERROR error ALICE DATE V3 User s Guide Injection of messages 95 C Synopsis Java Synopsis Description C Synopsis Java Synopsis Description C Synopsis Java Synopsis Description C Synopsis Java Synopsis Description include infoLogger h void ERROR char message void e
20. Disable time out interrupt Lae regPtrChar unsigned char corboBaseAddr CORBO_BIM2 OFST_VR1 13 regPtrChar 0 Clear time out vector 14 regPtrShort unsigned short corboBaseAddr CORBO_TEST2 15 regPtrShort 0 Prepare to trigger 16 regPtrShort FULLBYTE and then trigger once set busy ar LT regPtrShort unsigned short corboBaseAddr CORBO_CNT2 r 18 regPtrShort 0 Clear event counter Loy regPtrShort unsigned short corboBaseAddr CORBO_TOU2 r 20 regPtrShort 0 Clear dead time counter 2g regPtrShort unsigned short corboBaseAddr CORBO_CSR2 22 regPtrShort COUNTINPUT INPUTFRONTPANEL Enable front panel trigger Finally a flag is set to inform the routine EventArrivedCorbo that it is called for the first time as shown in Listing 7 8 The flag will instruct the routine EventArrivedCorbo when called to perform the same action as after an event Listing 7 8 Set flag for EventArrivedCorbo 1 Initialize trigger flag 23 eventArrivedFlag 1 3 4 7 4 3 End of run tidy up The only tidy up to be made at end of run is the releasing of the virtual memory space as shown in Listing 7 9 Listing 7 9 Releasing the virtual memory space DisarmCorbo UnmapVME iL 28 3 Unmap the memory window 4 5 ALICE DATE V3 User s Guide VME access and trigger system
21. If a functionality is not required dummy routines should still be provided Accessing the parameters All the functions in the library have access to two sets of parameters a The equipment specific parameters These are accessible via a pointer received as first parameter in the routine call see example in Listing 6 2 b The detector specific parameters They are unique in a given detector They are accessible via a global pointer see example in Listing 6 2 char globPar The order type and format of all the parameters is matter of convention within the experiment Coherence must be assured between what is specified in the detectors configuration file and the code in the equipmentList c library No check is performed by readList before calling the library The actual values passed to the routines are the ones specified in the detectors configuration file in the section gt DETECTORS ALICE DATE V3 User s Guide The equipmentList library 65 The values of the parameters are copied into memory at run time while parsing the configuration file following this convention on their formats 1d corresponds to long hdto short 1x to long hex value hd to short hex value s and to char and cto char To ease the use of the parameters it is suggested to cast their memory pointer into a pointer to a structure with proper fields according to what declared into the detectors config configuration file in the same order and f
22. List of Figures Figure 1 1 DATE features ae Figure 1 2 Dataflow architecture we Figure 1 3 Event builder architecture A Figure 1 4 Monitoring architecture 4 Figure 1 5 Run control architecture 5 Figure 1 6 Information logger architecture 6 Figure 1 7 Run bookkeeping 6 Figure 2 1 The main run control window 9 Figure 2 2 The main run control window after connection 9 Figure 2 3 The status display 10 Figure 2 4 The file menu 11 Figure 2 5 The view menu 12 Figure 2 6 The extended main run control window 12 Figure 2 7 The option menu 13 Figure 2 8 No update indicator 13 Figure 2 9 The windows menu 14 Figure 2 10 The configuration parameters window 14 Figure 2 11 The buffer status window 15 Figure 2 12 Example of infoBrowser window 16 Figure 2 13 The infoBrowser operator window 17 Figure 2 14 Example of statsBrowser window 18 Figure 3 1 The full event format 26 Figure 4 1 The DATE online monitoring local and remote configurations 28 Figure 4 2 The DATE offline monitoring 2 2 29 Figure 4 3 The DATE relayed monitoring 30 Figure 5 1 Maineventloop fg amp chitin ee ae ae ee SE Figure 6 1 Readout and the standard readList be Be wd he dh amp a a HOB Figure 6 2 Readout and the genericreadList 2 59 Figure 6 3 The readoutControl window 61 Figure 6 4 Sub event structure 6k Ao wi ae Se nk a Ok Se HO Figure 6 5 The equipment header
23. Listings Courier 9 pt Symbol Courier Bold Italic 10 5 pt Use of any trademark in this document is not intended in any way to infringe on the rights of the trademark holder Preface The ALICE DATE Data Acquisition and Test Environment system has been developed as a basis for prototyping the components of the DAQ system and for the support of the ALICE test beams The ALICE DATE system includes a set of programs and packages needed for a data acquisition system such as readout monitoring error reporting and run control and also prototypes of the components developed for the future ALICE DAQ system such as the event building based on a switched network The ALICE DATE system is based on widely accepted hardware and software industry standards such as VME boards and workstations running Unix Java Tcl Tk TCP IP socket library The ALICE DATE system is designed to run on two different types of machines e The Local Data Concentrator LDC has the following functions readout the front end electronic format the data fragments into sub events record the data or send them to a GDC e The Global Data Collector GDC has the following functions event building formatting of subevents into events and data recording The list of releases of the ALICE DATE User s Guide is the following April 98 first released version for DATE V2 November 98 updated version for DATE V3 the following chapters have been modified D
24. Mo tine ter JOREOR pases m LEE Tipped ppc Renoning mire Henitor onabie Bag i nyging level jo Figure 11 1 Run control configuration window 11 5 4 Run number The run number is automatically incremented at each start of run It cannot be easily changed by the operator The run control creates a file DATE_SITE_CONFIG runNumber config and stores there the last run number used by DATE If you want to modify the current run number in order to re start the series you may edit the file and replace the run number with the new number minus one An example of the file exists in the distribution date runCont rol runNumber config which may be copied in the right place and edited there The file contains the line shown in listing Listing 11 4 Listing 11 4 Repository of the run number 1 set runNumber 0 ALICE DATE V3 User s Guide 140 DATE installation guide 11 5 5 11 6 Listing 11 5 11 7 Make sure that the file has got write access for everybody chmod a tw DATE_SITE_CONFIG runNumber config Multiple run controls It is not allowed to have several run controls active at the same time If more than one run control try to connect to the same machines an alert window is opened offering as only option give up In the case of a crash of the run control while connected to the DAQ machines it would be impossible to have a new run control connected to previously connected machines This li
25. User s Guide 16 Guide to operate the system Tool Fonthiea Poreground Background HensageaForegrourd Nessageabackaroumd Halp MS HHA BE Festa DiD hokas Tasliity USArnam EE oe ee E Ae E F fulider Usage seata amaneni Haire LD za 2 19 40 59 cee panes Ey Erinin i U22 ta 06 01 ERA A arene eed eealagl z irr 14 10 123 presidii P E vent fro ar DEL 1442103 NE OD0DRSESLS event Bud ide wo Byvant polider canner 11 08 08 recede B visibla Click ta focus Pylip Palen Home End arn to browre around Bll gaverttder Balect all dogs Peaelect alli J ja Dump ALL HEA Clear dis lear abo Clear display Clap above qurrent Mavant fod ler Figure 2 12 Example of infoBrowser window 2 2 1 The infoBrowser window can be partioned in three sections 1 the top menu used to control the tool and the font used for the display 2 the messages area where the information is shown 3 the selection control used to control which type of information is shown on the display Messages are sorted by generation time and source host The resolution of the time stamp is one second In case messages are generated at high rate more then one per second it is not possible to sort them correctly if coming from different sources processes The stream generated by one given process is always sorted in the correct order Messages are presented with all their fields time stamp source computer source facility process ID i
26. User s Guide Index DisarmHw 51 EventArrived 48 ReadEvent 48 recorder program 3 5 20 routine ArmHw 44 47 routine DisArmHw 44 routine DisarmHw 51 routine error 94 routine errorAll 96 routine errorTo 95 routine EventArrived 48 routine fatal 95 routine fatalAll 96 routine fatalTo 95 routine info 94 routine infoAll 96 routine infoTo 95 routine log 93 routine logAll 94 routine logBook 108 routine logBookMarker 108 routine logTo 93 routine MONITOR_FREE_EVENT 37 routine monitorControlWait 38 routine monitorDeclareMp 34 routine monitorDeclareTable 34 routine monitorDecodeError 39 routine monitorFlushEvents 37 routine monitorGetEvent 36 routine monitorGetEventDynamic 36 routine monitorLogout 40 routine monitorSetNowait 38 routine monitorSetSwap 38 routine monitorSetWait 37 routine ReadEvent 48 routines readlist 47 RUN parameters 132 run End of 11 20 21 46 run Start of 8 9 10 11 46 60 S Start of run 8 9 10 11 46 60 ArmHw 63 64 ArmHw routine 48 command 11 conditions 9 files 24 53 in readout 44 45 operations performed at 45 parameters at 8 record 45 records 21 24 ALICE DATE V3 User s Guide 147 script 45 52 53 statsBrowser program 6 105 statsCollector program 6 103 T Template readlist 47 U User nobody 20 90 128 root 125 127 V VME Access to the VME bus 72 CORBO module 75 148 Index ALICE DATE V3 User s Guide
27. data and all the records inserted by the readout process in RECORDINGBUFFER across the network In this case the name set by the user as recording device is the name of the machine to which the data have to be sent usually the event building CPU The recorder process performs the following sequence of operations in the order described below e initialize RECORDINGBUFFER where the readout will insert data e open a local disk file or connect to the event building remote machine in this case the port number defined as environment variable DATE_SOCKET_EB is used e enter the event loop writing each event on the local disk file or sending each event over then network e inthe event loop the behavior of the recorder process is different depending on the value of the flag enableRecordingInBurst set in the Run Control parameters If it is set to 1 the recorder process competes all the time for the CPU with the readout process running in the same machine Setting it to 0 makes the recorder process go to sleep while the ALICE DATE V2 User s Guide The readList templates 47 variable DAQCONTROL gt event Count changes The variable DAQCONTROL gt eventCount does not move either out of burst i e no more trigger are arriving or when the RECORDINGBUFFER is full This implementation improves the performance when bursts are present since it forbids the recorder process to steal the CPU to the readout process inside the burst The t
28. dura The operator viathe Configuration Limit tion in seconds of run control parameters win any start of run and dow Common end of run phase parameters under Suggested value is gt CNFPAR in run 30 Used by Control config rcServer to abort the run if readout does not complete the phase in due time randEventMin Optional parame The operator viathe Optional Configu Size ter used exclu run control ration parameters sively by the window LDC example of readout parameters under program contained gt LDCPAR in run in the DATE distri Control config bution kit Mini mum event size for the random genera tor randEventMax Optional parame The operator viathe Optional Configu Size ter used exclu run control ration parameters sively by the window LDC example of readout parameters under program contained gt LDCPAR in run in the DATE distri Control config bution kit Maxi mum event size for the random genera tor randEventInter Optional parame The operator viathe Optional Configu val ter used exclu run control ration parameters sively by the window LDC example of readout parameters under program contained gt LDCPAR in run in the DATE distri Control config bution kit Time interval between events in microsec onds 136 DATE installation guide Table 11 3 Run control parameters Parameter name Description Set by To appear in window maxFileSize Each run may be The operator
29. f DATE_SITE_LOGS Any active daemon tool using the DATE_SITE_LOGS area at the moment of the cleanup should be able to sense the cleanup and must adapt itself to the new running conditions If not it may be necessary to stop the image and restart it once the cleanup has been performed The standard infoBrowser tool can re configure itself during the cleanup procedure and does not need to be restarted ALICE DATE V3 User s Guide Extracting portions of the log files 89 8 5 Extracting portions of the log files Whenever the content of the DATE_SITE_LOGS area becomes too big or too dispersed it is possible to filter out the unimportant bits to produce a single file containing only what is of most interest The selection can be based on any of the fields included in the messages time of day source facility username messages Standard Unix tools grep egrep sort can be used to perform rather complex selections The output of the procedure is a new log file that can be stored for future reference or simply discarded after use The situation is the following in DATE_SITE_LOGS we have a set of log files where a subset is of interest to us This subset can be selected via one or more fields using regular expressions We want to extract the subset of messages and save it in a new log file e g tmp logs result log The commands to run on the machine hosting the DATE_SITE_LOGS repository is the following gt ca
30. facilities belonging to a DATE system can generate run time statistics records to be collected via the stat sCollector daemon and browsed with the stat sBrowser tool The statsBrowser is an X11 tool written in Java capable to scan over the records describing a series of runs display them and update their content on the fly It is also possible to perform online searches of given text patterns over the available data Help on the usage of the tool and the statistics scheme is available via a context sensitive scheme and using an associated Help command To start the statsBrowser console the first step is to run the DATE setup procedure Once the X11 display is set and enabled it is possible to run the command statsBrowser from the shell level There are various run time flags available for special purposes use the he1p flag for a complete description In its default configuration the statsBrowser will allow browsing of the online run description records area An example of how the infoBrowser window looks like is given in Figure 2 14 Please note that the appearance might change between different versions of the tool ALICE DATE V3 User s Guide Guide to operate the system Tool FontSize Color op RUN 4264 910 4266 550 files AUTO FOLLOW Event Builder VO1 07C summary Number of Start of Run 2 Number of End of Run Files 2 Output Filename dev null Number of records 9608 0 errors Data recorded 290 MBytes
31. files End_of_run_files EORF End of link End_of_link EOL End of run End_of_run EOR Monitoring types Monitoring type Action all all events of this type are monitored 100 yes a sample of the events of this type is monitored ino no events of this type are monitored ALICE DATE V3 User s Guide 36 Guide to write a monitoring program Returns C Synopsis FORTRAN Synopsis Description Returns C Synopsis All declarations are case insensitive and can be shortened to the nearest unique string watch out for ambiguous shortening e g end of run can match either end of run or end of run files The single word aliases are used to comply with the FORTRAN argument passing convention The input parameter has the following FORTRAN syntax event_type monitoring_policy e g the table SOR yes EOR yes Physics_event yes can be used to request only start of run end of run and physics events The syntax for the event_type and monitoring policy fields is the same as for the equivalent C call see Tables 4 2 and 4 3 with the only difference that blanks are not allowed inside the keywords either use the short aliases or single word aliases where blanks are replaced by underscores The default table is the following char defaultTable 3 All events yes 0 C
32. from the stamp file DATE_SITE_CONFIG runNumber config The syntax for the trailer field is the following FEFFE EEE ETH 4444444 The trailer string must have at least 10 characters The bookkeeping calls As the bookkeeping transport method is based on the infoLogger package the same rules described for the infoLogger callable interface are valid also for the bookkeeping callable interface Please refer to the ALICE DATE Reference Manual for more details This section describes the specific bookkeeping entries available for C programs Java code Tcl code via the shell library and Unix shells LOGBOOK logBook logBook include infoLogger h void LOGBOOK char message void logBook String message gt logBook message The given message is sent to the bookkeeping stream of the infoLogger host The message can be a record header part of or the whole of a record body and a record trailer although trailers can be sent using a logBook marker call LOGBOOK_MARKER logBookMarker logBookMarker include infoLogger h void LOGBOOK_MARKER void logbookMarker gt logBookMarker An record trailer marker is sent to the bookkeeping stream of the infoLogger host ALICE DATE V3 User s Guide The bookkeeping callable interface 109 9 6 3 Single line records examples Listing 9 1 Listing 9 2 Listing 9 3 9 6 4 Here are some examples of single line bookk
33. in the table Indeed the fields of the event header are set by one of the three following packages the readout process the user code of the readout process or by the event builder A program is included in the DATE system to dump a data file recorded following the DATE data format This tool is available in the monitoring package and is called eventDump see Section 4 3 ALICE DATE V3 User s Guide Data recording and data format Listing 3 1 Event header structure 1 struct eventHeaderStruct 2 long32 size size of event in Bytes Si unsigned long32 magic magic number used for consistency check 4 unsigned long32 type event type 5 unsigned long32 headLen size of header in bytes 6 unsigned long32 runNb run number Fs unsigned long32 burstNb burst number 8 unsigned long32 nbInRun event number in run gs unsigned long32 nbInBurst event number in burst 10 unsigned long32 triggerNb trigger number for this detector TI unsigned long32 fileSeqNb File sequence number for multifiles run 12 detectorIdType detectorId MASK_LENGTH detector identification I3 unsigned long32 time Time in seconds since 0 00 GMT 1 1 1970 LAs unsigned long32 usec microseconds 15 unsigned long32 errorCode 16 unsigned long32 deadTime The unsigned long32 deadTimeusec Si he Table 3 1 Event header fields FIELD M
34. infoLogger library The DATE_SITE_LOG area may also contain a special stream dedicated to DATE statistics but this should be accessed only via its dedicated interface see the chapter on DATE statistics for more details Once the information stream has been clearly defined the developer must choose the information value of each message The infoLogger scheme proposes three levels of severities e Information used for messages concerning normal running condition time stamps important events debug statements e Error an abnormal situation has been encountered but execution can somehow continue after some optional recovery actions e Fatal an unrecoverable situation has been detected and normal execution cannot be guaranteed The combination of information stream and severity should provide enough flexibility for all types of filtering and grouping of messages The infoBrowser tool highlights Error and Fatal messages for easy detection Several DATE products readout recorder eventBuilder runControl use the above strategy plus a variable logging level this gives an extra level of freedom to the developer operator for the generation of different level of details according to the situation ALICE DATE V3 User s Guide 92 Guide to use the infoLogger system 8 6 1 8 6 2 Listing 8 1 The messaging system When a message is exchanged using the infoLogger library an atomic exchange takes place between the cl
35. new events This is the default behaviour of the monitoring library Zero in case of success otherwise a error code see 4 2 3 for more details monitorSetNowait include monitor h int monitorSetNowait void INTEGER MONITOR_SET_NOWAIT After this call completes if the monitoring program requests an event when the monitoring buffer and the monitoring channel are empty the monitoring program will continue and a error code will be returned Zero in case of success otherwise a error code see 4 2 3 for more details monitorControlWait include monitor h int monitorControlWait int flag INTEGER MONITOR_CONTROL_WAIT INTEGER The wait nowait behaviour of the monitoring library is set accordingly to the input parameter e true wait C 0 0 FORTRAN 1 e false nowait C0 1 FORTRAN 0 Zero in case of success otherwise a error code see 4 2 3 for more details monitorSetSwap ALICE DATE V3 User s Guide Monitoring and Analysis in CIFORTRAN 39 C Synopsis FORTRAN Synopsis Description Table 4 4 Returns C Synopsis FORTRAN Synopsis Description include monitor h int monitorSetSwap int 32BitWords int 16BitWords INTEGER MONITOR_SET_SWAP INTEGER INTEGER This entry controls the behaviour of the monitoring library when a network cha
36. of typical users of the central bookkeeping repository are e an offline reconstruction program looking for the location and the characteristics of a given run data file e an operator developer checking the characteristics of the hardware and the software involved in the acquisition e an operator checking the behavior of the Acquisition System and its interaction with the external resources beam cooling control network As the bookkeeping package is mainly used to collect statistics concerning the running characteristics of the Data Acquisition system the terms bookkeeping and statistics are often intermixed and in the context of this document may be used to point to the same entity The implementation of the current release of the bookkeeping package makes intensive use of the infoLogger scheme The sources of information for the bookkeeping repository processes belonging to the Data Acquisition system either from online LDCs GDCs or offline hosts PDS drivers experiment control ship their data using the infoLogger library This data is stored in a dedicated stream named DATE_SITE_LOGS logBook A specialized daemon called ALICE DATE V3 User s Guide The statsCollector 103 statsCollector scans the content of the logBook stream formats the data assembles the data blocks and builds the run records that will be saved in the central bookkeeping repository This daemon can run on any host with online access to
37. performs the following sequence of operations in the order described below e execute the common start of run scripts if any e execute the specific start of run scripts if any e call the user routine ArmHw e build the header of the SOR record and fill the data of the SOR record with the run conditions copying the Run Control shared segment e prepare the common SOR files if any one record per file e prepare the specific SOR files if any one record per file The readout process then initiates the physics events main loop First it reserves the amount of space specified as the Maximum event size parameter in the Run Control panel in RECORDINGBUFFER If it doesn t find it the readout process sleeps until the requested space is available Then it waits for a trigger by calling the user routine EventArrived the arrival of a trigger can signal a physics event or a start of burst SOB or an end of burst EOB The readout process then fills the event header fields for which it is responsible and lets the user fill the event data by calling the user routine ReadEvent and the other event header fields which are under the user s responsibility see Chapter 3 It then checks that the mandatory fields in the event header have been set by the user routines increments the trigger number only for physics events not for SOB and EOB records and fills the variables in the Run control shared segment to allow the updating of the Run
38. readList corbo_lib c Listing 7 4 Bits and masks for the CORBO registers 1 CORBO bits definition 2 define CSRMASK Oxfff 33 define FULLBYTE Oxff 4 define CHANNELENABLE 0x0 5 define CHANNELDISABLE 0x1 6 define BUSYLATCHED 0x0 7 define BUSYNOTLATCHED 0x2 8 define INPUTFRONTPANEL 0x0 9 define INPUTDIFFIN 0x4 10 define INPUTDIFFOU 0x8 TI define INPUTINTERNAL Oxc 12 s define LOCALBUSYOUT 0x0 13 define DIFFBUSYOUT 0x10 14 define COUNTINPUT 0x0 15 define COUNTBUSY 0x20 16 define FASTCLEARENABLE 0x0 Te define FASTCLEARDISABLE 0x40 18 define PUSHBUTTONENABLE 0x0 1 93 define PUSHBUTTONRESETIRO 0x80 20 define INPUTPRESENT 0x100 Zs define LOCALBUSYPRESENT 0x200 22 define DIFFBUSYPRESE 0x400 Zh define IRQPENDING 0x800 7 4 Thggering with the CORBO 7 4 1 Using the CORBO as LDC trigger module In the example described here we use two channels of the CORBO for different purposes 1 Channel 1 is used to trigger the LDC and activate the readout program Interrupts are not used The readout program polls the channel until a trigger has arrived The busy output is used to gate off subsequent triggers until removed by software The internal counter is used to count the triggers received and seen by the LDC The dead time counter is used
39. regPtrChar 0 Clear time out vector 14 regPtrShort unsigned short corboBaseAddr CORBO_TEST1 15 regPtrShort 0 Prepare to trigger 16 regPtrShort FULLBYTE and then trigger once set busy E 17 regPtrShort unsigned short corboBaseAddr CORBO_CNT1 r 18 regPtrShort 0 Clear event counter L93 regPtrShort unsigned short corboBaseAddr CORBO_TOU1 r 20 regPtrShort 0 Clear dead time counter 21 regPtrShort unsigned short corboBaseAddr CORBO_CSR1 r 223 regPtrShort COUNTBUSY INPUTFRONTPANEL Enable front panel trigger Then it initializes the channel 2 of the CORBO as shown in Listing 7 7 ALICE DATE V3 User s Guide Triggering with the CORBO 79 Listing 7 7 CORBO channel 2 initialization 1 Channel 2 initialization Ae Input event number output not used 3 event counter event number all the events 4 regPtrShort unsigned short corboBaseAddr CORBO_CSR2 as regPtrShort INPUTINTERNAL Enable internal trigger 6 regPtrChar unsigned char corboBaseAddr CORBO_BIM1 OFST_CR1 Ts regPtrChar 0 Disable event interrupt 8 regPtrChar unsigned char corboBaseAddr CORBO_BIM1 OFST_VR1 9 regPtrChar 0 Clear event vector LG s regPtrChar unsigned char corboBaseAddr CORBO_BIM2 OFST_CR1 EI regPtrChar 0
40. sc he ae Soe A A PE a OL Figure 7 1 VME to memory mapping 72 Figure 7 2 Trigger electronics ae ye ton e tls a ag SA Figure 8 1 The DATE infoLogger architecture fake Vig fe Xe h e i a EOD Figure 9 1 DATE bookkeeping schematic view 102 Figure 9 2 Detailed view of the bookkeeping system 103 Figure 9 3 The source run record alant e ya LOZ Figure 9 4 Example of standard DATE bookkeeping record f T EL Figure 10 1 A view of the file organization in DATE 115 Figure 11 1 Run control configuration window 1839 ALICE DATE V3 User s Guide 150 List of Figures ALICE DATE V3 User s Guide Listing 2 1 Listing 3 1 Listing 3 2 Listing 4 1 Listing 4 2 Listing 5 1 Listing 5 2 Listing 5 3 Listing 5 4 Listing 5 5 Listing 5 6 Listing 5 7 Listing 6 1 Listing 6 2 Listing 6 3 Listing 7 1 Listing 7 2 Listing 7 3 Listing 7 4 Listing 7 5 Listing 7 6 Listing 7 7 Listing 7 8 Listing 7 9 Listing 7 10 Listing 7 11 Listing 7 12 Listing 7 13 Listing 7 14 Listing 8 1 Listing 9 1 Listing 9 2 Listing 9 3 Listing 9 4 Listing 9 5 Listing 9 6 Listing 11 1 Listing 11 2 Listing 11 3 Listing 11 4 Listing 11 5 List of Listings Installation of the buffer status plug in Event header structure Event types Example of event dump i in iG Example of analysis in FORTRAN Calling EventArrived Calling ReadEvent Example of ReadEvent routine Bookke
41. sibility of monitor dow LDC and ing from a given GDC parameters machine Zero 0 under both gt LDC means disabled PAR and gt GDC one 1 enabled PAR in runControl config ALICE DATE V3 User s Guide Run control configuration 137 Table 11 3 Run control parameters Parameter name Description Set by To appear in window Time recorderSleep enableRecording InBurst logLevel The recorder goes to sleep while events are arriving if enableRecordin gInBurst is dis abled to give priority to read out The time inter val expressed in microsec is picked up from this param eter Default is tuned to 10 millisec Switch to enable and disable the recorder during a burst of events Zero 0 means dis abled one 1 enabled Default is disabled which is adapted to fixed target experiments with bursts Col lider experiments continuous beam should enable it It controls the gen eration of messages by all the date pro cesses running on a DAQ machine Optional the default is well tuned The opera tor via the run con trol Optional The default is adapted to fixed target experiments with bursts The opera tor via the run con trol The operator via the run control Optional Configu ration parameters window LDC parameters under gt LDCPAR in run Control config Optional Configu ration parameters window LDC parameters unde
42. the DATE_SITE_LOGS area The output of the daemon is a set of log files one for each run where in one snapshot it is possible to access a complete run record The log files are saved in a dedicated directory defaulting to DATE_SITE_STATS The whole scheme is illustrated in details in figure 9 2 DATE process DATE process DATE process DATE DATE_SITE_LOGS logBook eee DATE statsCollector DATE bookkeeping bookkeeping repository host DATE statsBrowser other utilities utility DATE operator control offline Figure 9 2 Detailed view of the bookkeeping system 9 2 The statsC ollector The sources of statistics information belonging to a DATE system use the infoLogger primitives to ship records to the infoLogger host This data has to be extracted from the associated log file assembled sorted and saved in a run ALICE DATE V3 User s Guide 104 Guide to use the bookkeeping system 9 2 1 description file This function is performed by the statsCollector daemon a standalone process that can be run on any of the DATE hosts called bookkeeping host in one and only one instance To start the statsCollector the following procedure should be followed 1 if not yet done run the DATE shell setup procedure 2 start the statsCollector from the shell prompt gt statsCollector amp The daemon can be customized using a set of run time flags to see their description use the help o
43. the command monitorClients see 11 7 3 can be used to check for registered clients All the other parameters can be changed at any time and will become active for all new clients producers and consumers started after the modification s When the same parameter is defined in multiple files a last given policy is followed that is e parameters defined in DATE_SITE_CONFIG monitoring config can be overridden by equivalent definitions from any of the other files e parameters defined in S DATE_SITE S DATE_HOSTNAME monitoring config are final for local monitoring and can be overridden by equivalent definitions from etc monitoring config for relayed monitoring e parameters defined in etc monitoring configare final and cannot be overridden Only exception to this scheme is the parameter LOGLEVEL where the highest given level is used e g if the values 0 10 and 20 are specified the used value will be 20 The parameters that can be specified in the configuration files are Table 11 4 Monitoring configuration parameters Parameter name Used for Description LOGLEVEL All types of monitoring Level for error information and debug statements gen erated by the monitoring scheme MAX_CLIENTS Online monitoring only Maximum number of cli ents allowed to be regis tered simultaneously MAX_EVENTS Online monitoring only Maximum number of events available for moni toring EVENT_SIZE Online monitoring on
44. the configuration file Keyword Following items GDCSTATUS The list of control parameters which will be displayed in the GDC part of the window labelled Status display Each item is the conventional name of the parameter followed by a label which will appear in the display window TCLEVAL The list of tcl statements to be executed by the run control program The execution of these statements will happen after the run control has processed the configuration file and is properly configured and before the optional parameter file is read The main use is to plug in additional tcl source code TCLEARLYEVAL The list of tcl statements to be executed by the run control program These statements are immediately passed to the tcl interpreter line by line A mistake in the statements will cause an abnormal exit of the run control program since the proper recovery mechanism is not yet available at this point in time The main use is to define environment variables nec essary to the run control but not available in some environ ments The keywords RUNPAR CNFPAR LDCPAR GDCPAR LDCSTATUS and GDCSTATUS are followed by the declaration of some control parameters The list of the parameters is in Table 11 3 with a description of the parameters and the place where they may be declared The parameter names are conventional and must be used in the declaration The complete list of all the parameters used in DATE can be found in
45. the messages coming from different sources facilities and hosts in centralized data streams e capability to classify the messages according to their source severity and information stream e possibility to follow the evolution of a DATE system while this develops e capability to handle information stream based on local and networked file systems e g AFS and NFS e possibility to grep the information streams in real time using standard Unix tools cat grep tail and or ad hoc filtering facilities e simplified installation use and maintenance e integrated with the DATE environment and the DATE file system structure e capability to gather data coming from all sorts of sources daemons child processes shell scripts interactive processes interactive shell commands written in any of the languages commonly used by DATE developers and DATE users C Java Tcl Tk Unix shells The choice of implementation made for the infoLogger scheme is based on TCP IP sockets see fig 8 1 Each source of messages a generic process running on any of the DATE hosts may open at any time a TCP IP link with a central infoLogger host where a dedicated daemon is started Once a message is shipped from the client process the daemon running on the infoLogger host accepts and validates the data reformats it and stores it into the appropriate information stream The messages are tagged using various bits of information such as the time of cre
46. the repository can be arbitrary big Fragmentation of information on more streams can result in more overhead on the infoLogger host although it may give easier and quicker filtering and browsing of the various records The quantity of information stored in the repository is limited by the file system and by some practical constraints each stream needs a separate polling thread process or process cycle per each browser running on the ALICE DATE V3 User s Guide 88 Guide to use the infoLogger system infoLogger host and this needs the allocation of resources from the O S When the content of the DATE_SITE_LOGS repository becomes too big the following effects can be observed 1 the file system may reach the maximum limit therefore blocking completely the logging scheme the Unix command gt df k DATE_SITE_LOGS shows a used near or equal to 100 2 the infoBrowser tool uses more system resources memory amp CPU especially at startup 3 the infoBrowser tool starts to throw away messages from the display area a warning message is shown on the display 4 the statsCollector daemon see the chapter relative to DATE statistics for more details on the statsCollector tool uses more system resources memory amp CPU at startup only 5 any other browsing filtering grepping tool on the content of the information repository may begin to consume too many system resources CPU memory both at startup and at run
47. them must be created in the library through a set of arrays types and macros defined into the readList_detectors h header file The rules to be fulfilled are the following see example in Listing 6 2 e Types declared in the gt EQTYPES section without NODATA the values returned applying the equipmentDat aType macro to the name of each of these types must be assigned to the equipmentDataTable array of elements of type equipmentDataTableType The variable nbDtEqps must then be set with the number of entries put into the array e Types declared in the gt EQTYPES section with NODATA the values returned applying the equipmentNoDataType macro to the name of each of these types must be assigned to the equipmentNoDataTab1e array of elements of type equipmentNoDataTableType The variable nbNoDtEgps must then be set with the number of entries put into the array e Types declared in the gt TRTYPES section without NODATA the values returned applying the triggerDataType macro to the name of each of these types must be assigned to the triggerDataTab1le array of elements of type triggerDataTableType The variable nbDt Trgs must then be set with the number of entries put into the array 4 The file DATE_ROOT readList readList_detectors h must be included in the library ALICE DATE V3 User s Guide The equipmentList library 67 e Types declared in the gt TRTYPES section with NODATA the values returned applying the triggerN
48. vent gt eventHeader nbInBurst DAQCONTROL gt burstCount vent gt eventH ader burstNb She 38 395 40 vent gt eventHeader siz eventHeaderStruct vlen sizeof struct gettimeofday amp timeStamp 0 vent gt eventHeader tim timeStamp tv_sec vent gt eventHeader usec timeStamp tv_usec ALICE DATE V2 User s Guide Report message to the infoLogger 51 Listing 5 3 Synopsis Description Returns 5 3 Example of ReadEvent routine 1 int ReadEvent struct eventHeaderStruct header_ptr int data_ptr 2 3 Called after a trigger has arrived 4 5 Inserts raw data into the area pointed by data_ptr 6 7 int firstWord data_ptr 8 int dataSize Our ant oF 10 11 fill the header T2 13 header_ptr gt nbInRun DAS o Eio erara A Tos 16 fill the raw data 17 18 dataSize T9 20 for i 0 i lt dataSize 1 i 21 data_ptr i 223 23 returns number of bytes actually readout 24 25 return unsigned long32 data_ptr unsigned long32 firstWord 26 27 DisArmHw include rcShm h include event h void DisArmHw The DisarmHw routine is called at each end of run after the execution of the end of run Unix scripts and before the transfer of the end of run files on the output medium In general it should perform all the actions need
49. viathe Configuration recorded on multi run control parameters win ple files This is the dow Common maximum size of parameters under each file in kBytes gt CNFPAR in run It is used either by Control config the recorder LDC alone or by the event builder GDC bytesInFile Counter of kBytes Either readout or Status display win recorded in the cur event builder dow under either rent recording file gt LDCSTATUS or It is set either by the gt GDCSTATUS in recorder LDC runControl config alone or by the event builder GDC fileCount Counter of number Either readout or Status display win of files recorded in event builder dow under either the current run It is gt LDCSTATUS or set either by the gt GDCSTATUS in recorder LDC runControl config alone or by the event builder GDC recordingDevice A character string The operator viathe Configuration indicating the run control parameters win recording device of dow LDC parame an LDC It may be ters under either a file name or gt LDCPAR in run a machine name Control config the latter followed by ebRecordingDe A character string The operator viathe Configuration vice indicating the run control parameters win recording device of dow GDC param an GDC It must be eters under a file name gt GDCPAR in run Control config monitorEnable Switch to enable The operator viathe Configuration Flag and disable the pos run control parameters win
50. x7 CO CO CO CO CO CO CO CO CO CO CO CO CO CO OFST_ OFST_ OFST OFST OFST OFST OFST OFST CO CO CO CO CO CO CO CO CO corboDef h RBO_CSR1 RBO_CSR2 RBO_CSR3 RBO_CSR4 RBO_CNT1 RBO_CNT2 RBO_CNT3 RBO_CNT4 RBO_TOU RBO_TOU RBO_TOU RBO_TOU BwN EP RBO_BIM1 RBO_BIM2 RO R1 R2 R3 RO R1 R2 R3 d j aaqaaqaa lt lt 1 lt lt RBO_TES RBO_TES RBO_TES RBO_TES BwON EF RBO_CLEAR1 RBO_CLEAR2 RBO_CLEAR3 RBO_CLEAR4 RBO_RAI OOO SO OOO OOOO Oro SC OGOGO ooo oO O OOG CORBO registers definition x00 status registers 16 bit x02 x04 x06 x10 x14 x18 XLC x20 x22 x24 x26 0x30 WYUWoONOWEeE x58 x5A 5C x60 counters 32 bit time out 16 bit Event interrupt registers Time out interrupt registers Interrupt control Interrupt vector Simulate input trigger 16 bit Clear busy 16 bit Static memory max OxFE 8 bit 16 bit ALICE DATE V3 User s Guide 76 VME access and trigger system The bits and masks used in the CSR and other registers are defined in Listing 7 4 An earlier definition file can be found in date readList corbo h An example of all the functions that can be performed on the CORBO can be found in date
51. 27 list of 131 parameters 131 132 gdcServer program 3 I info routine 94 infoAll routine 96 infoBrowser program 6 86 infoBrowser symbol definition 117 infoDaemon program 6 infoLogger daemon 87 infoTo routine 95 Installation First time 125 of new release 129 K kits Distribution 124 Index L LDC iii 2 3 4 10 20 21 23 25 58 60 70 74 76 78 81 110 124 127 130 131 134 135 136 137 138 list of 131 parameters 131 132 Log files extracting portions of 89 repository 87 log routine 93 logAll routine 94 logBook routine 108 logBookMarker routine 108 logTo routine 93 M MONITOR_FREE_EVENT routine 37 monitorControlWait routine 38 monitorDeclareMp routine 34 monitorDeclareTable routine 34 monitorDecodeError routine 39 monitorFlushEvents routine 37 monitorGetEvent routine 36 monitorGetEventDynamic routine 36 monitorLogout routine 40 monitorSetNowait routine 38 monitorSetSwap routine 38 monitorSetWait routine 37 monServer application 28 mpDaemon program 4 P program ebDaemon 5 program eventBuilder 3 5 program gdcServer 3 program infoBrowser 6 86 program infoDaemon 6 program monServer 28 program mpDaemon 4 program rcServer 5 program readout 4 5 45 program recorder 3 5 20 program statsBrowser 6 105 program statsCollector 6 103 R rcServer program 5 ReadEvent routine 48 readlist routines 47 readout program 4 5 45 Readout routine ArmHw 47 ALICE DATE V3
52. 2k aaa 76 7 5 Using the CORBO to control the trigger 81 ALICE DATE V3 User s Guide 72 VME access and trigger system 7 1 Access to the VME bus Figure 7 1 Listing 7 1 There are various ways of accessing the VME The one described here is the direct mapping from the application program It consists of assigning a virtual address window to a region of the VME space as shown in Figure 7 1 Virtual address space OxFFFFFFFF 32 bit VME OxFFFFFF 24 bit Device vmeWinSize aDeviceVmeBaseAddr vmeWinOffset aDeviceBaseAddr 0x0 vmeWinAddr VME to memory mapping We shall describe how to operate an hypotethic device called aDevice It is convenient to indicate the offset addresses of the device registers in a header file such as the one shown in Listing 7 1 Definition of the offset addresses of a device U BUMD aDevice h Register address definition define A_DEVICE_CSR 0x00 define A_DEVICE_INPUT_REG 0x04 define A_DEVICE_OUTPUT_REG 0x08 The application code maps a VME window onto a virtual address window and then accesses the registers with normal memory references as shown in Listing 7 2 Before termination the program unmaps the VME window The mapping and unmapping are performed by external routines Simple arithmetics leads to the calculation of the device base address ALICE DATE V3 User s Guide
53. 3 Browsing HIML help pages The infoBrowser help pages are written in plain Hypertext Upon request the infoBrowser starts the Help browser by trying to reuse any Netscape window already opened on the X11 server and if none are available starts a new local instance of Netscape This works only if Netscape has been installed on the online host We recommend to install Netscape on the online host this will allow quick and easy browsing of the complete documentation kit 8 4 The information repository The infoLogger daemons write their messages in log files stored in the DATE_SITE_LOGS repository The same area is used by the infoBrowser if used in the default configuration to get its input data The information repository contains 1 the common file DATE_SITE_LOGS runLog shared between all DATE facilities and processes and used for all sorts of generic information 2 an arbitrary number of package specific package streams per each DATE package the name of the stream is given at run time by the infoLogger client 3 anoptional DATE_SITE_LOGS logBook stream used for DATE statistics The number of streams and their identifier is decided at run time by each client of the infoLogger package It is obvious that not all the DATE processes packages need to open an infoLogger channel the decision to use a given stream can be taken on the fly and introduces no extra overhead The number of information streams stored in
54. 7 4 4 Trigger processing The routine EventArrivedCorbo is repeatedly called by the readout program The purpose of this routine is to return a non zero code when a trigger has fired the CORBO The method used is to check the status of the CSR to see whether the busy signal is on as shown in Listing 7 10 Initially the CORBO is ready not busy It reverts to busy when the trigger signal arrives to the input Then the return code being not null the readout program will call the ReadEvent routine to perform the readout operations When the he ReadEvent routine has completed the routine EventArrivedCorbo starts again being repeatedly called The first time it will reset the busy condition Listing 7 10 The routine EventArrivedCorbo int EventArrivedCorbo Returns a value gt 0 when a trigger has occurred int value if eventArrivedFlag 1 First time after a trigger or very first regPtrShort unsigned short corboBaseAddr CORBO_CLEAR1 regPtrShort FULLBYTE Clear busy regPtrShort unsigned short corboBaseAddr CORBO_CSR1 OB WN EF WO On AIO value regPtrShort amp LOCALBUSYPRESENT DIFFBUSYPRESENT if value 0 eventArrivedFlag else eventArrivedFlag 1 return eventArrivedFlag OMDANAUIHPWNE CO 7 4 5 Reading the CORBO counters The CORBO counters may be readout This is usually done in the routine ReadEv
55. A a a A eS ot 89 11 5 5 Multiple run controls aa amp amp see amp vy amp 2 MO 11 6 Information logger configuration 140 11 7 Monitoring configuration do amp apa Se ae ee ae AO 11 7 1 Creation of configuration files 141 11 7 2 Installation of the monitoring daemon 143 11 7 3 Monitoring of the online monitoring scheme 144 Index 3 ong oasi di kee fe ae he A Se ee s Se a i LD Listof Figures o yos m Se ok ces BE ch ewe oy we peda Oe we we AD Listof Listings soa e m oe d for aw Ee oe ee ee LDL Eistof Tables a mo a es ee aot Bhd Go ee me Se he me me eG So ABS ALICE DATE V3 User s Guide ALIC E DATE architecture This chapter gives an overview of the architecture of DATE The features of the system are described with the components that implement such features For each component a brief explanation of the underlying mechanism is given 1 1 OvervieW 6 6 bk eee eR Re ee hee ee eS 2 1 2 Readout and dataflow 22200 3 1 3 Event MOnitori g lt sses essere bvr nerti nisn 4 14 Runicontrol ede sie y en ra ee eek wR ee 5 1 5 Information logging we Ware oe BO eR 5 1 6 Run bookkeeping 2222444242 bet44h 244 bb dc 6 ALICE DATE V3 User s Guide ALICE DATE architecture 1 1 Overview DATE Data Acquisition and Test Environment is a software system that performs data acquisition activities in a multi processor dis
56. ALICE 98 44 Intemal Note DAQ ALICE DATE User s Guide November 1998 DATE V3 Copyright CERN Geneva 1998 Copyright and any other appropriate legal protection of this documentation and associated computer program reserved in all countries of the world Organisations collaborating with CERN may receive this program and documentation freely and without charge CERN undertakes no obligation for the maintenance of this program nor responsibility for its correctness and accepts no liability whatsoever resulting from its use Program and documentation are provided solely for the use of the organisation to which they are distributed This program may not be copied or otherwise distributed without permission This message must be retained on this and any other authorised copies The material cannot be sold CERN should be given credit in all references Creation date 11 10 98 This document has been prepared with Release 5 5 3 of the Adobe FrameMaker Technical Publishing System using the User s Guide template prepared by Mario Ruggier of the Information and Programming Techniques Group at CERN Only widely available fonts have been used with the principal ones being Running text Palatino 10 5 pt on 13 5 pt line spacing Chapter numbers and titles AvantGarde DemiBold 36 and Bold 56 pt Section headings AvantGarde DemiBold 20 pt Subsection and subsubsection headings Helvetica Bold 12 and 10 pt Captions Helvetica 9 pt
57. Book call The value of the delay timer of the statsCollector can be changed at will use the help flag for a complete description of the relative option Multi line records examples We will now see three examples of multi line bookkeeping records This mechanism can be used to transfer all the information in one block and with a single TCP IP message It is obvious that all the required information must be available at the time of the call ALICE DATE V3 User s Guide 110 Guide to use the bookkeeping system Listing 9 4 Example of multi line C bookkeeping record char line 1000 sprintf line Run d n First line n Second line n Third and last line n runNb runNb LOGBOOK line run number Listing 9 5 Example of multi line Java bookkeeping record diy logChannel logBook Run runNb First line n Second line n Third and last MEFHEHt tet tt tt n runNb run number line n Listing 9 6 Example of multi line Bourne shell bookkeeping record Daw wWNEH gt runNb gt logBook First line Second line Third and last line FEFEEEEFEFEFF ATT E run number Run S runNb The shell example should be taken with care as the way each shell breaks multi line strings may vary from one implementation to another The snipped in listing 9 6 is valid f
58. EANING size 32 bits size of event in bytes set by the readout process magic 32 bits number used for consistency check an to discover the byte ordering set by the readout process type 32 bits type of record see Table 3 2 headLen 32 bits size of header in bytes set by the readout process runNb 32 bits burstNb 32 bits nbInRun 32 bits nbInBurst 32 bits triggerNb 32 bits fileSeqNb 32 bits run number set by the readout process burst number initialized to 0 by the readout process This field is set by the user routine ReadEvent event number within run initialized to NOT_SET_TAG by the readout process and set by the user routine ReadEvent The readout process logs an error mes sage if this field is not filled upon return from the rou tine This is the unique number identifying the event event number within burst initialized to 0 by the read out process This field is set by the user routine ReadE vent trigger number for this detector this field is incremented by the readout process only for PHYSICS_EVENT type of records sequence number of the raw data file containing this event This field is set by the recording library in the recorder or the event Builder process ALICE DATE V3 User s Guide The event types 23 Table 3 1 Event header fields FIELD MEANING detectorId 96 bits time 32 bits usec 32 bits errorCode 32 bits deadTime 32 bi
59. HARACTER 14 DEFAULT_TABLE All_events yes Zero in case of success otherwise a error code see 4 2 3 for more details monitorGetEvent include monitor h int monitorGetEvent void buffer long size INTEGER MONITOR_GET_EVENT INTEGER INTEGER The next available event if any is copied in the region pointed by buffer fora maximum length of size bytes In case of failure a zero length event is returned Zero in case of success otherwise a error code see 4 2 3 for more details monitorGetEventDynamic include monitor h int monitorGetEventDynamic void buffer ALICE DATE V3 User s Guide Monitoring and Analysis in CIFORTRAN 37 FORTRAN Synopsis Description Returns FORTRAN Synopsis Description C Synopsis FORTRAN Synopsis Description Returns C Synopsis FORTRAN Synopsis INTEG ER MONITOR_GET a ENT DYNAMIC POINTER The next available event if any is copied on space reserved from the process heap and returned to the caller The caller must take care of properly disposing the event via the free system call C library or via the MONITOR_FREE_EVENT call FORTRAN library failure to do so will exhaust the resources associated to the process and can severely degrade the overall system performances If no data is ava
60. LICE DATE V3 User s Guide Guide to operate the system 2 1 The contol console Overview The control console is a machine from which the data acquisition is operated using the run control system The control console is unique and cannot be changed during a run One of the processors involved in the data acquisition may be used for this purpose It is though preferable to give the role of control console to another workstation to avoid mixing control and dataflow activities in one processor The selected workstation should have access on its mounted disks to the two file areas used by DATE which are pointed to by the environment variables DATE_ROOT and DATE_SITE Preliminary condition to start the run control system are the following 1 The machine which owns the display which is not necessarily the control console when a remote shell is used must authorize all the other DATE machines to open windows on the screen e g with the command xhost 2 The shell symbol DATE_SITE must point to the directory where the DATE experiment specific files reside 3 The procedure date setup sh must be executed once in order to define all the symbols required The run control system is started by the shell command gt dateControl which launches the run control program on the control console The visible effect of this command is the creation on the main run control window on the screen Figure 2 1 The specific items shown on th
61. RAN Synopsis Description Returns Pointer to a C zero terminated static read only string C library only monitorLogout include monitor h int monitorLogout void INTEGER MONITOR_LOGOUT The monitoring link is closed and all resources allocated for this monitoring program are freed The link will be automatically re opened when the monitoring program will request the next event This entry can be used whenever the monitoring program expects long pauses such as operator input It imposes a certain overhead on the monitoring scheme and therefore should not be used too frequently Zero in case of success otherwise a error code see 4 2 3 for more details 4 3 The eventDump utility program Part of the standard DATE kit is the utility eventDump This image allows easy monitoring of any stream useful for a quick check or for debug of a running system The standard DATE kit provides a version of the eventDump utility for each architecture fully supported or only with monitoring support see Table 11 1 To run the utility on DATE hosts execute the standard DATE setup and issue the command S eventDump buffer For non DATE hosts copy the utility in your PATH or declare a proper alias and then issue the same command as for DATE hosts A list of all available options can be shown via the command line flag Some of the parameters are e b brief output does not display event dat
62. ROOT pointing to the site independent directory DATE_SITE pointing to the site specific directory DATE_SITE_LOGS pointing to DATE_SITE logFiles DATE_SITE_CONFIG pointing to DATE_SITE configurationFiles DATE_HOST indicating the host machine name ALICE DATE V3 User s Guide Internet d mons 117 10 4 10 5 DATE_SYS indicating the operating system of the host machine In addition the following aliases are defined and may be used as commands dateCont rol to start the run control program readoutCont rol to start the readout configuration program eventDump to start the event dump program infoBrowser to start the browser of the messages of the logging system dateSetup to set up the specific environment for a package dateSymbols to display all the common DATE environment variables Intemet dae mons There are several daemons used in DATE They make use of the Internet services d mon inetd The inetd services are specified by the following files etc services which declares the port number of all the services etc inetd conf which declares the server program to be invoked for all the services In DATE the declared servers are shell scripts which first set up the environment variables for the package concerned and then call a program The setup sh script defines some environment variables corresponding to the inetd services DATE _SOCKET_RCS 6101 DATE _SOCKET_EB 6201 DATE SOCKET _INFOLOGGER 6301 DATE SOCKE
63. Server is created when the recorder opens the socket The gdcServer gets the events from the socket and fills a circular buffer A process called eventBuilder off loads the buffer and sends the events to whatever device has been indicated in the configuration It usually is a disk file The function of the eventBuilder process is to collect the sub events from the various LDC and to build the full event figure 1 3 In the GDC there is one gdcServer and one circular buffer per LDC The eventBuilder simply goes around all the buffers and picks up all the sub events ALICE DATE V3 User s Guide ALICE DATE architecture Figure 1 3 1 3 Figure 1 4 Event builder architecture Event monitoring The architecture of the monitoring function is shown in figure 1 4 An analysis process may request events from any data acquisition machine by calling the monitoring library routines A buffer reserved to the monitoring function is filled with the requested events either by the readout program if the machine is an LDC or by the eventBuilder if the machine is a GDC No copy is made if there is no pending request Monitoring architecture The analysis process may run locally on the machine producing the events in this case it will get the events directly from the monitoring buffer It may also run on any other workstation even if the workstation does not share the DATE file system it will then get the events over the networ
64. T _EBDAEMON 6401 DATE_SOCKET_MON 6601 The DATE packages use these variables when declaring the socket port Remark all the port numbers are found also in the file S DATE_COMMON_DEFS shellParams common Since the environment variables are defined by setup sh or setup csh you must not modify the port numbers in etc services unless you modfy date setup sh accordingly Package development By convention the packages access the resource they need at run time either files or sockets through environment variables ALICE DATE V3 User s Guide 118 Conventions and file organization Let us assume that the package is called package and its abbreviateed acronym is pk Before using a package it is necessary to invoke dateSetup date package packageParams As a result all the environment variables necessary to the package will be defined Their name will be of the type DATE_pk_variable Each package must use a private definition of the run time directories for example DATE_pk_CONFIG points to DATE_SITE_CONFIG DATE_pk_SITE points to DATE_SITE The same holds for the variables used by the GNUmakefile to locate the files For example DATE_pk_ROOT points to the package main directory DATE_pk_BIN points to DATE_pk_ROOT DATE_SYS which is the run time directory DATE_pk_INC points to the include files DATE_pk_OBJS points to the object files It is possible to develop and test a package in a private directory whith
65. TCP IP e The hosts may be of different brands and run different operating systems e The hosts may share the same file system e Several independent experiments may run at the same site They share the same cluster and the same file system Each host will be allocated to only one experiment and all the experiments will use the same version of DATE e The installation of the DATE software on a site is made by bringing the full set of DATE packages called the distribution kit which contains all the site independent run time files plus all the source files and some templates of the site specific files The distribution kit is complete i e the DATE software can be re built from scratch starting from the kit The installation is decided and performed by the site manager The procedure does not disrupt any of the DATE run time system files specific to the experiment e Each DATE package can be developed and tested in conjunction with the standard running system DATE remains fully operational while a developer runs own versions of one or more packages File organization The DATE run time system makes use of two distinct areas on the file system figure 10 1 The first area contains all the files belonging to the DATE distribution kit It must reside in a directory called date and is pointed to by the environment variable DATE_ROOT The site manager may decide to put it elsewhere then a symbolic link date must be created to the area T
66. VEN Header length detectorld 0000 0000 0800 Subdetector B data The full event format Full event from GDC Event length ype PHYSICS_EVENT Header length detectorld 8000 0000 8001 Subevent A length pe PHYSICS_EVENT detectorld 0000 0000 0001 Subdetector A data Subevent B length type PHYSICS_EVENT Header length detectorlId 0000 0000 0800 Subdetector B data ALICE DATE V3 User s Guide Guide to wnte a monitonng program This chapter describes how to write a monitoring program After a brief introduction to the monitoring in DATE the monitoring library is explained and its use from all the most commonly used programming languages 4 1 Monitoring in DATE 3 2 248 a kee be Poe os 28 4 2 Monitoring and Analysis in C FORTRAN 30 4 3 The eventDump utility program 40 ALICE DATE V3 User s Guide Guide to write a monitoring program 4 1 Monitoring in DATE A Data Acquisition system such as DATE may require monitoring of experimental data online and offline data on online and offline hosts Some possible applications for this kind of monitor tasks are e statistical analysis of the experimental stream to evaluate the quality of the physics conditions e detailed analysis of the experimental data to extract specific information such as configuration occupancy and efficiency of the hardware e occasional checking of the overall status of th
67. VME window onto a memory window A pointer to the first 32 bit word is returned unsigned long vmeWinAddr vmeWinAddr unsigned long MapVME vmeWinOffset vmeWinSize Set the vme base address of the CORBO It may depend on a switch setting on the board unsigned int corboVmeBaseAddr corboVmeBaseAddr 0x700000 Calculate the memory base address of the CORBO corboBaseAddr unsigned int vmeWinAddr corboVmeBaseAddr unsigned int vmeWinOffset Then it initializes the channel 1 of the CORBO as shown in Listing 7 6 ALICE DATE V3 User s Guide 78 VME access and trigger system Listing 7 6 CORBO channel 1 initialization 1 Channel 1 initialization 2 Input trigger output busy 3 event counter trigger number events on this LDC 4 regPtrShort unsigned short corboBaseAddr CORBO_CSR1 53 regPtrShort COUNTBUSY INPUTINTERNAL Enable internal trigger 6 regPtrChar unsigned char corboBaseAddr CORBO_BIM1 OFST_CRO Ts regPtrChar 0 Disable event interrupt 8 regPtrChar unsigned char corboBaseAddr CORBO_BIM1 OFST_VRO gs regPtrChar 0 Clear event vector 10 regPtrChar unsigned char corboBaseAddr CORBO_BIM2 OFST_CRO 11 regPtrChar 0 Disable time out interrupt 2s regPtrChar unsigned char corboBaseAddr CORBO_BIM2 OFST_VRO 13
68. a e c check events data against a pre defined data pattern test environment only e s use static data buffer rather then dynamic memory e a use asynchronous reads nowait mode e i interactive pauses after each event and proposes a mini menu with several options e t allows the declaration of a monitoring table e g ALICE DATE V3 User s Guide The eventDump utility program 41 t SOR all EOR all will show all Start of Run and End of Run events available skipping all the other events The buffer parameter must always be specified The syntax to be used is the same as for the parameter of the monitorSetDataSource entry see Table 4 1 ALICE DATE V3 User s Guide 42 Guide to write a monitoring program ALICE DATE V3 User s Guide Guide to prepare a readout program This chapter describes the software running in the front end crates In particular it explains how to customize it and how to build the readout program responsible for performing the hardware readout 51 OVERVIEW so ee eae Sane oS ee oe eS ee 44 5 2 The readList templates oaa Boke eek A ee 47 5 3 Report message to the infoLogger 51 5 4 Customization of the front end software 52 5 5 How to build and install a readout program 54 ALICE DATE V2 User s Guide 44 Guide to prepare a readout program 5 1 Overview When the run starts a process called reServer lau
69. able start a new local instance of Netscape This obviously works only if Netscape has been installed on the online host We recommend to install Netscape on the online host this will allow quick and easy browsing of the complete documentation kit The bookkeeping repository The statsCollector and the statsBrowser use if running in their default configuration the DATE_SITE_STATS repository to respectively write and get the bookkeeping records This area contains a flat set of files one per each run record whose filename follows the syntax DATE_SITE_STATS daqStats log where is the run number In theory the only limits imposed on the number of records and to the quantity of information stored in this repository are imposed by the host s file system features What we suggest to do is to create at the end of each beam period a save set of all the run records tar zip or any other archival tool will do Once the DATE_SITE_STATS area has been archived the run descriptor files can be all removed without problems How to re create the DATE_SITE_STATS repository To re create the content of the DATE_SITE_STATS repository it is necessary to recover the file DATE_SITE_LOGS logBook relative to the complete beam period to re create Multiple logBook files can be merged via the following command gt cat logBook1l logBook2 sort uniq gt logBookDir logBook At this point it is possible to re crea
70. al04 Front end 4 3 HES 16 gt GDC 17 mppcal05 Event builder 18 19 gt RUNPAR 20 maxEvents Max number of events 21 22 gt CNFPAR 23 maxBytes Max kBytes to record 24 maxEventSize Max event siz bytes 25 maxFileSize Max file size kBytes 26 phaseTimeoutLimit Max time for SOR EOR phases sec 27 28 gt LDCPAR 29 recordingDevice Recording device 30 monitorEnableFlag Monitor enable flag 31 randEventMinSize Randev min size 32 randEventMaxSize Randev max size 33 randEventInterval Randev interval 34 logLevel Logging level 35 36 gt GDCPAR 37 ebRecordingDevice EVB recording device 38 monitorEnableFlag Monitor enable flag 39 logLevel Logging level 40 41 gt LDCSTATUS 42 eventCount Number of events 43 triggerCount Number of triggers 44 eventsTransferred Events transferred 45 bytesTransferred kBytes transferred 46 readoutFlag SOR EOR phase 47 48 gt GDCSTATUS 49 eventCount Number of events 50 ventsRecorded Events recorded 51 bytesRecorded kBytes recorded 52 monitorEnableFlag Monitor enable flag 53 fileCount File count 54 bytesInFile kBytes in file DD 56 gt TCLEVAL 57 set ldcEventBufferSize 1024000 58 source date runControl bufferStatus tcl Valid keywords are the ones shown in table 11 2 ALICE DATE V3 User s Guide Run con
71. alysis in FORTRAN as PROGRAM HSIMPLE 2 ies INTEGER VECTOR 100000 aes INTEGER STATUS a INTEGER get_event 6 PARAMETER NWPAWC 10000 es 8 CALL monitor_set_data_source tmp runData Ore CALL monitor_declare_mp FORTRAN demo mp 10 CALL HLIMIT NWPAWC LIS CALL HTITLE Example bas CALL HBOOK1 10 Event size dist 50 0 20000 0 13 DO 10 I 1 100 14 STATUS get_event VECTOR 100000 4 5 s IF STATUS LEs 0 STOP 16 CALL HF1 10 FLOAT VECTOR 1 1 hs 10 CONTINUE 18 CALL HPRINT 10 T9 STOP 20 END Other examples are available in the directory date monitoring namely the files eventDump c and eventDumpFtn f The monitoring package files The distribution point for the monitoring package is DATE_MONITOR_DIR defined by the DATE setup procedure In this area it is possible to find the following files e DATE_MONITOR_DIR monitor h prototypes and definitions for monitoring programs written in C e DATE_MONITOR_DIR DATE_SYS libmonitor a monitoring library for any language capable of calling C code e g C and FORTRAN e DATE_MONITOR_DIR DATE_SYS libmonitorstdalone a monitoring library with reduced functionality for non SHIFT hosts see below Compilation of C monito
72. arations refer to types of equipments for which the driving software is provided in the equipmentList c library The actual instantiation of the equipments is declared under the various detector declarations Trigger equipments types are declared using the same format as for the readout equipment types described above 6 6 1 The difference between trigger equipments and readout equipments concerns the number of routines to be provided in equipment List c and the how they will be called by the generic readList see 6 5 1 and 6 5 2 ALICE DATE V3 User s Guide 70 The generic readList 6 6 3 6 6 4 The detector parameters The list of detector specific parameters is declared under the heading gt DEPARAMS These are parameters that can be accessed from all the equipment handling routines of a detector The list of parameters is common to all the detectors while the values are assigned detector by detector The parameters are declared one per line by specifying for each of them the name and the format either hd 1d shx 31x s or c according to the C language definition of these constants The detectors The detectors description appears under the heading gt DETECTORS The description of a detector consists of its name the detector specific parameter values and the list of its equipments with their equipment specific parameter values The description begins with a line with the format gt detName Label wh
73. art of run in the shared memory segment of the run control of all the machines LDCs and GDCs Each item is the conventional name of the parameter followed by a label which will appear in the main window The list of control parameters that will appear in the window labelled Configuration parameters under the head ing Common parameters The treatment is the same as for the RUNPAR parameters the only difference is the location on the screen which implies less visibility for the operator The list of control parameters which will appear in the LDC part of the window labelled Configuration parame ters These parameters will be set at start of run in the shared memory segment of the run control of all the LDCs Each item is the conventional name of the parameter fol lowed by a label which will appear in the parameter window The list of control parameters which will appear in the GDC part of the window labelled Configuration parame ters These parameters will be set at start of run in the shared memory segment of the run control of all the GDCs Each item is the conventional name of the parameter fol lowed by a label which will appear in the parameter window The list of control parameters which will be displayed in the LDC part of the window labelled Status display Each item is the conventional name of the parameter followed by a label which will appear in the display window 132 DATE installation guide Table 11 2 Keywords of
74. ata recording and data format Guide to prepare a readout program and DATE installation guide the following chapter have been completely rewritten Guide to write a monitoring program two new chapters have been added The generic readList and on VME access and trigger system This User s Guide can be found in the ALICE web site ALICE home page gt Documents gt Internal Notes DAQ gt INT 98 44 or http consult cern ch alice Internal_Notes 1998 44 abstract ALICE DATE V3 User s Guide iv Preface ALICE DATE V3 User s Guide Contents Preface Chapter 1 ALICE DATE architecture 1 1 Overview 1 2 Readout and data flow 1 3 Event monitoring 1 4 Run control 1 5 Information logging 1 6 Run bookkeeping Chapter 2 Guide to operate the system 2 1 The control console 2 1 1 Overview 2 1 2 The menu bar 2 1 2 1 The File menu 2 1 2 2 The View menu 2 1 2 3 The option menu 2 1 2 4 The windows menu 2 2 The infoBrowser console 2 2 1 The infoBrowser operator window 2 3 The statsBrowser console Chapter 3 Data recording and data format 3 1 Data recording 3 1 1 Data recording from i a LDC 3 1 2 Data recording froma GDC 3 2 Data files 3 3 The data format 3 4 The event types 3 5 The full event format Chapter 4 Guide to write a monitoring program 4 1 Monitoring in DATE 4 2 Monitoring and Analysis in C FORTRAN 4 2 1 Some simple examp
75. ation the client DATE hostname the source process ID the source facility the package or utility that created the message the severity of the message and the user owner of the DATE process if applicable Each facility usually associated to a DATE package can open at any given time several information streams from processes running simultaneously on multiple hosts e g readout b from processes running simultaneously on the same host e g a common library used by two or more packages c from a single process separate streams for different purposes debug statements statistics records run time errors etc ALICE DATE V3 User s Guide Introduction 85 Each DATE process can have at any given time at most one TCP IP channel open with the infoLogger host DATE host1 DATE host2 DATE hostN DATE Process DATE Process DATE Process infoLogger daemon DATE infoLogger DATE infoLogger daemon daemon DATE_SITE_LOGS l l l l gt logFilel q 4 gt m runLog zt logFile2 l _logFileN infoLogger Beg ee ge information repository other utilities DATE infoBrowser utility DATE operator developer Figure 8 1 The DATE infoLogger architecture The set and the syntax of the information stream identifiers is completely free and can be reconfigured at will at any time Therefore all tools who need t
76. atsCollector will create the statistics records relative to the runs stored in the file offlineLogs logBook saved copy of an old run It is also possible to direct the output of the daemon to a separate repository area via the outputDir flag This avoids possible confusions between online and offline stats records The statsCollector configuration file It is possible to fine tune the stat sCollector daemon using a dedicated configuration file One of the optional parameters for the daemon can be used to specify the path to this file use the help flag for a complete description of the relative option The content of the configuration file can be 1 Empty or comment lines starting with 2 Output order directives outputOrder facility ALICE DATE V3 User s Guide The statsBrowser This directive can be used to order the contributions coming from the various facilities using a fixed order e g outputOrder readout recorder eventBuilder pds will create records with the contribution of readout from all hosts followed by recorder and eventBuilder Finally all other contributions will be appended in the given order excepted for the section generated by the pds facility that if present will be always put as last 9 3 The statsBrowser The information gathered from the statsCollector can be browsed using the statsBrowser This tool can run on any system capable to open a link towards an X11 server and where t
77. bol ALICE DATE V3 User s Guide use of 52 53 55 70 DATE_INFOLOGGER_LOGHOST symbol use of 127 DATE_ROOT symbol definition 116 use of 8 66 114 115 116 125 129 DATE_SITE symbol definition 116 use of 8 52 53 54 55 60 87 114 116 118 125 126 128 DATE_SITE_CONFIG symbol definition 116 use of 8 52 53 60 62 68 108 127 129 139 DATE_SITE_LOGS symbol definition 116 use of 15 87 88 89 91 98 99 102 103 104 106 107 DATE_SITE_STATS symbol use of 103 104 106 107 DATE_SYS symbol definition 117 use of 93 118 128 dateControl symbol definition 117 dateSetup symbol definition 117 dateSymbols symbol definition 117 DisArmHw routine 44 DisarmHw routine 51 Distribution kits 124 E ebDaemon program 5 End of run 11 20 21 46 command 11 DisarmHw 64 DisarmHw routine 51 146 files 24 53 in readout 44 45 maximum file size 21 operations performed at 45 records 24 script 52 53 Error messages browsing 15 86 102 110 from the readout 51 error routine 94 errorAll routine 96 errorTo routine 95 Event type CALIBRATION_EVENT 24 35 END_OF_BURST 24 35 END_OF_RUN 24 35 END_OF_RUN_FILES 24 35 PHYSICS_EVENT 23 24 35 START_OF_BURST 24 35 START_OF_RUN 24 35 START_OF_RUN_FILES 24 35 EventArrived routine 48 eventBuilder program 3 5 eventDump symbol definition 117 F fatal routine 95 fatalAll routine 96 fatalTo routine 95 G GDC iii 2 3 4 10 19 20 21 25 110 124 1
78. ceOutputReg 36 aDeviceCsr 37 unsigned short aDeviceBaseAddr A_DEVICE_CSR 38 aDevicelInputReg 39 unsigned short aDeviceBaseAddr A_DEVICE_INPUT_REG 40 aDeviceOutputReg 41 unsigned short aDeviceBaseAddr A_DEVICE_OUTPUT_REG 42 43 Beginning of processing 44 int value 45 aDeviceCsr 0 Clear CSR 46 value aDeviceInputReg 0x2 Read input 47 aDeviceOutputReg 0x2 Output bit 1 48 End of processing 49 50 Unmap the memory window Siz UnmapVME 52 The routines MapVME and UnmapVME depend on the operating system The routines to be used in AOS the IBM AIX for the Motorola VME processor boards can be found in the file date readList vme2utils c ALICE DATE V3 User s Guide VME access and trigger system 7 2 The tiggersystem The trigger system provides the synchronization between the experiment and the data acquisition When an event has been collected by the detectors a trigger signal is sent to each LDC to activate the readout program The EventArrived routine handles the trigger hardware and provides the synchronization The EventArrived strongly depends on the triggering method adopted and on the hardware modules It is possible to wait for interrupts if the trigger modules can generate them DATE has been design in a way that permits to poll the status of a register instead of waiting for an interrupt Polling is much faster tha
79. cl eval rsald01 IBM1 ALD lab Figure 2 5 The view menu The main run control window with the additional panels is shown in Figure 2 6 1 Parameters file name If enabled an additional panel will display the name of the parameters file that was used last Commands that set this name are the menu options File Save and File Load and the PARFILE statement in the configuration file The file name can be changed in the interactive entry field the new value is used as default in the menu options File Save and File Load Trace If enabled an additional panel will display a list of information messages internally generated by the control program The most recent messages are on top of the list Buttons allow the operator to pause and clear the trace for debugging purposes and to get more details by enabling a debug mode Tcl eval If enabled an additional panel will allow the operator to dispatch an expression to the Tcl interpreter for evaluation The expression may be written on the entry field then it will be evaluated when the eval button will be depressed This facility has the same effect as a TCLEVAL statement in the configuration file It may be used to add new functions to the control program while it is running by evaluating source Tcl statements referring to the plug in Tcl source file Parameters file name Tcl command TEE PA ITIO bere Cr eee eves Guitar loppi OK ry ew feb ITZE bec Prom mrbin Dent bade so
80. cs stream tcp nowait nobody date runControl OS rcServer sh rcServer date_site 5 date_info stream tcp nowait nobody date infoLogger OS loggerDaemon loggerDaemon 6 date_evb stream tcp nowait nobody date eventBuilder OS GDCserver sh GDCserver date_site 7 date_evbd stream tcp nowait nobody date eventBuilder OS eventBuilderDaemon sh eventBuilderDaemon date_site 8 date_mon stream tcp nowait nobody date monitoring OS mpDaemon sh mpDaemon date_site Note that the symbol date_site represents the real name of the directory indicated by the symbol DATE_SITE the symbol DATE_SITE cannot be used in this context Conversely the symbol OS points to the machine s operating system available within the symbol DATE_SYS The service defined in line 4 is needed on GDCs and LDCs The service defined in line 5 is required by the infoLogger host The services defined in line 6 and 7 are needed on the GDCs Finally the service defined in line 8 is needed by all machines who wish to offer monitoring streams The file etc services of each machine where the INETD services are defined will contain the code fragment in listing 11 2 Listing 11 2 INETD services o O a e w a He DATE servers date_rces 6101 tcp DATE Run Control Server date_evb 6201 tcp DATE Event Builder GDC server date_info 6301 tcp DATE info logger daemon date_evbd 6401 tcp DATE Event Builder Daemon
81. ctivity or a significant part of it It is used to single out the messages generated by one activity when they are either intermixed with other activities in the same stream or dispersed in several streams whose routines may be used by many different programs Usually for a given program package stream and facility have the same name Filtering the logged messages at the source the log level Critical real time processes may require to select at run time whether to log messages In case of testing or debugging there should be the possibility of selectively logging some of the messages of any severity at variable levels of detail The designer will be aware that additional information will increase the overhead This selection mechanism called log level can only be implemented in the code of each program and cannot be generalized The selection cannot be done within the logging routines since the designer may want to dynamically exclude fragments of his own code as well There is a convention between the run control and various programs in the data acquisition to share a variable called logLevel in the rcShm shared memory segment The run control sets this variable at start of run the other programs decide what to do according to the value they find The logLevel space is divided into ranges with a standard meaning with thresholds defined by the symbols LOG_NORMAL_TH 10 LOG_DETAILED_TH 20 LOG_DEBUG_TH 30 1 0 lt 1logLev
82. ded with the distribution kit and describe the electronic set up in another library called equipment List Figure 6 2 The generic readList introduces the following new features a The raw data of an LDC also called detector may be further divided into smaller parts called equipment data The equipment that generates the data may be an electronic board or a set of electronic boards depending on how the readout software is written Each equipment data block begins with an equipment header followed by the equipment raw data In summary a fully built event contains the sub events from the various detectors which in turn contain the data blocks from the equipments see Figure 6 4 b The experiment specific readout software can be separately written for each equipment A set of equipment handling routines similar to the ones of the readList used to arm and disarm the hardware wait for a trigger and read the event deal now with one single equipment thus the code is more modular and readable The code for all the equipments must be merged in one file conventionally called equipment List c c A textual configuration file detectors config unique for the whole experiment contains the list of available equipments for each detector An equipment may be repeated several times in a detector each run time call will be distinguished by a different set of parameters The configuration file specifies the selection of the active equipments and
83. dled by the offline analysis for a single run The data may be saved on several files according to the maxFileSize parameter See also the description in Table 11 3 maxEvent Size It should indicate the maximum size in bytes of the sub event in the LDCs It is not used in the GDCs The ReadEvent routine must make sure not to read more bytes than maxEvent Size otherwise memory corruption may happen On the other hand maxEvent Size cannot be indefinitely large since the same amount of bytes will be wasted at the top of the event buffer in any case it should be one order of magnitude smaller than the event buffer phaseTimeoutLimit See description in Table 11 3 randEventMinSize Optional parameter See description in Table 11 3 randEventMaxSize Optional parameter See description in Table 11 3 randEvent Interval Optional parameter See description in Table 11 3 maxFileSize See description in Table 11 3 recordingDevice To be set for all the LDCs If not set the LDC works in isolation and discards the events It may indicate either a disk file for local recording or the name of the event builder GDC In the latter case it must terminate with ebRecordingDevice Event builder recording device This indicates the directory and the name of the disk file where the full events data should be written There is a special naming convention concerning the recording devices both recordingDevice and ebRecordingDevice when th
84. e Data Acquisition system e g operator status panel To perform these and other functions DATE provides the monitoring package whose objective is to offer a uniform interface for the development and the support of user written monitoring programs tailored to specific needs The monitoring interface implements access to events coming from the live experimental stream or from a Permanent Data Storage PDS media with statistical or strict monitoring purposes on online part of the Data Acquisition system or offline near the online or totally detached hosts When monitoring is performed in its full online configuration see Figure 4 1 top diagram the monitoring program gets the data from a local monitoring buffer filled from the online data producer the readout process on LDCs and the event builder process on GDCs This approach is the most efficient for what concerns the use of system resources but might impose an unacceptable load on the online host already charged with acquisition and control tasks ONLINE host LDC or GDC local readout OR monitoring buffer monitoring event builder program ONLINE host LDC or GDC OFFLINE host remote readout OR event builder monitoring buffer monitoring program Figure 4 1 The DATE online monitoring local and remote configurations ALICE DATE V3 User s Guide Monitoring in DATE 29 Fig
85. e hosting the event builder Create under DATE_SITE one directory for each CPU in your system The name of each directory must be the name of the corresponding machine e g mppcal01 Give write access for the group and for others to all the directories previously created gt chmod R atw DATE_SITE Create the file rcShmKey onto the site dependent area gt touch DATE_SITE runControl rcShmKey Make sure that rcShmKey has read access for all If not change the access rights gt chmod atr DATE_SITE runControl rcShmKey Internet services Some INETD services are needed by DATE date_recs Run control service used to start the reServer It should be added to all the ALICE DATE V3 User s Guide First time installation 127 machines declared as either LDC or GDC e date_evb Event builder service used to start the gdcServer It should be added to the machines declared as GDC e date_info Info logger d mon used to start the infoDaemon It should be added to the machine declared as DATE_INFOLOGGER_LOGHOST in DATE_SITE_CONFIG infoLogger config e date_evbd Event builder control service used to start the ebDaemon It should be added to the machine declared as GDC e date_mon Monitoring service used for hosts who wish to offer monitoring stream via network It should be added to any machine online or offline who wish to offer monitoring streams to the outside world There is a procedure that
86. e it prior to shipping Messages should also contain printable characters only the library will remove non printable characters prior to shipping Setting the facility name One of the tags used to identify each message exchanged via the infoLogger package is the package name This is a unique identifier that can be specified by the owner of the client process and where some default values are proposed A shell process by default sends messages tagged operator This can be overridden setting the environment variable DATE_FACILITY to the desired value A C program by default sends messages tagged with the package name for DATE packages or with the image filename for non DATE packages This behavior can be overridden by setting the C preprocessor variable DATE_FACILITY to the desired value This must be done before including the infLogger h file e g Setting the Facility name in C programs 1 define DATE_FACILITY myFacility 2 include infoLogger h Java programs are tagged by default as javaCode This behavior can be changed by calling the setFacility method of the LogChannel class Please note that the ALICE DATE V3 User s Guide Injection of messages 93 8 6 3 C Synopsis Java Synopsis Description C Synopsis setFacility method is static one call and all instances of the LogChannel class will use the same facility identifier The infoLogger callable interface This section d
87. e private directory rename it GNUmakefile and customize the makefile target detectors to compile the newly created equipmentList c file 4 Invoke the command gt dateSetup date readout packageParams 5 Invoke the make procedure for the target detectors gt gmake detectors 6 Copy the readout executable file into the DATE_SITE hostName directory for each of the hosts using this specific readout program The detector configuration may be changed between runs just by modifying the detectors config file in DATE_SITE_CONFIG an interactive tool to do that is provided See Chapter 6 3 Changes are taken into account at the next start of run The modifications may concern only the detectors description part of the configuration file below the gt DETECTORS heading and not the other declarative parts which are linked by convention to the handling software The following items can be changed 1 Detectors can be added or removed 2 Equipments assigned to a detector can be added or removed ALICE DATE V3 User s Guide The readout control 61 3 Equipments assigned to a detector can be activated or de activated 4 The value of the parameters provided to both detectors and equipments can be changed Modifications to any other declaration in the configuration file require changing the equipment List software and building again the readout program 6 3 The readout contol An interactive tool is provided to manipula
88. e window are determined by the declarations in the configuration file DATE_SITE_CONFIG runControl config see Chapter 11 5 1 Remark throughout this chapter the term configuration file refers to the file DATE_SITE_CONFIG runControl config Still there are on the window various panels with a well defined functionality e The menu bar is used to perform operations on the run control program itself e The machine selection panel shows all the available machines The ones participating in the data acquisition can be selected here their status will be displayed e The run parameters panel contains the most important parameters to be set by the operator These parameters will be used at start of run e The run status panel shows the run number the run status and a line providing some basic information about the run operation e The run conditions panel allows the operator to select the machine hosting the trigger system and to enable disable the event recording From this panel it is ALICE DATE V3 User s Guide The control console also possible to automatically start one run after the other The run control panel contain all the buttons used by the operator to control the run machine selection menu bar P aaan mn A parameters run status run conditions run control Figure 2 1 The main run control window Initially the only active button is the one labelled Connect Beforehand the ope
89. eader runNb DAQCONTROL gt runNumber vent gt eventHeader nbInRun NOT_SET_TAG vent gt eventHeader burstNb 0 vent gt eventHeader nbInBurst 0 for mIndex 0 mIndex lt MASK _ LENGTH mIndex vent gt eventHeader detectorId mIndex 0 vent gt eventHeader fileSeqNb 0 vent gt eventHeader errorCode 0 vent gt eventHeader deadTime 0 vent gt eventHeader deadTimeusec 0 let the user fill the header and the record by calling the user routine raw data of the event vlen ReadEvent amp event gt eventHeader if LOGLEVEL gt 20 sprintf lineLog LOG_TO fileLog LOG_INFO Read event of lenght d amp event gt rawData 0 evlen lineLog if LOGLEVEL gt 10 amp amp NOT_SET_TAG sprintf lineLog event gt eventH ader nbInRun ReadEvent has not set the event nbInRun lda LOG_TO fileLog number LOG_ERROR increment trigger number only for da vent gt eventHeader nbInRun lineLog ta records not sob eob if event gt eventHeader type PHYSICS EVENT DAQCONTROL gt triggerCount 1 vent gt eventHeader triggerNb DAQCO fill the control shared section vari the status display TROL gt triggerCount ables used to update DAQCONTROL gt eventsInBurstCount
90. eclarations in the configuration file Coreen ian Keyes Ho recent TDN ee asri ine Er uaa me ar ahy EE Mes Wri For ORTEGA phasi 11 Figure 2 10 The configuration parameters window All the parameters can be modified at any time with no harm for the data acquisition since the values will be used exclusively at start of run 2 Status display This option recalls the status display window see Figure 2 3 The status display shows a list of parameters for each machine involved in the data acquisition The lists are established according to the declarations in the configuration file The values are updated at regular intervals according to the settings of the Options menu 3 Buffer status This menu option is not normally present It is a feature added by the buffer status plug in The buffer status plug in may be installed either at the program startup by adding the following fragment of code to the configuration file ALICE DATE V3 User s Guide The infoBrowser console 15 Listing 2 1 Figure 2 11 Installation of the buffer status plug in 1 TCLEVAL 2 source date runControl bufferStatus tcl or at any time by typing the same source command into the Tcl command window and pushing the Eval button A new window will then be shown Figure 2 11 which indicates the occupation level of the circular buffers used in the data acquisition This window is regularly updated at the same time as t
91. ed at end of run such as for example the switching off the high voltages the saving of error statistics that may have been collected in the readList routines The routine does not return any value It should use the variables readList_errorand readList_errorSource to signal error conditions which will provoke the log of a message Report message to the infoLogger Both the readout and the recorder processes use the infoLogger package facilities to report and trace error or abnormal conditions and to trace state changes The ALICE DATE V2 User s Guide 52 Guide to prepare a readout program Table 5 1 readout process also updates the bookkeeping information at the end of the run through the LOGBOOK facility The user can tailor these features to the required needs by setting the value of the variable LOGLEVEL in the Run Control panel according the Table 5 1 LOGLEVEL definitions for the readout and the recorder processes LOGLEVEL Meaning 0 Switch off both infoLogger and bookkeeping facilities lt 10 Switch off the use of the infoLogger facility gt 10 Only report major traces and errors gt 20 Add the report of minor state changes gt 30 Add the report of debugging messages The information reported by the readout package in the run bookkeeping at the end of each run is shown in the Listing 5 4 Listing 5 4 Bookkeeping information reported by the readout package
92. eeping record Example of single line C bookkeeping record char line 1000 sprintf line Run d runNb runNb run number LOGBOOK line LOGBOOK First line LOGBOOK Second line LOGBOOK Third and last line LOGBOOK_MARKER YUYHNUBWNE Example of single line Java bookkeeping record logChannel logBook Run runNb runNb run number logChannel logBook First line logChannel logBook Second line logChannel logBook Third and last line logChannel logBookMarker OP WN ER Example of single line shell bookkeeping record logBook Run runNb runNb run number logBook First line logBook Second line logBook Third and last line logBookMarker opt WN Fe VVVVV In the above examples all logBook calls are consecutive and belong to the same block this is not a requirement Between each logBook call it is possible to add as many foreign instructions as needed It is not necessary to group all calls in consecutive statements However the statsCollector will timeout after a fixed delay this means that when too much time elapses between logBook calls the default timer is 30 seconds the stat sCollector will flush whatever has been received so far to the final log record Therefore it is better to leave as little time as possible between the first and the last Log
93. el lt LOG_NORMAL_TH No message will be logged 2 LOG_NORMAL_TH lt logLevel lt LOG_DETAILED_TH Normal error and information messages will be generated It should allow for normal operation of the data acquisition with the necessary indications in case of troubles The message generation should have no impact on the performance 3 LOG_DETAILED_TH lt logLevel lt LOG_DEBUG_TH Detailed information messages will be logged A performance degradation should be expected 4 LOG_DEBUG_TH lt logLevel Debugging information messages will be logged A severe performance degradation should be expected 2 The severity is an attribute of the messages that makes it possible to make selections on the seriousness of the problem reported 3 The facility is an attribute of the messages that makes it possible to select sets of messages across streams ALICE DATE V3 User s Guide Logging information 121 Each program may have its own internal convention within these ranges Other programs not accessing rcShm should use another mechanism to pass the logLevel value but still use the same log level convention ALICE DATE V3 User s Guide 122 Conventions and file organization ALICE DATE V3 User s Guide DATE Insta lla tion guide This chapter describes how to retrieve the DATE software from the reachieve and how to install DATE on the target systems 11 1 Hardware and software platforms 124 11 2 Getti
94. ent Since this routine may readout other interfaces besides the CORBO we Shall refer here to a routine called ReadEvent Corbo which may be called by ReadEvent An example of such routine is shown in Listing 7 11 It reads the trigger counter from channel 1 and the event counter from channel 2 The way these counters are used is not shown in this code fragment The trigger counter must be equal to the value of the variable triggerNb in the sub event header The latter is set by the software the former by the hardware A consistency check may be performed and the suitable diagnostics generated The event counter should be copied into the variable nbInRun in the sub event header As explained above see 7 4 1 this is necessary to be able to build the complete event ALICE DATE V3 User s Guide Using the CORBO to control the trigger 81 Other counters may also be read from the CORBO such as the dead time counter Reading the CORBO counters 1 ReadEventCorbo 23 int triggerNb nbInRun 3 s 4 Read local trigger counter from channel 1 twice 16 bits 5 regPtrShort unsigned short corboBaseAddr CORBO_CNT1 6 nbInRun regPtrShort lt lt 16 TA regPtrShort Bs nbInRun regPtrShort 9 10 Read all event counter from channel 2 twice 16 bits a Das e regPtrShort unsigned short corboBaseAddr CORBO_CNT2 7 T2s triggerNb regPtrShort lt lt 16 IESE regPtrShort
95. entBuilder process The full path of the directory and filename has to be specified as recording device of the GDC in the Run Control panel for example tmp my_raw_data dat The output filename may include any character except a The directory in which the filename resides should have write access for nobody one can either give write access for everybody or change the owner of the directory The event Builder process uses the infoLogger facility to report and trace error or abnormal conditions and to trace state changes Data fies The data are recorded in unix data files It is possible to limit the total amount of information to be recorded in a run by setting the parameter maxBytes in kBytes in the run control configuration parameters panel A 0 value of the maxBytes parameter means that there is no limit If maxBytes is different from 0 when this ALICE DATE V3 User s Guide The data format 21 3 3 limit is reached the program recording the data recorder in the LDC or the event Builder in the GDC will request the run control to stop the run The data of a run may be recorded to one or several disk files The maximum size of each file is fixed by the parameter maxFileSize in kBytes It allows to limit the file size independently of the run duration A 0 value of maxFileSize parameter means that there is no limit and that all the data will be recorded in a unique file If maxFileSize is diffe
96. eping information reported by the tasdout package SOR commands script SOR commands SOR files The equipment h file Example of equipmentList c library for the sonneur Hier in iisbne 6 367 Example of detector config configuration file Definition of the offset addresses of a device Example of an application accessing a VME device CORBO registers Bits and masks for the CORBO registers Mapping the virtual memory window to the CORBO VME address space CORBO channel 1 initialization CORBO channel 2 initialization Set flag for EventArrivedCorbo Releasing the virtual memory space The routine EventArrivedCorbo Reading the CORBO counters CORBO channel 4 initialization Clear CORBO busy Set CORBO busy Setting the Facility name in E programs Example of single line C bookkeeping record Example of single line Java bookkeeping record Example of single line shell bookkeeping record Example of multi line C bookkeeping record Example of multi line Java bookkeeping record Example of multi line Bourne shell bookkeeping record INETD service configuration INETD services Example of the file runControl contig Repository of the run number Repository of the host of the information logger d mon ALICE DATE V3 User s Guide 15 122 24 31 32 48 50 51 52 753 753 54 63 69 72 73 75 76 77 78 79 79 79 80 81 82 82 82 92 109 109 109 110 110 110
97. ere det Name is the name of the LDC as defined in DATE_HOSTNAME on the LDC and Label is an optional string describing the detector On the following line the detector parameters values may be specified separated by spaces according to what is declared in the detector parameters section Then the detector equipment list follows Each equipment is declared in a line with the following format S EqTypeName Label where S is the equipment status for active equipments for inactive equipments EqTypeName is the type of this equipment one of those declared in the equipment types or trigger types section and Labe1 is an optional string describing the equipment Following each equipment declaration there may be a line with the specification of the equipment parameters values separated by spaces according to what declared in the equipment type section ALICE DATE V3 User s Guide VME access and tngger system This chapter gives some indications on how to set up the trigger system It shows what electronic equipment is required and how the software should handle it An example based on VME equipment is discussed down to the details The introductory section explains how to access the VME bus from a program using a direct mapping mode 7 1 Access to the VME bus 442 6 2 eb oa hee oe oe de 72 7 2 ANCAP SESSION s x sang ie Ce a ee a BO 74 7 3 The CORBO module aa aaa aaa a 75 7 4 Triggering with the CORBO lt
98. escribes the macros and the methods available for programs written in C or Java and running on any DATE host Compilation of C programs require the inclusion of the file infoLogger h part of the infoLogger distribution kit pointed by the environment variable DATE_INFOLOGGER_DIR The proper include path is automatically set if a standard DATE makefile is used Linking of C programs require the libInfo a library available on DATE_INFOLOGGER_DIR DATE_SYS directory this is done automatically if a standard DATE makefile is used to link the executable image Some system libraries are also needed for run time support the environment variable DATE_SYSLOADLIBES defined by the standard DATE shell setup can be used to resolve all symbols needed by the infoLogger library Compilation and execution of Java programs requires the inclusion of the directory S DATE_INFOLOGGER_DIR in the CLASSPATH environment variable LOG log include infoLogger h void LOG char severity char message severitycan be one of LOG_INFO LOG_ERROR LOG FATAL void log char severity String message j severitycan be one of DB severityInfo DB severityError DB severityFatal The given message is sent to the infoLogger host runLog stream with the given severity LOG_TO logTo include infoLogger h void LOG_TO char stream char severity char message severitycan be one of LOG_INFO ALICE DATE V3 User s Guide
99. escription of the readout equipment types 2 gt TRTYPES followed by the description of the trigger equipment types 3 gt DEPARAMS followed by the description of the list of parameters passed to all the detectors The last section contains the list of equipments associated to each detector and the actual literal values for the parameters 4 gt DETECTORS followed by the detectors description An example of the configuration file is shown in Listing 6 3 The readout equipment types The types of readout equipments see 6 5 2 available are declared under the heading gt EQTYPES These declarations refer to types of equipments for which the driving software is provided in the equipmentList c library The actual instantiation of the equipments is declared under the various detector declarations The description of a readout equipment type begins with a line of the format gt EqTypeName Id NODATA where gt is a literal character EqTypeName is the type name Id the type identifier a short integer that will be copied in the equipment header if this type produces data and NODATA an optional keyword specifying that the equipments of this type do not produce any data when they are read out If NODATA is omitted the equipments of this type are expected to produce data If NODATA is specified there will be no trace of these equipments in the events even the equipment headers will be omitted The type declaration may be followed by t
100. ey indicate a file name At start of run these names are communicated to the respective machines before that they are parsed and the following substitutions are applied 1 The first occurrence of the character is replaced by the machine name 2 The first occurrence of the character is replaced by the run number This feature allows us to generate data files with different names for each machine and each run For example if ehRecordingDevice is declared to be Data raw the data of the run 321 will be stored on the file Data 321 raw the next run will be on Data 322 raw and so on The actual communicated file name may be forced to be dev nul11 if the recording is disabled on the main run control window ALICE DATE V3 User s Guide Run control configuration 139 a If the name of the recording device ends by it is supposed to indicate a remote machine instead of a file therefore the substitutions do not occur e monitorEnableF lag See description in Table 11 3 e logLevel It controls the generation of messages by all the date processes running on a DAQ machine Standard value is 10 All the possible values are the following 0 No message is generated 10 Normal error and information messages No impact on the performance 20 Detailed information messages Small performance degradation 30 Debugging information messages Big performance degradation fanmo E md
101. f available username if applicable severity Information Error or Fatal and the text As the browser is written in Java it can run on any host were the standard DATE set is installed The display runs via a standard X11 link and therefore it is vital to have the X11 display setup and running At this purpose it is necessary to setup the DISPLAY shell variable either via the shell or opening a new X terminal and authorize the client on the server prior to run the infoBrowser The infoBrowser operator window The infoBrowser may be run in a special mode intended for the operator of the data acquisition by the command infoBrowser operator amp ALICE DATE V3 User s Guide The statsBrowser console 17 The infoBrowser will then show a different window Figure 2 13 which will not allow the interactive selection of the message streams Only the runLog stream will be shown which by convention contains only the messages concerning the general aspects of the operations and provides a rough indications on the behaviour of the system In case of troubles it is of course necessary to obtain more detailed information from the various packages therefore the standard infoBrowser window should be used It is a good practice to keep the infoBrowser operator window permanently open on the operator console yD py psa ey ral Figure 2 13 The infoBrowser operator window 2 3 the statsBrowser console Using the infoLogger scheme
102. f full event data read out by one stand alone LDC data read out by several LDCs and assembled by one GDC In the first case the detector identification mask is set to 0 In the second case the detector identification mask in the header is used to distinguish between an event and a subevent and to identify the subevents of the same event see Table 3 3 Table 3 3 Usage of the detector identification mask header field Event generated by Event generated by aLDC aGDC Subevent Type of event Event with no sub Event with atleast Subevent event one subevent Detector mask 0 1 0 bit 95 Detector mask 0 Logical OR of the Detector mask of bits 0 to 93 detector masks of the corresponding the subevents detector In a full event there will be one header for the event itself and one header for each subevent The size field in the header will be used as indicated in Table 3 4 Table 3 4 Usage of the header len and of the size Event generated by Event generated by a LDC a GDC Seren Header len Displacement to Displacement to Displacement to the event data first subevent the subevent data Size Event size Full event length Subevent size ALICE DATE V3 User s Guide 26 Data recording and data format Figure 3 1 Subevent from LDC A Subevent A length ype PHYSICS_EVEN Header length detectorld 0000 0000 0001 Subdetector A data Subevent from LDC B Subevent B length ype PHYSICS_E
103. g and analysis platform How to build and install a readout program The timerRand readList After the installation of the distribution kit the file readList_timerRand c sits in the date readList directory To build the readout program using this readList the following procedure has to be followed 1 Execute gt dateSetup date readout packageParams 2 Copy date readList readList_timerRand c ina private directory 3 Copy date readList GNUmakefile_readout in the same directory and rename it GNUmakefile 4 Execute the makefile procedure giving timerRand as target gt gmake timerRand This will produce a new executable program called readout 5 Copy the newly built readout program on the directory pointed to by S DATE_SITE DATE_HOSTNAME The custom readList If you wish to define your own readList implementing a library with the functions described in the above templates and build your readout program the following procedure has to be followed 1 Execute gt dateSetup date readout packageParams 2 Create a readList c file providing containing the four routines described in the templates ALICE DATE V2 User s Guide How to build and install a readout program 55 3 Copy date readList GNUmakefile_readout in the same directory and rename it GNUmakefile 4 Edit the GNUmakefile to add your own target and the dependencies for it private libraries include files etc 5 Execute the ma
104. g structure Utility scripts such as ALICE DATE V3 User s Guide 116 Conventions and file organization GNUmakefile packageParams and testParams configuration files use by dateSetup e All the source files e One directory per operating system named upon the operating system name as returned by which contains the run time system for all the hosts running that type of system Files in this directory are of type obj lib bin sh and tcl1 Currently supported systems are AIX SunOS 10 2 3 Structure of DATE SITE This area contains the following items e The directory configurationFiles which contains all the configuration files of an experiment runControl config runNumber config runControl params SOR commands SOR files infoLogger config e The directory logFiles which contains a general log file called runLog package specific log files e One directory per package named upon the package name which contains all the private configuration files of a package e One directory per host named upon the host name which contains any host specific software such as readout enableTrigger disableTrigger 10 3 Environment vanables and aliases Any interactive session or shell using DATE must define an environment variable DATE_SITE pointing to the site specific directory and then invoke date setup sh As a result of this operation the following environment variables are defined DATE_
105. he declaration of the equipment specific parameters These parameters can be accessed from all the equipment handling routines of a detector The list of parameters is common to all the equipments of a given type The values are assigned equipment by equipment ALICE DATE V3 User s Guide The detectors configuration file 69 The parameters are declared one per line by specifying for each of them the name and the format either hd 1d shx 31x s or c according to the C language definition of these constants Listing 6 3 Example of detector config configuration file DAINAUTAPWNRFPDOOWMATNAAIHWNE WWWWWWWNHNNNNNNNDND NO DOPWNHRFDOWOWMDAAAABWNHER OW WW co N gt EQTYPES gt VmeWindow 0 NODATA vmeWinOffset 1x vmeWinSize 1x gt Rand 1 EvMinSize ld EvMaxSize ld Eqid shd gt TRTYPES gt Corbo 0 NODATA vmeBaseAddress 1x gt Timer 0 NODATA Evinterval ld gt DEPARAMS name s gt DETECTORS gt mppcal03 LDC1 LDC1 VmeWindow 0x700000 0x60 Corbo CorboLDCl 0x700000 Timer TimerLDCl 1000000 t Rand RandLDCl 10 20 11 t Rand RandLDCl 1000 1010 22 gt mppcal04 LDC2 LDC2 VmeWindow Ox OxF00000 Corbo CorboLDC2 0x700000 Timer TimerLDC2 5000000 Rand RandLDC2 20 30 11 6 6 2 The trigger equipment types The types of trigger equipments see 6 5 1 available are declared under the heading gt TRTYPES These decl
106. he front end software 53 Wherever a run is started the start of run scripts are executed before calling ArmHw Similarly when a run is stopped the end of run scripts are performed before calling DisArmHw Examples of start of run scripts may be the programming of high voltages and the calibration of ADCs and TDCs If there is the need to record the results of these programs on the output medium one can simply let these programs writing their results on the start of run files There is also the possibility to call the infoLogger facility from a shell script as shown in Listing 5 5 It is recommended to call the error macro of the infoLogger facility from the user scripts to signal and report eventual errors in order to get a trace in the runLog stream of the operator InfoBrowser window Listing 5 5 SOR commands script i bin sh 2 3 date setup sh 4 log SOR commands DATE_SITE SDATE_SITE DATE_HOSTNAME SDATE_HOSTNAME pwd gt pwd W 5 4 2 The Start of Run and End of Run files A facility is provided to describe in special files the list of filenames to be copied to the output medium as special records at the beginning and at the end of each run These files may have been created by the above Unix shell scripts and may again be common to all front end processors or specific to each front end processor These special files are called respectively SOR
107. he machine hosting the software to enable disable the trigger c Enable disable autostarting the runs one after the other 2 The run parameters panel of the main run control window Figure 2 1 Frequently changed run parameters such as the number of events to collect Other parameters may be presented here as specified in the configuration file These parameters will be set at start of run into all the machines involved in the data acquisition 3 The configuration parameters window See Figure 2 10 In this window there should be parameters that the operator is not supposed to change since they concern rather the system configuration than the run configuration a Common parameters these parameters will be set at start of run into all the machines involved in the data acquisition b LDC GDC specific lists of parameters are presented for all the available machines Only the machines involved in the data acquisition will receive their own parameters at start of run ALICE DATE V3 User s Guide The control console 11 2 1 2 1 Figure 2 4 2 1 2 2 The run operation is controlled through the buttons of the run control panel Figure 2 1 1 Start run It initiates the procedure to start the run on all the machines involved to synchronize them and to eventually enable the trigger 2 Stop run It initiates the procedure to disable the trigger and then stop the run on all the machines involved 3 Pause trigge
108. he other status information The buffer status window 2 2 the infoBrowser console All the components a DATE system generate all sorts of messages diagnostics debugging statistics logging These messages can be browsed online filtered and selected using the infoBrowser an X11 tool written in Java The main features of the infoBrowser are e small run time overhead e memory based browsing and filtering no use of disk resources e browsing messages stored in a standard online area DATE_SITE_LOGS as well as in archives areas e display customizable in font size and color e allows printing of the messages or regions of interest e online context sensitive and generic help e selection of messages according to the source facility the messages severity and the repository logFile e display of messages starting from a given time of day To start the infoBrowser console the first step is to run the DATE setup procedure Once the X11 display is set and enabled it is possible to run the command infoBrowser from the shell level There are various run time flags available for special purposed use the help flag for a complete description In its default configuration the infoBrowser will allow normal browsing of the online logs area An example of how the infoBrowser window looks like is given in Figure 2 12 Please note that the appearance might change between different versions of the tool ALICE DATE V3
109. he standard DATE kit has been installed and properly configured To start the statsBrowser the following procedure should be followed 1 set the DISPLAY environment variable to the appropriate value gt DISPLAY displayHostName displayNum export DISPLAY or whatever is the syntax of your preferred shell the above example works for bin sh Alternatively one might do gt xterm display displayHostName displayNum amp This will create a brand new X11 terminal emulator where the environment is properly set The string displayHostName displayNum must correspond to a valid X11 display identifier e g ion01 cern ch 0 0 2 if not done yet run the DATE shell setup procedure 3 start the statsBrowser from the shell prompt gt statsBrowser amp For a complete list of the run time flags issue the command gt statsBrowser help An online help is available via the stat sBrowser Help online menu The help pages are written in HTML and can be browsed via Netscape see section 9 4 for more details The stat sBrowser documentation can also be accessed offline using any HTML browser at the protected URL ALICE DATE V3 User s Guide 106 Guide to use the bookkeeping system 9 4 9 5 9 5 1 http aldwww cern ch documents DATE DATE infoLogger statsBro wser docs html Browsing HIML help pages The statsBrowser tries to reuse any Netscape window already opened on the X11 server and if none are avail
110. hell the above example works for bin sh Alternatively one might do gt xterm display displayHostName displayNum amp This will create a brand new X11 terminal emulator where the environment is properly set The string displayHostName displayNum must correspond to a valid X11 display identifier e g ion01 cern ch 0 0 2 if not done yet run the DATE shell setup procedure 3 start the infoBrowser from the shell prompt gt infoBrowser The infoBrowser tool allows a set of run time flags to enable special functions For a complete list of these flags and to get their description issue the command gt infoBrowser help Once the tool is started an online help is available via Help menu The help pages are written in HTML and can be browsed online using Netscape see section 8 3 for more details The infoBrowser documentation can also be accessed offline using any HTML browser at the protected URL http aldwww cern ch documents DATE DATE infoLogger infoBrow ser docs html ALICE DATE V3 User s Guide Browsing HTML help pages 87 8 2 1 The infoBrowser operator mode The infoBrowser can run in operator mode Main features of this mode are e browsing of the standard information stream dedicated to run time operators e reduced functionality for simpler and faster usage The operator mode can be selected when the infoBrowser is started via the operator switch e g gt infoBrowser operator 8
111. here may be only one such area in a given file system This area is totally overwritten when installing a new version of the distribution kit therefore no site specific files may reside here At run time this area can be considered as read only The second area contains all the site specific files such as the configuration files and all the files created at run time by the DATE packages It may reside in any directory decided by the site manager and is pointed to by the environment variable DATE_SITE This area is left untouched by the installation procedure so that all the site specific files be preserved There may several site specific areas on the same file system one per experiment ALICE DATE V3 User s Guide File organization 115 Figure 10 1 A view of the file organization in DATE 10 2 1 Structure of DATE ROOT This area contains the following items Utility scripts These are scripts used for the environment management such as Install setup sh and setup csh README The hidden directory commonScripts which contains a library of scripts The directory commomDefs which contains all the common h files The directory java The directory makefiles The directory ReleaseNotes One directory per DATE package named upon the package name dataRecorder eventBuilder infoLogger monitoring readList readout runControl simpleFifo 10 2 2 Package directory Each package directory has the followin
112. ient process who generates the message and the daemon running on the infoLogger host This operation can run concurrently with other requests coming from several processes This means that no guarantee is given on the order of delivery of messages coming simultaneously from several DATE hosts and processes The only facts that can be given for certain are 1 all log requests will block the issuer until the request has been accepted by the TCP IP communication library maybe before the request is actually sent see the next point 2 several messages coming from the same or from different infoLogger client processes might be packed by the TCP IP communication library on the client host before the actual delivery takes place 3 no feedback is returned by the daemon running on the infoLogger host therefore it is not possible to guarantee the proper delivery of a given message 4 consecutive messages coming from the same process will be delivered and recorded in the same order as they have been issued 5 multi line messages containing embedded carriage returns will be delivered in one single exchange and will be recorded with one atomic operation thus excluding all interleaving with other process All the messages exchanged via the infoLogger library must be native strings and can contain embedded carriage returns to create multi line messages However they should not be terminated with a carriage return the library will remov
113. ilable and the channel set in noWait mode the pointer returned will be NULL in this case the event does not need to be disposed Zero in case of success also if no event is available otherwise a error code see 4 2 3 for more details MONITOR_FREE_EVENT a INT POINTER SUBROUTINE MONITOR_FREE The given event obtained via MONITOR_GET_EVENT_DYNAMIC is released and all the resources associated to the event are returned to the system monitorFlushEvents include monitor h int monitorFlushEvents void SUBROUTINE MONITOR_FLUSH_EVENTS All the data available in the monitoring buffer is discarded The next event transferred over the monitoring channel will be injected in the monitoring stream after this call terminates Zero in case of success otherwise a error code see 4 2 3 for more details monitorSetWait include monitor h int monitorSetWait void INTEG ER MONITOR_SET_WAIT ALICE DATE V3 User s Guide 38 Guide to write a monitoring program Description Returns C Synopsis FORTRAN Synopsis Description Returns C Synopsis FORTRAN Synopsis Description Returns After this call completes if the monitoring program requests an event when the monitoring buffer and the monitoring channel are empty the monitoring program will stop and wait for
114. ime during which the recorder process sleeps can be controlled by the variable DAQCONTROL gt recorderSleepTime allowing for more precise tuning depending on the burst length Its value expressed in microseconds can be set in the Run Control parameters The default is 10 milliseconds each event is sent separately over the network while in the event loop the recorder process continuously checks for the arrival of the end of run command It exits the event loop if one of the following conditions is met the maximum number of bytes to be written has already been reached in case of recording of a local file or there have been too many errors in writing the file or in the transfer over the network or if somebody asked to stop the run e close the file or the socket connection after exiting the event loop 5 2 The readList templates The template for the sources of the experiment specific code can be found in the directory date readList readList_timerRand c The template can be used as a working example The readout program distributed in the DATE distribution kit uses this template The readList_timerRand generates dummy events of random size between DAQCONTROL gt randEventMinSize and DAQCONTROL gt randEventMaxSize at a simulated trigger interval of DAQCONTROL gt randEvent Interval microseconds All these variables can be set in the Run Control panel The data produced are monotonously increasing and start at 0 The user may
115. implements all these services It must be invoked from root Make sure that the DATE symbols are still defined otherwise execute again the procedure described in 11 3 1 point 8 The command to invoke the procedure is the following gt dateNetwork c i g GDC1 GDC2 1 LDC1 LDC2 J G InfoLoggerl InfoLogger2 m MON1 MON2 where c performs the check that the network is properly setup to run DATE no modifications performed g indicates the list of GDCs G indicates the list of machines where the infoLogger may run i installs the setup needed by DATE 1 indicates the list of LDCs m indicates the list of machines from which event monitoring is allowed The list of machines must be a blank separated sequence of hostnames e g gt dateNetwork c g hostl1 host2 host3 1 host4 host5 Ghost6 m host7 host8 host9 All hosts shall be reachable directly by rsh and the caller must be authorized to remote shell as user root on the remote machines This is normally the case if you have followed the instructions in 11 3 1 The command dateNetwork performs the following steps 1 The file etc inetd conf of each machine where the INETD services are required will contain the code fragment in listing 11 1 ALICE DATE V3 User s Guide 128 DATE installation guide Listing 11 1 INETD service configuration BwWN Ee DATE service handlers date_r
116. in by readout to dow Common reserve space in the parameters under event buffer If this gt CNFPAR in run number is too large Control config space is waisted at the end of the buffer If it is too small the ReadE vent routine in readList may overwrite and cor rupt valuable data or code triggerCount Incremented by Readout Status display win readout at each dow under gt LDC physics event not STATUS in for the other types runControl config of events after call ing ReadEvent and then stored in eventHeader trig gerNb eventCount Incremented by Readout and event Status display win readout foralltypes builder dow under gt LDC of events before STATUS and calling ReadEvent gt GDCSTATUS in It is compared to runControl config maxEvents to stop the run Incre mented by the event builder as well burstCount Set by readout at Readout Status display win each event after dow under gt LDC calling ReadEvent STATUS in to the value found runControl config in event Header burstNb 134 DATE installation guide Table 11 3 Run control parameters Parameter name Description Set by To appear in window eventsInBurst Set by readout at Readout Status display win Count each event after dow under gt LDC calling ReadEvent STATUS in to the value found runControl config in event Header nbInBurst burstFlag Not used by DATE Either the operator It depends on the It may be used to or
117. ion based on the F selects all messages messages severity with severity level Fatal I Info E Error F Fatal username Selection based on the nobody selects all mes originator username sages generated from automatic daemons text Selection based on the Error 23 selects all messages text messages starting with the string Error 23 All sorts of egrep regular expressions can be used with the above field Some examples are given in Table 8 2 Table 8 2 Some examples of regular expressions between 10 00 and 10 59 Messages generated October 14 between 20 00 and 22 59 Messages generated October 14 between 06 00 and 12 59 Messages coming from the package monitor Messages coming from monitor or readout Error messages Error and Fatal messages What to select How Messages generated on October 12 1012 Messages generated between October 12 101 2 4 and October 14 Messages generated October 14 101410 10142 0 2 1014 0 6 9 1 0 2 monitors monitor readout E EIF Multiple selection patterns can be specified in the regular expression e g expressionl amp expression2 expression3 will select messages matching expressionlandexpression2 or expressions Once the selection has been preformed it is possible to browse them via the command gt infoBrowser dir directory ALICE DATE V3 User s Guide Injection of messages
118. isable event interrupt 7 regPtrChar unsigned char corboBaseAddr CORBO_BIM1 OFST_VR3 8s regPtrChar 0 Clear event vector iS regPtrChar unsigned char corboBaseAddr CORBO_BIM2 OFST_CR3 10 regPtrChar 0 Disable time out interrupt Eiz regPtrChar unsigned char corboBaseAddr CORBO_BIM2 OFST_VR3 T2 regPtrChar 0 Clear time out vector 13 regPtrShort unsigned short corboBaseAddr CORBO_CNT4 14 regPtrShort 0 Clear event counter LSs regPtrShort unsigned short corboBaseAddr CORBO_TOU4 F 16 regPtrShort 0 Clear dead time counter Listing 7 13 Clear CORBO busy 1 regPtrShort 2 regPtrShort unsigned short corboBaseAddr CORBO_CLEAR4 FULLBYTE Clear busy The program DisableTrigger does the same but instead of clearing the busy signal it sets it as shown in Listing 7 14 Listing 7 14 Set CORBO busy 1 regPtrShort unsigned short corboBaseAddr CORBO_TEST4 2 regPtrShort 0 Prepare to trigger 3 regPtrShort FULLBYTE and then do it set busy ALICE DATE V3 User s Guide Guide to use the infoLogger system A typical distributed data acquisition system is made of several processes running on one or more CPUs often belonging to incompatible domains The development and operation phases of such a system need a set of primitive
119. it integer with the length in bytes of an optional extension of the equipment header default 0 Equipment identifier a 16 bit integer number identifying the equipment declared in the configuration file The identifier is a number which uniquely identifies an equipment within a detector It is a convention of the experiment Byte alignment an 8 bit integer specifying the length of the word read from the hardware in bytes Reserved 8 bit word Raw data length a 32 bit integer with the length in bytes of the data block header excluded headerExtLen type rawByteAlign equipmentid 31 32 bit 0 Figure 6 5 The equipment header The equipment header is defined in the file DATE_COMMON_DEF S equipment h see Listing 6 1 ALICE DATE V3 User s Guide The equipmentList library 63 Listing 6 1 The equipment h file 1 struct equipmentHeaderStruct 2 short headerExtLen header extension in bytes 3 short type th quipment type identifier 4 char reserved reserved byte Or char rawByteAlign word length in bytes 6 short equipmentId equipment identifier 7 long rawDataLen data block length in bytes Sr OE7 9 10 struct equipmentStruct ADA lE struct equipmentHeaderStruct equipmentHeader Loss unsigned short rawData 1 TS cE 6 5 The equipmentiist library The functions to handle the equipments declared in the configuratio
120. itoring library is available as S DATE_MONITOR_DIR S DATE_SYS libmonitorstdalone a This library will not reference the RFIO system and can be used for local file I O only Hardware and software platforms not part of the standard DATE distribution but possible clients of the DATE monitoring scheme see Table 11 1 can still use the monitoring library by copying the necessary files and performing a local compilation and link Error codes The entries belonging to the monitoring library may return a monitoring specific error code This code can be either zero for success or non zero for failure To decode an error code please refer to the DATE_MONITOR_DIR monitor h file or call the entry monitorDecodeError described in the next section The monitoring callable library This section describes the entries available in the monitoring library Each entry is described in the C version and if available in the FORTRAN equivalent For the decoding of error codes eventually returned by the entries please refer to 4 2 3 monitorSetDataSource include monitor h int monitorSetDataSource char ALICE DATE V3 User s Guide 34 Guide to write a monitoring program FORTRAN Synopsis Description Table 4 1 Returns C Synopsis FORTRAN Synopsis Description Returns C Synopsis INTEGER MONITOR_SET_DATA_SOURCE CHARACTER The source of events to monito
121. k via an Inetd server called mpD mon The monitoring library provides several other features such as delivery of events stored on disc either local or remote ALICE DATE V3 User s Guide Run control 1 4 Run contol The run control architecture is shown in figure 1 5 The whole DATE system is controlled from a central point which may be either one of the processors involved in the data acquisition or more suitably an independent workstation It is preferable that the run control workstation shares the DATE file system even though this is not strictly necessary The run control is performed by a process called runCont rol which is made aware of the hardware configuration by reading a configuration file This process opens sockets to all the machines involved where Internet daemons are created The d mon called reServer controls the readout and recorder processes Two deemons ebDaemon and reServer control the eventBuilder process These daemons make use of shared memory segments to communicate with the controlled processes The runControl process displays in a dedicated window the status of the machines involved in the data acquisition and the values of selected variables in the shared memory segments Work aiaiai Figure 1 5 Run control architecture 1 5 Information logging Any process in the system may generate messages A system is provided figure 1 6 to collect and display these messages in an orderly way
122. kefile procedure giving the target that you have defined 6 Copy the newly built readout program on the directory pointed to by S DATE_SITE DATE_HOSTNAME 5 5 3 The generic readList For building a readout program using the generic readList you should refer to the Chapter 6 ALICE DATE V2 User s Guide 56 Guide to prepare a readout program ALICE DATE V2 User s Guide The genenc readList This chapter describes how to use the generic readList which allows you to organize the readout as a collection of equipments Equipments can be programmed independently they can be selected activated and de activated without changing the readout code 6 1 6 2 6 3 6 4 6 5 6 6 ALICE DATE V3 User s Guide OOVEIVIEW 6 ok n bee es a ene oe ee ee a i 58 Using the generic readList 4 uu aw exe SORE eR RN 59 The readout control god ee gee ee ee HR 61 The equipment header 44 ebb acd PE oe RAM 62 The equipmentList library 2 2 4 0 4660 000 s ee adn 63 The detectors configuration file 68 58 The generic readList 6 1 Overview Figure 6 1 In Chapter 5 it is described how the readout program accesses the hardware by calling the routines of the readList library which contains all the code specific to a given electronic set up Figure 6 1 Readout and the standard readList Instead of writing the experiment s own readList it is possible to use a generic readList that is provi
123. les 4 2 2 The monitoring package files ALICE DATE V3 User s Guide iii AJIJ RAUNG 11 11 11 13 lt 1 15 16 17 19 20 20 20 20 21 x 23 24 27 28 30 31 32 vi Contents 4 2 3 4 2 4 Error codes The monitoring callable library 4 3 The eventDump utility program Chapter 5 Guide to prepare a readout program 5 1 Overview 5 1 1 5 1 2 5 1 3 The main event loop The readout process The recorder process 5 2 The readList templates 5 3 Report message to the infoLogger 5 4 Customization of the front end software 5 4 1 5 4 2 The Start of Run and End of Run scripts The Start of Run and End of Run files 5 5 How to build and install a readout program 5 5 1 5 5 2 5 5 3 Chapter 6 The generic readList Overview Using the generic readList The readout control The equipment header The equipmentList library 6 1 6 2 6 3 6 4 6 5 6 5 1 6 5 2 6 5 3 6 5 4 6 5 5 6 5 6 6 5 7 6 5 8 The timerRand read List The custom readList The generic readList Trigger equipments Readout equipments Accessing the parameters Arming the equipments Reading the equipments Disarming the equipments Triggering The function references 6 6 The detectors configuration file 6 6 1 6 6 2 6 6 3 6 6 4 Chapter 7 VME access and trigger system Access to the VME bus The trigger system The CORBO module
124. ly Average event size ALICE DATE V3 User s Guide Monitoring configuration 143 Table 11 4 11 7 2 Monitoring configuration parameters Parameter name Used for Description EVENT _BUFFER_SIZE Online monitoring only Size of buffer used to store events data EVENTS_MAX_AGE Online monitoring only Maximum age in sec onds of the events avail able for monitoring MONITORING_HOSTS Online monitoring Comma separated list of Networked monitoring hosts allowed to perform monitor when available from this host MUST_MONITORING_HOSTS Online monitoring Comma separated list of Networked monitoring hosts allowed to perform all types of monitoring from this host For the MONITORING_HOSTS and MUST_MONITORING_HOSTS parameters a comma separated list of hosts should be given e g MONITORING_HOSTS localhost suxy mppcxy05 In the above example the hosts allowed to perform normal monitoring are the local host all hosts whose name begins with suxy suxy01 suxy02 and so on plus the host mppexy05 A host who is defined within the MONI TORING_HOSTS list can only perform monitoring when available To be able to perform 100 monitoring a host must be in the MUST_MONITORING_HOST list If the parameter MUST_MONITORING_HOSTS is not specified all hosts can perform 100 monitoring on the monitored machine Conversely if the parameter MONITORING_HOSTS is not specified all hosts can perform monitoring functions
125. me in milliseconds for the readout of a particular equipment Up to DAQCONTROL gt maxEvent Size bytes are available to fill the event data and the event header this parameter can be changed in the main panel of the Run Control program when setting the maximum event size The main readout program calls this routine Listing 5 2 after the initialization of some fields in the event header Upon return from this routine it checks that the user has filled the mandatory fields in the event header updates some variables used in the run status display and sets the time in the event header The routine must return the number of bytes actually readout It should use the variables readList_error and readList_errorSource to signal error conditions which will provoke the log of a message and the termination of the run An example on the easiest way to keep the count in a varying length event is given in Listing 5 3 ALICE DATE V2 User s Guide 50 Guide to prepare a readout program Listing 5 2 Calling ReadEvent ADNDUTBPWNHRFDWDOWOWAATAAIHRWNE 20 Zs 223 23 24 Zoe 26 21 28 29 30 31 32 33 34 35 36 fill the header of event record vent gt eventHeader magic EVENT_MAGIC_NUMBER vent gt eventHeader type PHYSICS_EVENT vent gt eventHeader headLen sizeof struct eventHeaderStruct vent gt eventH
126. ment header 4 DisarmEquipment ALICE DATE V3 User s Guide The generic readList 6 5 2 6 5 3 It is called by DisarmHwat end of run It should reset the hardware It cannot generate any data If a functionality is not required dummy routines should be provided Readout equipments Readout equipments deal with the readout hardware and collect the equipment information They usually generate data therefore the flag NODATA see 6 6 1 is not specified It may though be convenient to isolate some specific processing in an equipment even though this processing generates no data The equipment will be marked with the flag NODATA in such a way no data relative to this equipment will be added to the event still all the routines will be called The routines to be provided for each equipment are the following replace Equipment with the equipment name 1 ArmEquipment It is called by ArmHw at start of run It should initialize the hardware It cannot generate any data 2 ReadEventEquipment It is called by ReadEvent at each event even if NODATA is specified In the latter case the pointer to the header and to the data are not provided It should perform the event related processing of the equipment hardware and if NODATA is omitted read the equipment data block and update the equipment header 3 DisarmEquipment It is called by DisarmHwat end of run It should reset the hardware It cannot generate any data
127. messages that are processed by the infoLogger system The following conventions apply to such messages Use of streams Any DATE activity may send messages to three information streams its own stream the common runLog stream and the common logBook stream By activity it is meant a set of programs processes co operating for a given purpose The largest collection of programs should be a DATE package the smallest may be a single program A library of routines may also be considered as an independent activity and have its own stream when it has a purpose specific to the data acquisition e g the monitoring library 1 The activity own stream The name of the stream should be the most significant name indicating the related activity It may be the name of the package or the name of the program or any other All the messages of any severity should go to this stream The routine calls to be used to inject messages into this stream are of the type logTo 2 The runLog stream This stream is used by all the programs to address messages to the operator The messages may have any severity The messages on this stream must be thought to be particularly significant to the operator they are usually not the same messages intended to the expert Only a small fraction of all the occurrences generating messages should be reported to the operator as well in this case the designer should consider generating a stripped down message for the operator con
128. mitation can be overcome by doing the following operations 1 Select in the main run control window the option View Tcl eval 2 Inthe Tcl command entry area type the following command set bypassMastership 1 3 Click on the Eval button It is then possible to connect to the machines since the take over option is proposed as well This statement can also be permanently set in the file DATE_SITE_CONFIG runControl config under the gt TCLEVAL keyword Then it would be possible to have several run controls active at the same time This practice is strongly discouraged Information logger configuration The information logger makes use of a deemon running on a host machine The name of the machine must be indicated in the file DATE_SITE_CONFIG infoLogger config This configuration file should contain one line as Repository of the host of the information logger d mon 1 DATE_INFOLOGGER_LOGHOST infologger_host where infologger_host is the name of the machine where you want to run the infoLogger daemon Monitoring configuration Functionally hosts participating toa DATE monitoring scheme can be defined as ALICE DATE V3 User s Guide Monitoring configuration 141 11 7 1 1 monitoring hosts running specificmonitoring programs either written by a specific experience or part of the standard monitoring package e g the utility eventDump 2 monitored hosts offering monitoring streams to monitoring pr
129. monitoring host We will now review the configuration needed on monitored and relayed hosts to let them perform their function Creation of configuration files The monitoring scheme can be configured using three separate files e S DATE_SITE_CONFIG monitoring config this file is optional and can be used to control a complete DATE site all types of hosts e S DATE_SITE DATE_HOSTNAME monitoring config this file is mandatory for online hosts and must be created by the DATE system administrator It is not required for offline relayed or monitoring hosts e etc monitoring config this file is optional and can be used to control the behaviour of relaying hosts it is useless for online or offline monitored hosts The above files should be created using the following commands ALICE DATE V3 User s Guide 142 DATE installation guide Listing 11 6 Creation of configuration files 1 gt touch file 2 gt chmod u rw g rw o r file where file is the full path of the file to be created Once created the configuration files can be edited and parameters can be specified as a list of names followed by their associated values Comments can be inserted via the sign e g This is a comment PARAMETER VALUE comment These files can be changed at any time Some of the parameters those labelled in Table 11 4 as Online monitoring only require the acquisition to be stopped and no active clients
130. n file must be provided as the equipmentList c library An example is shown in Listing 6 2 There exist two types of equipments the trigger equipments and the readout equipments Their handling routines have a different calling sequence in the two cases as described below 6 5 1 Trigger equipments Trigger equipments deal with the trigger hardware Many trigger equipments may be declared in the detectors configuration file but only one per detector should be active If there are more of them no warnings are given but only the first is kept Trigger equipments usually do not generate data therefore the flag NODATA see 6 6 1 is specified Optionally trigger equipments may generate data The routines to be provided for the trigger equipment are the following replace Equipment with the trigger name 1 ArmEquipment It is called by ArmHw at start of run It should initialize the hardware It cannot generate any data 2 EventArrivedEquipment It is called by EventArrived when waiting for the next event It should check the trigger status and return 1 if a trigger has arrived 0 otherwise It cannot generate any data 3 ReadEventEquipment It is called by ReadEvent at each event even if NODATA is specified In the latter case the pointer to the header and to the data are not provided It should perform the event related processing of the trigger hardware and if NODATA is omitted read the equipment data block and update the equip
131. n interrupt handling but it waists CPU cycles this is not a problem when the processors are dedicated to the readout The trigger electronics should implement the basic scheme given in Figure 7 2 LOC TRGGER MODULE TRIGKER M TIGGER C Pa p THIER ERABLE C 1 k FAH OUT MRa MODLLE i Emr LOC Teas at MODLLE THI EGER Fi WaT oT BAME DETECTOR Figure 7 2 Trigger electronics This circuit meets the following requirements 1 All the LDCs handle a software controlled trigger module The trigger module is essentially a set reset flip flop indicating whether the LDC is enabled to accept triggers status ready or is busy processing an event status busy The status of the module is accessible as an output signal and can be interrogated by software as well When a trigger arrives the module must flip its status from ready to busy within nanoseconds In the time window between the trigger arrival and all the output signals being flipped the trigger must be hold off by other electronic means The trigger module status is reset to ready by software 2 The trigger may be gated off by several veto sources In the circuit the complementary logic is used the trigger is enabled by the presence of the following signals a A general trigger enable signal This signal is used to have clean start of run and end of run procedures by blocking the triggers at the source It is also used to pause and continue the data acquisiti
132. nches the readout and the recorder processes in all the front end crates participating in the data taking and declared in the Run Control panel The readout process contains the experiment dependent code used to perform the front end electronic readout This code is specified in a separate software module called readList which has to be compiled and linked with the readout main program The readlist module consists of the following four routines e ArmHw called at each start of run to perform the initialization e EventArrived called in the main event loop to discover whether a trigger has occurred e ReadEvent called in the main event loop after the arrival of a trigger to perform the readout of the hardware e DisArmHw called at each end of run to perform the hardware rundown 5 1 1 The main event loop The structure of the readout program and the main event loop are shown in Figure 5 1 be SOR scrip files E ArmHw f t Y 2 3 YEN EOR request p i ma Sak Fi F EventArrived Oo DWsArmHw i bg lt evnt wrie EOR script files yes i E fill circular buffers Reade vent i x emil ee eee Figure 5 1 Main event loop ALICE DATE V2 User s Guide Overview 45 5 1 2 The readout process The readout process is responsible for performing the hardware readout It inserts data into two buffers called RECORDINGBUFFER and monitoring buffer At start of run the readout process
133. ng the SORW AIO se 6 ee Wade eS ee eK Be EH 124 11 3 First time installation 22 125 11 4 Installation of a new release aooaa aaa 129 11 5 Run control configuration oaoa aa aa 129 11 6 Information logger configuration 140 11 7 Monitoring configuration 42 weve ewe wee a 140 ALICE DATE V3 User s Guide 124 DATE installation guide 11 1 Hardware and software platfomms Table 11 1 The system supports the platforms listed in table 11 1 Hardware and software platforms Hardware Operating system LDC GDC Monitoring Run control Motorola MVME 2600 AOS y Vv y y IBM 43P AIX v J y v SUN SPARC Solaris v y y HP HP Unix J PC Windows 95 v PC Linux y 11 2 Getting the software The distribution kits of DATE are accessible from the web An authorization is required The kit repository is located at the following URL http aldwww cern ch Collaborators DATE Releases There are two types of kits called respectively core and full this is indicated in the file name 1 The full kit contains the DATE distribution and the distribution of Java Its size is around 80 MB It should be used for the first installation and in the successive installations when the Java version has changed Since several DATE packages are written in Java it is important to make sure that the Java release on the machine is in step with the one used by DATE 2 The core kit contains exclusivel
134. nnel is opened with a host of different endianness e g PC DEC vs Motorola IBM Sun The two parameters are used to control the swapping algorithm to be used for the data portion of the incoming events their possible use depends on the actual content of the data payload portion of the event and can be summarized as follows Bytes swapping control Data buffer data type sear psc as 8 bit entities signed or unsigned characters FALSE FALSE 32 bit entities e g VMEbus data TRUE FALSE 16 bit entities e g CAMAC data FALSE TRUE In case swapping is not known beforehand monitoring programs should set the two flags to FALSE and swap the data manually once their type is known this will avoid unnecessary double swapping at run time The values that can be given to the two flags are e true perform swapping C 0 0 FORTRAN 1 e false do not swap C 0 1 FORTRAN 0 Zero in case of success otherwise a error code see 4 2 3 for more details monitorDecodeError include monitor h char monitorDecodeError int code SUBROUTINE MONITOR_DECODE_ERROR INTEGER CHARACTER The entry returns the pointer to a string describing the given error code C library or stores the same string into a user given character array FORTRAN library ALICE DATE V3 User s Guide 40 Guide to write a monitoring program Returns C Synopsis FORT
135. o handle the infoLogger information streams must be able to perform a quick online reconfiguration on demand of the infoLogger client processes Streams can suddenly appear creation of new streams and disappear cleanups filtering from the information repository area All this is under control of the DATE processes and of the DATE operator its behavior is therefore unpredictable ALICE DATE V3 User s Guide 86 Guide to use the infoLogger system 8 2 The various streams can then be merged either online or offline and can be used for browsing or processing A standard browser called infoBrowser is available via the DATE infoLogger package see The infoBrowser on page 86 The infoLogger package is also used to implement the DATE statistics scheme where multiple sources can contribute to common statistics records concerning the whole Data Acquisition system For more details see the chapter dedicated to DATE statistics The infoBrowser The infoBrowser is a tool that allows browsing and filtering of messages generated via the infoLogger package It can run on any system capable to open a link with a X11 server and where the standard DATE kit has been installed and properly configured To start the infoBrowser the following procedure should be followed 1 set the DISPLAY environment variable to the appropriate value gt DISPLAY displayHostName displayNum export DISPLAY or whatever is the syntax of your preferred s
136. o the infoLogger host to the given information stream with Fatal severity INFO_ALL infoAll include infoLogger h void INFO_ALL char stream char message void infoAll String stream String message The given message is sent to the infoLogger host to the default and to the given information st ream with Info severity ERROR_ALL errorAll include infoLogger h ERROR_ALL char stream void char message void errorAll String stream String message The given message is sent to the infoLogger host to the default and to the given information st ream with Error severity FATAL_ALL fatalAll include infoLogger h void FATAL_ALL char stream char message j void fatalAll String stream String message The given message is sent to the infoLogger host to the default and to the given information st ream with Fatal severity ALICE DATE V3 User s Guide Injection of messages 97 8 6 4 Constructors Description Public fields Java Synopsis Returns Description The Java LogChannel Class constructors fields and methods This section describes the LogChanne1 Class basic element for all the exchange of information using the infoLogger Java callable interface Please note that all this information is available online at the URLs file date infoLogger LogChannel html http aldwww cern ch documents DATE DATE infoLogge
137. oDataType macro to the name of each of these types must be assigned to the triggerNoDataTable array of elements of type triggerNoDataTableType The variable nbNoDtTrgs must then be set with the number of entries put into the array Listing 6 2 Example of equipmentList c library for the configuration file in Listing 6 3 DAAIADAUTAPWNFPDWOOWMAATNAAIHWNHE BWWWWWWWWWWNNNHNNNNNDN O COMDAIDUABAWNFDWOWOWATAAIHPWNHER CW DDNDUUNUNUNUOUUOHHOP RP PP PPP PAD NP OWOWMDANAHDABWNHFDOOWMATAAIHPWNHE Code for Rand and Corbo equipments PARAMETE KOR K KK OK OK OK OK G L O B A Th R S OK OK OK OK eK eK typedef struct char namePtr GlobParType KOR K KK OK iC O RB O FER A A A A A A A A A I I I typedef struct unsigned int vmeBaseAddressPtr CorboParType void ArmCorbo char parPtr CorboParType corboPar CorboParType parPtr printf Arming corbo with vme base address 1x n corboPar gt vmeBaseAddressPtr void DisArmCorbo char parPtr int EventArrivedCorbo char parPtr void ReadEventCorbo char parPtr struct eventHeaderStruct header_ptr KOR KK KK KK OK OK RAND Fe I I I I I I I typedef struct long eventMinSizePtr long eventMaxSizePtr short eqIdPtr RandParType void ArmRand char parPtr RandParType randPar RandParType parPtr GlobParType gpar GlobParType globPar printf detector name s n
138. ograms these streams can be online streams from a live DAQ system or offline streams typically files available from permanent data storage 3 relaying hosts offering a liaison between monitored and monitoring hosts that cannot establish a direct link due to the presence of network firewalls or gateways It is possible to have any combination of those three functions e g hosts who are monitoring are monitored and offer relayed monitoring to other hosts The DATE monitoring scheme needs configuration only for monitored and relaying hosts in the following situations 1 online monitored hosts LDCs or GDCs who wish to offer online offline or relayed monitoring to itself and or to other hosts 2 hosts part of a DATE system who wish to offer offline or relayed monitoring to other hosts No setup is required for hosts only wishing to perform monitoring either on the same or on remote hosts and a complete DATE installation is not required For the developer of the monitoring program itself a library is available and can be used in stand alone mode Otherwise monitoring programs can be exported to any type of hosts within the set of supported architectures with no need for extra files or special setups No daemons are necessary and no configuration is required on the monitoring hosts If an upgrade from DATE V2 or earlier is performed it is recommended although not necessary to remove all existing setups eventually present in the
139. ointer to the equipment header and a pointer to the raw data block to fill in If the equipment produces data the size in bytes of the data read must be returned 2 Defined in the DATE_COMMON_DEFS event h header file to be included in the library 3 Defined in the DATE_COMMON_DEFS equipment h header file to be included in the library ALICE DATE V3 User s Guide 66 The generic readList 6 5 6 6 5 7 6 5 8 Disarming the equipments If NNN is the name of a type either readout or trigger declared in the configuration file the DisArmNNN function to disarm the equipment of that type must be provided with the following signature void DisArmNNN char The function receives a pointer to a memory region containing the sequence of pointers to the values of the parameters of the component being disarmed see 6 5 3 Triggering If NNNis the name of a trigger equipment type declared in the configuration file the EventArrivedNNN function to signal the trigger arrival must be provided with the following signature int EventArrivedNNN char The function receives a pointer to a memory region containing the sequence of pointers to the values of the parameters of the trigger see 6 5 3 The function must return the value 1 if a new event has arrived 0 otherwise The function references In order to make the functions contained in the library accessible from the generic readList references to
140. ollowing the format convention 6 5 4 Arming the equipments If NNN is the name of an equipment type either readout or trigger declared in the configuration file the ArmNNN function to arm the equipment of that type must be provided with the following signature void ArmNNN char The function receives a pointer to a memory region containing the sequence of pointers to the values of the parameters of the component being armed these values are read at run time before arming the detector from the detector config file and are assigned to the equipment see 6 5 3 6 5 5 Reading the equipments If NNN is the name of a type either readout or trigger declared in the configuration file the ReadNNN function to read the equipment of that type must be provided Since the type may produce data or not if the type NNN is declared as NODATA in the configuration file the function signature is void ReadEventNNN char struct eventHeaderStruct Otherwise if NODATA is omitted the function signature must be int ReadEventNNN char struct eventHeaderStruct struct equipmentHeaderStruct int All the parameters in the call are input parameters The first parameter is a pointer to a memory region containing the sequence of pointers to the values of the parameters of the component being read see 6 5 3 The second parameter is a pointer to the event header Two further parameters appear in the second case a p
141. omplish the following steps in order to perform its function ALICE DATE V3 User s Guide Monitoring and Analysis in CIFORTRAN 31 declare the source providing the data to monitor declare itself to the monitoring scheme declare if necessary the monitor policy he wishes to use declare if necessary the wait nowait policy to be followed OL aR ON get the available event s from the monitoring stream This chapter describes the callable interface available within the DATA monitoring package and its characteristics 4 2 1 Some simple examples In Listing 4 1 we have a very simple example of a monitoring program written in C Listing 4 1 Example of event dump in C 1 include lt stdio h gt 2 include lt stdlib h gt 3 include event h 4 include monitor h gi 6 void printError char where int errorcode ae fprintf stderr 8 Error in s s n 9s where monitorDecodeError errorcode 10 exit 1 11 End of printError 123 13 int main 14 int status 15 Los status monitorSetDataSource EZ if status 0 Ss printError monitorSetDataSource status Loy status monitorDeclareMp C demo mp 20 if status 0 24 printError monitorDeclareMp status Zot for Start of endless loop 23 void ptr 24 struct eventHeaderStruct event 25 268 status monit
142. on b A burst signal when triggers must be gated off during the inter spill period c A ready signal indicating that all the LDCs are ready to collect events This signal is an AND of the ready signals from each LDC ALICE DATE V3 User s Guide The CORBO module 75 LDCs not involved in a given run are prevented to affect the ready signal by the individual disable detector signals The disable detector signals should be generated in accordance with the running conditions The fan out module distributes the triggers to the LDCs There may be events that require reading out some of the LDCs but not all of them The fan out may implement a logic decision and send the trigger only to those LDCs that are involved in the event 7 3 The CORBO module The CORBO module CES RCB 8047 is a VME read out control board It has four identical channels containing a trigger input a busy output two VME interrupt generators and two counters The CORBO registers are defined in Listing 7 3 Listing 7 3 CORBO registers OMDANADUABPWNFPDOOWATAGAIHWNHE 20 Zils 22 2 3 24 254 263 27 28 29 30 Sis S25 33 34 35 36 Ss 38 39 40 define define define define define define define define define define define define define AE define 8 bit define define define define define define define define define define define define define define define define define
143. on the given machine Installation of the monitoring daemon All machines wishing to offer monitoring online offline or relayed to other hosts should install the monitoring daemon mpDaemon This can be done either via the dateNetwork command or manually see 11 3 2 for more details concerning the installation of the daemon The proper setup of the daemon can be verified as follows From any host wishing to perform monitoring run the command ALICE DATE V3 User s Guide 144 DATE installation guide Listing 11 7 Listing 11 8 11 7 3 Remote monitoring test procedure gt telnet host DATE_SOCKET_MON Trying Connected to host Escape character is YAN BWNE Connection closed Substitute host with the name of the machine offering monitoring functions What reported above is a good example When things go wrong one should expect some error messages that should lead to the culprit of the problem e g Example of failed remote monitoring installation 1 gt telnet host DATE_SOCKET MON 2 Trying host 3 telnet connect to address xxx xxx xxx xxx Connection refused 4 telnet Unable to connect to remote host Connection refused In the above example the mpDaemon service has not been properly installed Other useful debugging information may be available within the system console of the machine target the host mentioned above In our previous exam
144. or shells such as sh and bash on Solaris and AIX On those two OSs C shells use a different syntax where each line has to be terminated by a character Other OSs may use different syntax for multi line breaking 9 7 The standard DATE bookkeeping record As delivered in its standard configuration a DATE readout system will produce a standard bookkeeping record available automatically for each run The record will contain the following information e for each LDC readout version number hostname where readout executes number of data events start of burst events end of burst events combined number of physics events number of bursts in the run e for each GDC ALICE DATE V3 User s Guide The standard DATE bookkeeping record 111 event builder version number number of start of run events number of end of run events number of start of run files events number of physics events number of end of link events output file name and number of files written number of records written number of bytes written An example of the standard DATE bookkeeping record is available in Figure 9 4 Tool FontSize Color oy Pum 5695 5446 57 Readout Y 2 04 summary on Bepcalla Data s20 and an sount 231285 Priysles date samt 231785 Etat crunt O Readout V 1 04 sumaary on mopcalt Data 22b and b cunt Z31235 Piiysies date count 231135 ura
145. orGetEventDynamic amp ptr 21 if status 0 26 printError monitorGetEventDynamic status 29s event struct eventHeaderStruct ptr 30 printf Run ld Event 1ld Type ld Length ld Data size d n 3s vent gt eventHeader runNb 32 vent gt eventHeader nbInRun 33 vent gt eventHeader type 34 vent gt eventHeader size 3 5 vent gt eventHeader siz vent gt eventHeader headLen 36 free ptr 37 End of endless loop 38 End of main The program consists of a declaration phase followed by an endless loop where events are fetched from the monitoring stream and their header is printed More in details we can observe the following features line 3 inclusion of the DATE event declaration module ALICE DATE V3 User s Guide 32 Guide to write a monitoring program Listing 4 2 4 2 2 line 4 inclusion of the DATE monitoring declaration module line 16 declaration of the source of monitoring data in this case the online local host line 19 declaration of the monitoring program line 26 the next available event is transferred from the monitoring buffer Similarly in Listing 4 2 we can see a simple example written in FORTRAN Main features of this simple program are line 8 declaration of the source of data in this case the local file tmp runData line 14 transfer of the next available event from the monitoring buffer Example of an
146. ository 9 6 The bookkeeping callable interface 9 6 1 The run record structure 9 6 2 The bookkeeping calls 9 6 3 Single line records examples 9 6 4 Multi line records examples 9 7 The standard DATE bookkeeping record Chapter 10 Conventions and file organization 10 1 DATE environment 10 2 File organization 10 2 1 Structure of DATE_ROOT T 10 2 2 Package directory 10 2 3 Structure of DATE_SITE E 10 3 Environment variables and aliases 10 4 Internet daemons 10 5 Package development 10 6 Logging information 10 6 1 Use of streams 10 6 2 Use of the severity 10 6 3 Use of the facility 10 6 4 Filtering the logged messages at the s source the log level Chapter 11 DATE installation guide he 11 1 Hardware and software platforms 11 2 Getting the software 11 3 First time installation l 11 3 1 Setting up the file base 11 3 2 Internet services 11 4 Installation of a new release 11 5 Run control configuration 11 5 1 Run control windows configuration 11 5 2 The LDC event buffer size ALICE DATE V3 User s Guide 84 86 87 87 87 89 91 92 92 93 97 98 101 102 103 104 105 106 106 106 107 107 108 109 109 110 113 114 114 115 115 116 116 117 117 119 119 119 120 120 123 124 124 125 125 126 129 129 129 137 Viii Contents 11 5 3 Run control configuration parameters 138 11 5 4 Run number le one 2 a R
147. out disturbing the other users DATE To do so the developer keeps in his private directory a replica of date package and a replica of the date_site area with only the files necessary to the package The usage of the private version of the package is obtained by re defining the environment variables DATE _pk_variable to point to the private area The private definition is kept by convention in the file testParams and may be activated by the command dateSetup private_dir package testParams If the program to be tested is a daemon there is a further complication The convention adopted is the following For each d mon two services with different numbers and names are declared one to be used by the standard system and the other for testing purposes The declaration in inetd conf will either indicate two different scripts or the same script with two different parameters In the latter case it is up to the script to sort out the two cases Each DATE package refers to the service via an environment variable giving the socket port number It is sufficient to re define this variable in the testParams file to automatically switch to a private version of the d mon The socket ports used for development are DATE_SOCKET_RCS 6102 DATE_SOCKET_EB 6202 DATE_SOCKET_EBDAEMON 6402 ALICE DATE V3 User s Guide Logging information 119 10 6 Logging information 10 6 1 10 6 2 The DATE packages may generate at run time information
148. own in Listing 11 3 Another example exists in the distribution date runControl runControl config which may be copied in the right place and edited there The file runCont rol config is structured as a sequence of keywords prefixed with either gt or followed by a list of definitions of the items related to the keyword Each item occupies one line and is made of a sequence of parameters The syntax rules are the following e String parameters must be enclosed within quotes e Blank lines are ignored e Blank characters are used as parameter separators otherwise they are ignored e Lines starting with are ignored comment lines e Lines starting with either gt or must specify a valid keyword e A keyword followed by another keyword is ignored even if not valid e Keywords may be given in any order The file is interpreted in one pass A syntax error stops the interpretation of the file remaining lines are ignored ALICE DATE V3 User s Guide 130 DATE installation guide Listing 11 3 Example of the file runControl config 1 Example of configuration file for the runControl program 2 Compliant with DATE v3 2 3 3 4 gt WINDOWTITLE 5 DATE run control 6 7 gt PARFILE 8 runControl params 9z 10 gt LDC 11 rsald01 Front end 1 0 12 rsald02 Front end 2 1 13 mppcal03 Front end 3 2 14 mppc
149. pdated I Figure 2 8 No update indicator 4 Update interval The operator can choose among a set of pre defined update intervals Unnecessary frequent polling should be avoided in order to interfere as little as possible with the data taking At each update messages are sent to all the machines involved which reply sending back the variable values 2 1 2 4 The windows menu This menu Figure 2 9 allows the operator to recall windows other than the main run control window that may have been either closed or iconized or hidden under other windows The content of this menu may vary since plug ins may add options corresponding the windows they may have created ALICE DATE V3 User s Guide Guide to operate the system File View Options windows Configuration parameters Run para status display Max number of event Buffer status rsald01 IBM1 ALD lab BE Figure 2 9 The windows menu 1 Configuration parameters This option recalls the configuration parameters window This window is not usually shown since the operator need not modify it when running in stable conditions It is though the place where most of the configuration parameters can be manipulated The configuration parameters window Figure 2 10 shows a list of common parameters and lists of parameters specific to each machine available to but not necessarily involved in the data acquisition The lists are established according to the d
150. ping system and how to use it 9 1 The DATE bookkeeping package 102 9 2 The statsCollector 44 4 ae sy ee bade Pees Bae 103 9 3 The statsBrowser oes we EM ae ee eo ee ae 105 94 Browsing HTML help pages 6062 0 4e eee wees 106 9 5 The bookkeeping repository 106 9 6 The bookkeeping callable interface 107 9 7 The standard DATE bookkeeping record 110 ALICE DATE V3 User s Guide 102 Guide to use the bookkeeping system 9 1 The DATE bookkeeping package The role of the DATE bookkeeping package is to assist all the processes belonging to a DATE Data Acquisition system to create and update an online database were all the relevant information concerning the development of a given period of data taking called run are stored The flow of information handled by the bookkeeping package can be summarized as follows DATE process DATE process bookkeeping user central bookkeeping repository bookkeeping user Figure 9 1 DATE bookkeeping schematic view An arbitrary number of sources all belonging to the same DATE Data Acquisition System contribute to the editing of a central bookkeeping repository where the relevant data concerning the hardware and the software aspects of each run can be stored The central repository can then used to retrieve these records and to extract specific or generic parameters concerning each run Some examples
151. ple we can see the message host inetd 4692 execv date monitoring SunOS monServer sh No such file or directory A more complete validation of the remote monitoring scheme can be done using the eventDump facility part of the standard DATE monitoring scheme Hosts who wish only to perform monitoring functions do not need to install the monitoring daemon Monitoring of the online monitoring scheme The online monitoring scheme can be monitored using the following commands e monitorClients displays a list of all registered clients eventually with the name of the host they are running on e monitorSpy displays a more detailed snapshot of the monitor scheme A graphical monitoring tool is currently under development It is not possible to monitor the status of an offline or relayed monitoring host ALICE DATE V3 User s Guide Index A application monServer 28 ArmHw routine 44 47 B Bookkeeping callable interface 107 Guide to use 102 host 104 interactive addition of messages 108 package 102 107 records 107 109 110 repository 102 C Configuration Detectors configuration file 68 End of Run file 53 End of Run script 52 Information logger 140 parameters 132 readout 54 Run control 129 Start of Run file 53 Start of Run script 52 D daemon infoLogger 87 Data format definition 21 in C Fortran monitoring program 31 used in the readout program 49 DATE_HOST symbol definition 116 DATE_HOSTNAME sym
152. ppari y a a 1h Run 16S cise Dato Fe TAA GDE a epee P 1 Pot Ae LS eno ippa Praia Pet Pe 17 2 bie Frees arran raand commen 3 TOE i manr qnam Ssa t Figure 2 6 The extended main run control window ALICE DATE V3 User s Guide The control console 13 2 1 2 3 The option menu The option menu Figure 2 7 provides facilities to change the internal behaviour of the control program Timer Bille View Options Windows F Log the info I Log the trace Max numbe E Update remote status W rsald01 IBM1 ALD lab Update interval 1 gald02 IBM2 ALD lab 2sec jppcal03 PPCO3 5 sec 10 sec 20 sec GDC ppeal04 PPC04 30 sec jald01 IBM1 ALD lab 1 min vald02 IBM2 ALD lab B Figure 2 7 The option menu 1 Log the info If enabled all the messages appearing in the information field of the run status panel will be sent to the Information logger as well 2 Log the trace If enabled all the messages appearing in the trace panel will be sent to the Information logger as well 3 Update remote status If enabled the status of the machines in the machine selection panel and the variables in the status display are updated at regular intervals The internal timer displays a rotating segment on the right of the menu bar see Figure 2 7 If disabled a warning appears on the menu bar Figure 2 8 Fle View Options Windows Status not u
153. ption The daemon can be started both from interactive and detached shells as long as the proper environment is available If by accident more then one instance of the stat sCollector is started within one DATE cluster from the same or from different hosts the daemon initiates a self termination procedure A non deterministic algorithm is followed to kill all the instances of the daemon excepted one which continues to perform its normal function When first started the stat sCollector daemon scans the whole content of the DATE_SITE_LOGS logBook stream if necessary updating or creating the relative stats records in the DATE_SITE_STATS repository Once the scan is completed the daemon keeps polling the DATE_SITE_STATS logBook stream waiting for the submission of new records It is possible to force the daemon to immediately terminate the polling cycle once the end of the input file is reached via the nopo11 flag The statsCollector uses by default the online DATE_SITE_LOGS repository It is possible to scan an alternate repository area via the inputDir command line option In this case the use of the nopo11 flag is also recommended there is no use to start polling an offline logBook file as no updates are expected The above procedure can be used to restore the stats records for a given run period saved in an old logBook file e g gt statsCollector nopoll inputDir offlineLogs In the above example the st
154. quipments involved in the readout of the detector saves them in a table and then calls the Arm routines for each active equipment in the order specified in the configuration file 2 ReadEvent This routine generates the equipment header and then invokes the ReadEvent routine for each active equipment in the order specified in the configuration file 3 DisArmHw It calls the Disarm routine for each active equipment in the opposite order as specified in the configuration file 4 EventArrived This routine calls the EventArrived routine of the trigger equipment selected in the configuration file see 6 5 1 6 2 Using the generic readList The use of the generic readList requires the preparation of two files containing the description of the detectors and their equipments 1 The equipmentList c contains the code handling the readout of all the equipments that may be activated in all the detectors Chapter 6 5 gives the detailed description on how to prepare it An example of equipmentList c can be found in date readList 2 The detectors config contains a textual declaration of all the equipments followed by a description of how the equipments are shared by the detectors ALICE DATE V3 User s Guide 60 The generic readList Chapter 6 6 gives the detailed description on how to prepare it An example of detectors config canbe found in date readList These two files are strictly correlated and must match one ano
155. r gt LDCPAR in run Control config Configuration parameters win dow LDC and GDC parameters under both gt LDC PAR and gt GDC PAR in runControl config 11 5 2 The LDC event buffer size The event buffer size in the LDCs is given a default value of 1024000 bytes This value can only be changed by adding the following statements in the file DATE_SITE_CONFIG runControl config gt TCLI EVAL set ldc EventBufferSize x where x is the desired buffer size in bytes 1 The parameter recordingBufferSize defined in date runControl rceShm h cannot be changed by the operator It is used to save the size of the event buffer after the shared memory segment has been created ALICE DATE V3 User s Guide 138 DATE installation guide 11 5 3 When the value is changed it is necessary to destroy all the shared memory segments of the event buffers with IPCS a simpler way is to bootstrap all the LDCs Run control configuration parameters The parameters declared under the keywords RUNPAR CNFPAR LDCPAR and GDCPAR will appear in the run control windows either the main one or the one labeled Configuration parameters Figure 11 1 The operator is supposed to set the values of these parameters in the corresponding entry fields on the windows maxEvents See description in Table 11 3 maxBytes Maximum number of kBytes to be collected in a run It is used to limit the amount of data to be han
156. r Continue trigger It respectively either disables or enables the trigger 4 Disconnect Connect It respectively either disconnects or connects the control console from to all the machines involved in the data acquisition The menu bar The File menu The options of the file menu Figure 2 4 are the following File View Options Windows Save parameters Load parameters LDC W rsald01 IBM1 ALD lab rann amean armani el The file menu 1 Save parameters Saves on a file all the parameters switches and menu options selected on the main run control window and on the configuration parameters window see Figure 2 10 The saved conditions can be restored with the menu option File Load parameters 2 Load parameters Restores from a file all the parameters switches and menu options 3 Quit Stops the run control program The same effect is achieved by the close window button of the window frame If there is a run active an alert window will ask for a confirmation since the action will compromise the run The View menu The options of the view menu Figure 2 5 allow the operator to display additional panels onto the main run control window The usage of these additional panels is reserved to specialists therefore it is suggested to keep these options disabled ALICE DATE V3 User s Guide 12 Guide to operate the system Bile View Options Windows I Parameters file name I Trace Max T
157. r LogChann el html1 The first URL is relative to the local machine we suppose that the DATE has been locally installed The second URL is protected and belongs to the DATE developer team the VID field can be used to validate the coherency of what is installed with what is documented in the above URL Class LogChannel LogChannel channel new LogChannel LogChannel channel new LogChannel String logHost where logHost is the name of the infoLogger host A LogChannel object is created ready to send messages to default infoLogger host or to a given alternative logHost static final String VID The version ID of the LogChannel Class static boolean quiet Disable when t rue and enable when false generation of error and warning messages static boolean verbose Disable when t rue and enable when false generation of debug messages validate static final boolean validate True if the LogChannel1 object can be used to transfer log records false if the state of the LogChannel object is not correctly set The LogChannel is validated and set ready for use if possible All the fields and parameters are checked and all the TCP IP channels are validated and if necessary created This call can be used to ensure that the class can be immediately used to log new information It does not ensure that the infoLogger host daemon will actually perform the operation however it en
158. r has arrived The user is supposed Returns to insert the readout data into the area pointed to by the second parameter and to fill the following fields in the event header pointed to by the first parameter e event gt eventHeader nbInRun initialized to NOT_SET_TAG mandatory This variable is the event number in the run The readout process logs an error message if this field is not filled upon return from the routine e event gt eventHeader burstNb initialized to 0 optional This variable is the burst number e event gt eventHeader nbInBurst initialized to 0 optional This variable is the event number within burst e event gt eventHeader type initialized to PHYSICS_EVENT mandatory This variable is the type of record The readout process increments the trigger number in the variable event gt eventHeader triggerNb only for PHYSICS_EVENT type of record and not for other types of records such as SOB and EOB data e event gt eventHeader errorCode initialized to 0 optional This variable is an experiment dependent error code that the user may set to signal any kind of error occurred during the readout phase e event gt eventHeader deadTime initialized to 0 optional This variable may be set by the user to measure the dead time in seconds for the readout of a particular equipment e event gt eventHeader deadTimeusec initialized to 0 optional This variable may be set by the user to measure the dead ti
159. r is declared The syntax of the monitor source parameter is the following Monitor source parameter syntax a local online default file local file both full and relative paths are accepted full path recommended host remote online on node host file host remote file on node host the full path to the file should be given host1 host2 remote online on node host1 via the relay host host2 file host1 host2 remote file on node host1 via the relay host host2 the full path to the file should be given If a remote monitoring is specified and the remote hostname points to the local host then local monitoring is assumed and no transfer take place over TCP IP The monitoring library is able to resolve host aliasing and multi interface hosts Zero in case of success else a error code see 4 2 3 for more details monitorDeclareMp include monitor h int monitorDeclareMp char INTEGER MONITOR_DECLARE _MP CHARACTER The given string is used to declare the monitoring program This can be used for debugging for fine tuning and to monitor the monitoring scheme Zero in case of success else a error code see 4 2 3 for more details monitorDeclareTable include monitor h int monitorDeclareTable char ALICE DATE V3 User s Guide Monitoring and Analysis in CIFORTRAN 35
160. rator may choose in the machine selection panel the pattern of machines involved in the data acquisition If the configuration has not changed the operator will leave the pattern previously saved and automatically retrieved at the program startup Otherwise a new configuration pattern will be selected Figure 2 2 The main run control window after connection ALICE DATE V3 User s Guide 10 Guide to operate the system The Connect button instructs the run control program to get in touch with all the selected machines The main run control window will change as shown in Figure 2 2 The pattern of machines in the machine selection panel cannot be changed any longer all the buttons are disabled In addition a new window will appear on the screen called the status display window Figure 2 3 The status display shows a list of variables for each machine involved in the data acquisition The list is established according to the declarations in the configuration file The variables are updated at fixed intervals Figure 2 3 The status display Before starting a run the operator may need to change the run conditions There are various means of doing it some of the conditions are expected to be changed more frequently than others therefore they have been given more visibility 1 The run conditions panel of the main run control window Figure 2 1 a Enable disable recording global switch b Select t
161. rds inserted in the monitor buffer The Start Of Run and the End Of Run sequences have been split into phases corresponding to the points enumerated above At each time the timeout limit set by the operator in the Run Control window and expressed in seconds is restarted in order to allow for long initialization or ending procedures Diagnostics messages have been added in case of time out to know which phase originated it The recorder process The recorder process performs one of the two following functions e if the recording device in the Run Control panel does not terminate with the recorder process is responsible for writing the readout data as well as all the records inserted by readout in RECORDINGBUFFER to an output file on the local disk This is typically used if the data acquisition system is composed by a single front end processor without any event building functionality In this case the name set by the user as recording device is the full path and the filename on disk which will contain the data for example tmp my_raw_data dat The directory in which the filename resides should have write access for the user nobody one can either give write access for everybody or change the owner of the directory sy e if the recording device field ends wit the recorder process assumes that it is the name of the remote event builder machine and tries to open a TCP IP connection to it in order to send the readout
162. re must be a run identifier the record itself and a trailer marking the end of the record The union of these three elements constitutes a single logical message to be sent to the LogBook stream This message can be sent in one single packet or split over several packets The stat sCollector daemon will use the run identifier to associate the record to the proper run descriptor the record trailer to know when to use the data received and the record body to compile the final record Intermixing with similar data coming from other DATE processes does not constitute a problem each client process is uniquely identified by the package hostname PID tags used by the statsCollector for its record building phase The run record structure sew Run record a ROD Re OR Se The source run record Each contribution to a run record must have a header specifying the run ID relative to the record the body and a trailer marking the end of the record A bookkeeping data source can give the required information either via multiple calls or with a single multi line call The syntax for the Run ID is the following ALICE DATE V3 User s Guide 108 Guide to use the bookkeeping system 9 6 2 C Synopsis Java Synopsis Shell Synopsis Description C Synopsis Java Synopsis Shell Synopsis Description Run where is the unique ID of the run This ID can be extracted either from the DATE shareable data segment or
163. readout it purpose indicate the burst depends on the condition to be purpose passed either from the operator to readout or the other way round eventsRecorded Counter of events Either readout or Status display win recorded on disk event builder dow under either either by the LDC gt LDCSTATUS or if alone or by the gt GDCSTATUS in GDC runControl config bytesRecorded Counter of kBytes Either readout or Status display win recorded on disk event builder dow under either either by the LDC gt LDCSTATUS or if alone or by the gt GDCSTATUS in GDC runControl config eventsTransferred Counter of events Readout Status display win transferred from the dow under gt LDC LDC to the GDC STATUS in runControl config bytesTransferred Counter of kBytes Readout Status display win transferred from the dow under gt LDC LDC to the GDC STATUS in runControl config readoutFlag A number indicat Readout Status display win ing the phase of the dow under gt LDC start of run and end STATUS in of run procedures At the beginning of the procedure it is set to 1 then it will run from 1000 to 6000 at start of run and from 11000 to 17000 at end of run Zero means com pletion runControl config ALICE DATE V3 User s Guide Run control configuration 135 Table 11 3 Run control parameters ALICE DATE V3 User s Guide Parameter name Description Set by To appear in window phaseTimeout Maximum
164. recording 3 284 ho oe He a 20 3A Data tiles s sse 20404 6 wed ee RB ey ade Hee S a 20 3 3 Thedataformat 002 0 05004 21 pe Th eeventtypes aud ce bo Bee Chee Cee whee ER ES 23 3 5 The full event format 5004 24 ALICE DATE V3 User s Guide 20 Data recording and data format 3 1 3 1 1 3 2 Data recording Data recording from a LDC The data generated in a LDC can be recorded in two different ways either by recording them to a local disk file or by sending them to a GDC If the data acquisition system is composed by a single front end processor without any event building functionality the recorder process can write the readout data to an output file on disk The full path of the directory and filename has to be specified as recording device of the LDC in the run control configuration parameters panel for example tmp my_raw_data dat If the data acquisition system includes a GDC for the event building the recorder process can send the data to the GDC This option is selected by specifying a recording device that ends with In this case the recording device name must be the host name of the event builder machine for example mppcna57eb01 The recorder process uses the infoLogger facility to report and trace error or abnormal conditions and to trace state changes Data recording from a GDC The data assembled in a GDC can be recorded to a local disk file by the ev
165. rent from 0 when this limit is reached the program recording the data recorder in the LDC or the event Builder in the GDC will open a new file The full path of the directory and filename has to be specified as recording device of the GDC in the Run Control panel Some characters have a special meanings is replaced by the event builder host name and is replaced by the current run number With the following value of the recording device data run_ raw the data of the run 1020 will be recorded into the file data run_1020 rawif there is no limit on the filesize If there is a limit on the maximum file size the data will be recorded to a sequence of files Their filenames will be formed by the addition of the original filename for this run and a sequential number data run_1020_000 raw data run_1020_001 raw data run_1020_002 raw etc The sequential number 0 is reserved for the data recorded at start of run It includes the records of the types START_OF_RUN and START_OF_RUN_FILES The data format The data format is described by the event Struct structure defined in the file date commonDefs event h An event is constituted by an event header described by the eventHeaderStruc structure followed by the event data The fields in the event header are shown in Listing 3 1 The meaning and the size of each field in the event header are explained in Table 3 1 The DATE package which writes each field is also indicated
166. ring programs should include the prototypes declaration monitor h either via the directive I DATE_MONITOR_DIR DATE machines or copying the prototypes declaration locally non DATE machines ALICE DATE V3 User s Guide Monitoring and Analysis in CIFORTRAN 33 4 2 3 4 2 4 C Synopsis Linking monitoring programs require the following libraries 1 S DATE_MONITOR_DIR S DATE_SYS libmonitor a or the equivalent for non SHIFT hosts S DATE_MONITOR_DIR DATE_SYS libmonitorstdalone a see below for more details 2 network libraries on SunOS lsocket 1ns1 eventually required for socket I O man 3n socket should normally give pointers to those libraries 3 on machines where the CERN Remote File I O RFIO library is available the SHIFT library usually available as usr local lib libshift a plus all the libraries eventually required by the SHIFT library itself not needed if the libmonitorstdalone a library is used The SHIFT library referenced by the monitoring I O package is used to access hosts whose PDS is available on SHIFT servers via Remote File I O RFIO system e g the CERN ALICE WorkGroup server ION the CERN batch processing facility SHIFT and the CERN High Performance Storage System HPSS servers If access to any of these three facilities is not required the inclusion of the libshift a library is not necessary In this case a special version of the mon
167. ror and readList_errorSource to signal error conditions which will provoke the log of a message and the termination of the run EventArrived include rcShm h include event h int EventArrived The EventArrived routine is called to know whether a trigger has occurred It is up to the user to either poll and return immediately with 0 if no trigger has occurred or to wait for an interrupt with an appropriate driver call for the hardware The main readout program calls this routine as follows in a strictly close loop without sleeping as shown in Listing 5 1 Upon occurrence of a trigger the routine should return a value different from 0 If the trigger is to be regarded a valid event the value must be gt 0 The routine should return 0 if no trigger has occurred It should use the variables readList_error and readList_errorSource to signal error conditions which will provoke the log of a message and the termination of the run Calling EventArrived 1 while waiting for a trigger check that nobody stopped the run E 2 3 while triggerArrived EventArrived 4 if DAQCONTROL gt readoutFlag goto finish Si ReadEvent include rcShm h ALICE DATE V2 User s Guide The readList templates 49 include event h int ReadEvent struct eventHeaderStruct unsigned short Description The ReadEvent routine is called after a trigge
168. rror String message The given message is sent to the infoLogger host standard information stream with Error severity FATAL fatal include infoLogger h void FATAL char message void fatal String message The given message is sent to the infoLogger host standard information stream with Fatal severity INFO_TO infoTo info include infoLogger h void INFO_TO char stream char message void infoTo String stream String message void info String stream String message The given message is sent to the infoLogger host to the given information stream with Info severity ERROR_TO errorTo error include infoLogger h void ERROR_TO char stream char message j void errorTo String stream String message void error String stream String message The given message is sent to the infoLogger host to the given information stream with Error severity FATAL_TO fatalTo fatal ALICE DATE V3 User s Guide 96 Guide to use the infoLogger system C Synopsis Java Synopsis Description C Synopsis Java Synopsis Description C Synopsis Java Synopsis Description C Synopsis Java Synopsis Description include infoLogger h void FATAL_TO char stream char message void fatalTo String stream String message j String message void fatal String stream The given message is sent t
169. s for logging all sorts of information debug run time statistics plus the tools to manipulate format and browse the resulting messages The DATE infoLogger package gives to both the developer and the operator a set of tools to perform the above functions This section describes the facilities available within the infoLogger how to use them and how to exploit their features 8 1 Introduction s e see des Re Seed a te Be vtec ds eRe 84 G2 TheinfoBtoWsS er s es siss sia Me a boeg EERE OS 86 8 3 Browsing HTML help pages 24 naaa aaa 87 8 4 The information repository ooo 87 8 5 Extracting portions of the log files naaa aaa 89 8 6 Injection of messages o oo d aw bade dwnes 91 ALICE DATE V3 User s Guide Guide to use the infoLogger system 8 1 Introduction A DATE Acquisition System is composed on several processes running on one or more CPUs These processes are often detached from the operator s and the developer s environment they can be started as network daemons they can be child of network daemons or they can run as standalone images However all DATE processes should be able to generate asynchronous information such as debug statements run time information and statistic records Furthermore the DATE operator developer should be able to gather sort and browse these messages in real time All this can be achieved via the DATE infoLogger package whose basic features are e capability to merge
170. start from the example provided to write a custom readList in which the four routines included in the template and described below must be supplied or use the generic readList see next chapter Upon return from the user routines the readout process checks the content of the global variable readList_error whose value allows the user to signal error conditions If its value is different from 0 the readout process logs an error message containing the value of the variable and the name of the routine originating the error set by the user in the global variable readList_errorSource and asks to stop the run When using the generic readList the variable readList_errorSource is filled by the readout process ArmHw ALICE DATE V2 User s Guide 48 Guide to prepare a readout program Synopsis Description Returns Synopsis Description Returns Listing 5 1 Synopsis include rcShm h include event h void ArmHw The ArmHw routine is called at each start of run after the execution of the start of run Unix scripts and before the transfer of the start of run files on the output medium The routine should perform all the actions needed at the beginning of the run such as the initialization of the hardware and of the trigger and the pre encoding of hardware addresses to be saved in global static variables The routine does not return any value It should use the variables readList_er
171. sures that the remote daemon has been started and is responding as expected ALICE DATE V3 User s Guide 98 Guide to use the infoLogger system Java Synopsis Returns Java Synopsis Description Java Synopsis Description Java Synopsis Returns Description Java Synopsis Returns 8 6 5 getError final String getError A string describing the last error encountered clearError final void clearError The last error string is cleared see getError setFacility static final void setFacility String facility The facility used to identify the source of messages is set to the given facility value default value LogChannel DEFAULT_FACILITY getFacility static final String getFacility The facility identifier associated with the object The facility identifier used for the LogChannel Class is returned toString final String toString A string describing the object Interactive injection of messages At any moment the DATE operator can inject messages in the information stream The only pre requisites are to have executed the DATE setup procedure which enables the use of the standard logging primitives log and logTo and to have defined the DATE_SITE_LOGS environment variable ALICE DATE V3 User s Guide Injection of messages 99 The log command sends a message to the common log stream DATE_SITE_LOGS runLog By default
172. t count O Event Bullder VOl T36 ary Humber of Start of Run z Humber of Erd of Fun 2 Humber of Start of Pun Files 2 Humber of Prysies event Z31265 0 incomplete Humber of of link L Sutput Filename p Humber of records 731293 D errors Data recorded 77 MEytes 7 Disable uto follow Figure 9 4 Example of standard DATE bookkeeping record ALICE DATE V3 User s Guide D sutevert mismateh ddata raun5655 r an 2 fila sii 112 Guide to use the bookkeeping system ALICE DATE V3 User s Guide Conventions and file Organization This chapter describes the organization of directories in DATE and the conventions adopted for naming files and directories Other conventions concern the usage of Inet dzemons 10 1 DATE environment 0 02000 114 10 2 File organization cies oes Ge aa HES 114 10 3 Environment variables and aliases 116 10 4 Internet d mons aaa 020005 117 10 5 Package development nonoo dada 117 10 6 Logging information osaa Ee ow BO 119 ALICE DATE V3 User s Guide 114 Conventions and file organization 10 1 10 2 DATE environment The operation of DATE requires obeying a number of conventions concerning the file organization and the Internet daemons These conventions have been established in order to fulfil the following requirements e DATE runs at a site made of a cluster of host machines communicating via network
173. t input egrep pattern sort unigq gt tmp logs result log The input parameter is the name of the input information stream s wild cards allowed such as DATE_SITE_LOGS runLog or DATE_SITE_LOGS It is possible to specify an explicit list of two or more files The selection pattern is a standard egrep regular expression Details can be found in the Unix egrep manual page Several fields following the infoLogger standard formatting scheme can be used to restrict the filtered set The available selection fields together with some examples are illustrated in table 8 1 Table 8 1 Selection fields Pattern Action Example AMMDDhhmmss Selection based on the 1012110000selects all time of messages generated on generation of the mes 12 October 11 00 00 AM sage MM month 01 12 DD day of month 01 31 hh hour of day 00 23 mm minutes of hour ALICE DATE V3 User s Guide source facility 00 59 ss seconds of minute 00 59 hostname Selection based on the mppcal07 selects all source host name messages coming from the host named mppcal07 PID Selection based on the 19672 selects all mes source PID Process IDen sages coming from the tifier process whose PID is 19672 decimal facility Selection based on the readout selects all messages coming from the facility readout Guide to use the infoLogger system Selection fields Pattern Action Example severity Select
174. taining only the essential information while all the details are given in the message to the own stream The messages in this stream are also sent to the own stream therefore the routine calls to be used to inject messages into this stream are of the type logA11 3 The LogBook stream This stream may be used by any program to record messages on the run bookkeeping file The messages can only have the severity information Ar run time each process may produce messages addressed not only to the stream assigned to its own package but also to all the streams assigned to the libraries it uses Use of the severity The severity must be chosen and specified for each message according to the definitions given in the chapter on the infoLogger system The infoTo 1 Information streams are recorded by the infoLogger onto files each stream has got its own file ALICE DATE V3 User s Guide 120 Conventions and file organization 10 6 3 10 6 4 infoA11 call type force the developer to make the choice while in the logTo type the severity is an optional parameter The three possible severities are specified by the symbols LOG_INFO LOG_ERROR LOG_FATAL A message should be flagged as fatal only if it signals a catastrophic action such as a run abort A fatal message should mark the end of a series of unrecoverable errors that have generate normal error messages Use of the facility The facility name identifies an a
175. te the DATE_SITE_STATS area via the command gt statsCollector inputDir logBookDir outputDir stats nopoll The above command instructs the statsCollector to pass the logBookDir logBook file once to create the corresponding run records and to ALICE DATE V3 User s Guide The bookkeeping callable interface 107 9 6 9 6 1 Figure 9 3 save the result in the directory stats It is possible to use alternative file names via the inputFile flag The same operation can be used to restore all the run records relative to the current DATE_SITE_LOGS logBook all the operator has to do is to stop the statsCollector currently running if any and to start a fresh copy of the daemon This will recreate all the run records relative to the current logBook and can be used to restore a lost or corrupted DATE_SITE_STATS area The bookkeeping callable interface The bookkeeping package proposes a callable interface for C Java and shell Tcl Tk code can make use of the shell interface As the bookkeeping transport layer is based upon the infoLogger package the same rules and definitions used for the infoLogger are also valid for the bookkeeping For more details concerning the features and the use of the infoLogger package please refer to Chapter 8 The basic principle for the injection of bookkeeping records is the following A DATE process provides the information for each record on a run by run basis for each record the
176. te the file detectors config froma window without using an editor The program may be activated by invoking the following command gt readoutControl The program will open a window as shown in Figure 6 3 The tool is allowed to perform the following operations 1 Equipments assigned to a detector can be activated or de activated 2 The value of the parameters provided to both detectors and equipments can be changed Coe fed See ees Cece ee Dele Figure 6 3 The readoutControl window The modifications are applied to the file only if the menu option File Save configuration is invoked They do not have any influence on the current run ALICE DATE V3 User s Guide 62 The generic readList since they will be taken into account only at the next start of run provided that the edited file is DATE_SITE_CONFIG detectors config 6 4 The equipment header The equipment header provides a further division of the sub event raw data Within a detector it defines blocks of data generated by an equipment Figure 6 4 Seb cven data Figure 6 4 Sub event structure The equipment header has the following structure Figure 6 5 Equipment type a 16 bit integer number identifying the type of the equipment declared in the configuration file The type is a number which uniquely identifies a class of identical equipments It is a convention of the experiment Header extension length a 16 b
177. the connection phase monitoring program can declare themselves to the monitoring scheme This allows easy tracing of each client and makes possible to fine tune the runtime parameters of the monitoring system When a monitoring program connects itself with the experimental stream it has the capability to declare a monitoring policy for any given event type This policy can require all events for monitoring must policy a random share of events for monitoring yes policy or no monitoring at all none policy It is important to understand the impact of a given monitoring policy on the Data Acquisition system and on the monitoring environment A monitoring program requesting a must policy must process the information as fast as it will be offered or it might stall the entire data acquisition system On the other hand the exclusion of certain classes of events unwanted for a given type of monitoring will reduce the overhead on the 1 The term PDS defined in the ALICE technical proposal is used here with a wider meaning also covering permanent semi permanent and temporary storage usually located in the physical path between the Data Acquisition online buffer and the final PDS ALICE DATE V3 User s Guide Guide to write a monitoring program online host and on the interconnecting network as less data will be stored and transferred between the online producer readout or event builder and the consumer the monitoring program
178. the file date runControl rcShm h The parameters declared under the keywords CNFPAR LDCPAR and GDCPAR will appear in the run control window labeled Configuration parameters Figure 11 1 These parameters are the ones that are relatively stable and do not need to be changed by the operator in normal production runs The parameters declared under the keyword RUNPAR will appear in the main run control window These parameters are supposed to be possibly changed from run to run such as maxEvents Table 11 3 Run control parameters Parameter name Description Set by To appear in window maxEvents Maximum number The operator viathe Run control main of events tobe col run control window Run lected in a run parameters under When a DAQ gt RUNPAR in run machine hits the Control config limit an end of run request is issued Zero means no limit ALICE DATE V3 User s Guide Run control configuration Table 11 3 Run control parameters ALICE DATE V3 User s Guide Parameter name Description Set by To appear in window maxBytes Maximum number The operator viathe Configuration of kBytes tobe col run control parameters win lected in a run dow Common When a DAQ parameters under machine hits the gt CNFPAR in run limit an end of run Control config request is issued Zero means no limit maxEventSize Maximum event The operator viathe Configuration size in bytes Used run control parameters w
179. the message is logged as Information but this behavior can be changed via command line flags To get a complete list of these flags use the help option An example of use of the Log command is gt log This is a log message The logTo command performs the same function of the log command but the message is directed to a named information stream e g gt logTo infoStream This is a log message will store the log messages in the file DATE_SITE_LOGS infoStream As usual the help flag can be used to get a description of all the available options For more details on all the above subjects please refer to the file date infoLogger README or to the protected document at the URL http aldwww cern ch documents DATE DATE infoLogger ALICE DATE V3 User s Guide 100 Guide to use the infoLogger system ALICE DATE V3 User s Guide Guide to use the bookkeeping system To get the maximum results from a Data Acquisition system run the experimental data may not be enough Other information has to be stored and made available for post run analysis the status of the beam of the hardware of the software the identifier of the data files and so on The DATE bookkeeping system provides a way to edit a run record describing all the relevant facts concerning a given run by assembling the information coming from all the components of the Data Acquisition system This section describes the features of the bookkee
180. the setting of the parameters that will be passed to the readout routines Therefore it is possible to modify the readout program behaviour without changing the readout executable code 1 Remark The LDCs data blocks can be identified in the event structure as separate sub events It is desirable to have a dedicated LDC for each detector in order to be able to distinguish its data In general though a detector with a large number of channels may be served by several LDCs Still the blocks of data from each LDC will be separately identified they may be considered as sub detectors Throughout this chapter the term detector will be used as synonym of LDC and sub detector it will indicate a front end processor generating identifiable blocks of data ALICE DATE V3 User s Guide Using the generic readList 59 d An interactive program provides a graphical interface to manipulate the configuration file mane Figure 6 2 Readout and the generic readList The equipment handling routines must be provided in a single file called equipment List c The routine names are fixed by convention the name is obtained by concatenating the prefix Arm Disarm ReadEvent and EventArrived with the name of the equipment type as declared in the detectors configuration file The generic readList library implements the following functions see the readList template in Chapter 5 2 1 ArmHw It scans the configuration file identifies the e
181. ther There is no tool to make sure that this is the case Error conditions due to a mismatch are discovered at the start of run and will immediately stop the run There are essentially three conventions tying equipmentList c and detectors config 1 The name of the equipments in detectors config constrains the name of the readout routines in equipment List c a prefix is added to the equipment name as explained in 6 5 1 and 6 5 2 2 The readout routines in equipmentList c must be declared using a set of macros provided in readList_detectors h as explained in 6 5 8 These macros provide the link between the equipment name read by readList from detectors config and the address of the readout routines to be called by readList 3 There are two sets of parameters that can be passed to the readout routines the first is specific to the detector parameters specific to a given LDC the second is specific to a given equipment the parameters values may be different for each declaration of the same equipment in a detector The parameter declarations in detectors configand equipmentList c must match as described in 6 5 3 To prepare a readout program using the generic readList the following steps must be followed 1 Create equipmentList c into a private directory 2 Create the detectors config configuration file into the directory pointed to by DATE_ SITE CONFIG 3 Copy the date readList GNUmakefile_readout makefile into th
182. time When any of the above conditions is met it is a good idea to start a cleanup of the information repository We also suggest to plan a periodic cleanup either by manual or automatic procedure As the information streams are implemented as plain text files any tool capable of moving and archival of such entities can be used tar zip gzip backup We suggest not to remove the files as their content can be used at any time to trace back running conditions and or statistics for any given time of a run The files should instead be moved to an offline area for long term archival eventually packed and compressed to gain flexibility and to save resources With the acquisition system preferably in quiescent state no run in progress no processes active dump the content of the DATE_SITE_LOGS in a save set e g gt tar cf offline logs xxx tar DATE_SITE_LOGS gt gzip 9 offline logs xxx tar The field xxx should be a unique identifier for the save set content e g run dates run numbers shift number The offline directory can be another disk or a permanent storage device according to the architecture of the data acquisition system infrastructure It should be noted that whenever the DATE_SITE_LOGS partition becomes too full the offline area MUST be on a different partition where enough disk space shall be made available prior to issue the above commands The whole content of the repository can now be removed gt rm
183. to measure the readout dead time 2 Channel 2 is used to count all the events whether seen by the LDC or not This number should be written in the sub event header since it uniquely identifies the event An event can only be built by putting together all the sub events with the same event number since the trigger number may vary from one LDC to another ALICE DATE V3 User s Guide Triggering with the CORBO 77 7 4 2 Start of run initialization The CORBO module must first be initialized This is done in the routine ArmHw Since this routine may initialize other interfaces besides the CORBO we shall refer here to a routine called ArmCorbo which may be called by ArmHw The code of ArmCorbo performs a series of operations First it maps a virtual memory window to the VME address space occupied by the CORBO as shown in Listing 7 5 Listing 7 5 Mapping the virtual memory window to the CORBO VME address space OPWNHNFDOWO ATAU HAWNE onw7o Zs 223 23s 24 Zoe 26 Zils 28 29 detectorList c include corboDef h insigned int corboBaseAddr unsigned int eventArrivedFlag register unsigned char regPtrChar register unsigned short regPtrShort ArmCorbo Set the offset and the size of the VME window unsigned long vmeWinOffset vmeWinSize vmeWinOffset 0 Start from the bottom vmeWinSize O0xF00000 Map part of the 24 bit the VM space EJ Map the
184. tributed environment The basic dataflow architecture is organized along parallel data streams working independently and concurrently followed by an event builder stage where data are merged and eventually recorded as a complete event A view of this architecture is depicted in the middle of figure 1 1 where terms such LDC and GDC used throughout this guide are illustrated The LDC Local Data Concentrator is the front end processor whose main purpose is to readout the front end electronics of a given detector or section of a detector The LDCs manage concurrent streams of data the triggering system provides the necessary synchronization with the physics events The LDC data streams are injected on the data acquisition network shown in figure 1 1 as an Ethernet switch through which they reach the GDC The GDC Global Data Collector is a processor that performs the event building function It collects the various sub events from the LDCs puts them together and encapsulates them with the proper event structure It also performs the recording function The conditions imposed to the hardware architecture in order to support DATE are minimal The operating system of all the processors must be Unix b All the processors must share the same file system c All the processors must be linked to a network supporting the TCP IP stack and the socket library Fan centred Riim aama Rira configuration liferation reporting Figure 1
185. trol configuration 131 Table 11 2 Keywords of the configuration file Keyword Following items COMMENT PARF ILE LDC GDC RUNPAR CNFPAR LDCPAR GDCPAR LDCSTATUS ALICE DATE V3 User s Guide WINDOWTITLE Catch all all the following lines are ignored until the next keyword The title of the run control program main window The directory and file name of the run control parameters If specified the run control program will try to read the param eters from this file and will propose this file as default loca tion for further parameter saving and restoring The list of the machines where the front end software runs Each item is the name of the machine followed by a label which will appear in the run control main window The last parameter is the detector bit which will be written in the detector mask field of the event header Valid detector bits are 0 up to 126 This list contains all the machines among which the active machines will be chosen The list of the machines where the event builder and the monitor process should be launched at start of run Each item is the name of the machine followed by a label which will appear in the run control main window This list contains all the machines among which the active machines will be cho sen The list of control parameters that will appear in the main window under the heading Run parameters These parameters will be set at st
186. ts deadTimeusec 32 bits detector identification mask initialized to 0 by the read out process Each bit identifies one LDC The event builder process will write in this field the detector bit a as declared by the user in the runControl config file Valid detectors bit are from 0 to 94 Time in number of seconds since 0 00 GMT 1 1 1970 set by the readout process time in milliseconds to be added to the previous field set by the readout process error code for the event initialized to 0 by the readout process This field is an experiment dependent error code that the user may set to signal any kind of error occurred during the readout phase dead time for the event readout in seconds initialized to 0 by the readout process This field may be set by the user to measure the dead time in seconds for the read out of a particular equipment dead time for the event readout in milliseconds initial ized to 0 by the readout process This field may be set by the user to measure the dead time in milliseconds for the readout of a particular equipment 3 4 The eventtypes The event types or record types are defined in the same include file date commonDefs event h as shown in Listing 3 2 The first usage of the event type is to identify each type of event or record The event type is also used by the event builder to determine whether the event building has to be applied on a given event Only the sub events of the type
187. ure 4 2 To off load the online environment it is possible to run the monitoring program on another host linked to the first via LAN or WAN see Figure 4 1 bottom diagram The result is similar to what we achieved in the first configuration with the advantage of freeing resources on the Data Acquisition host at the price of an increased load on the interconnecting network between the two machines The same Data Acquisition system can have without reconfiguration several local and remote monitoring programs all running simultaneously and getting their data from the same source However each monitoring program can receive its data to monitor from one source at a time It is possible to switch forth and back between different data sources within the same monitoring program although this practice is not recommended Another operating mode of the monitoring library shown in Figure 4 2 allows the same functions on offline streams usually coming from the experiment s Permanent Data Storage PDS This setup allows direct monitoring from the PDS server or from other hosts batch server desktop or workstation not connected to the PDS media This configuration can optionally make use of the CERN Remote File I O system library RFIO to access SHIFT or HPSS disks servers available at CERN PDS attached host Remote host local remote monitoring monitoring program program The DATE offline monitoring During
188. y the DATE distribution Its size is around 15 MB therefore it is faster to retrieve It may be used for upgrading the DATE versions within a major release The name of the kit contains the date of generation and is suffixed with either tar or tar gz Before retrieving the distribution kit on your file system e g on tmp make sure that the tmp directory has the read write execute protections for everybody otherwise set the privileges in the following way gt chmod a rwx tmp ALICE DATE V3 User s Guide First time installation 125 11 3 Arsttme installation 11 3 1 Setting up the file base 1 Login as root 2 Authorize root to open a remote shell on all the machines that will be running DATE This is done by editing or creating the rhosts file to contain a line with the name of the machine where you ar logged in 3 Create the directory where the DATE software should be installed You should call it date If you choose another name e g date_root you must set up a symbolic link in the following way gt In s date_root date Such a link must be declared in all the machines in the system 4 Create a directory where your experiment specific files will reside 5 Define the symbol DATE_SITE to point to the new directory above In Bourne shell do the following gt export DATE_SITE DATE_SITE directory In C she11 do the following gt setenv DATE SITE directory In BASH do the following gt declare

Download Pdf Manuals

Related Search

Related Contents

English User Manual  EnGarde Secure Linux 3.0 -- Quick Start Guide  Proyector Manual del usuario  Bedienungsanleitung franz-366sisch 20l 07-2007  LWG/LWX Gebrauchsanleitung LWG/LWX User Manual  User Manual and Spare Parts Catalogue Notice d  AT 2050(47.7 cm3) - AT 2080(72.4 cm3)  PHD-4 Portable Helium Detector  Whirlpool ET8WTMXKT00 User's Manual  Owners Manual - LS120P (French)  

Copyright © All rights reserved. Failed to retrieve file