Solaris virtualization options
Contents
1. First you need to choose your VM name and its OS type then assign the memory size to the VM which for normal Solaris installation requires at least 768 MB Next you need to specify the VM disk image If it is your first time to create a VirtualBox File image you have to click the NEW but ton You must also specify a file for the disk image You can either specify LOGIN VOL 33 NO 5 LOGIN OCTOBER 2008 the size of the disk file or set it to grow dynamically over time A standard Solaris SXCE requires 12 GB for the image file If you want to further con figure the VM s devices you can click the Settings button One important note is that so far VirtualBox only supports NAT network topology If you want to install from the CD choose the CD ROM option there you can click to mount the CD drive from the host CD ROM or from a specific ISO image Once done you can simply click the Start button to start the VM The VM will boot from CD ROM and the rest of the installation process is the same as normal Solaris installation When the virtual machine is created and booting VirtualBox will prompt you with a VNC window to display the screen of the guest VM you can either click the mouse or press the Enter key to get into the guest VM If you want to get out of the VM box you can press the right Ctrl key to release yourself from the VM In VirtualBox if you want to run a previously installed VM you can operate on2. Time category FIGURE 1 PERFORMANCE COMPARISON BY PERCENTAGE AGAINST BASE SOLARIS SYSTEM COMPILE OF APACHE In Figure 2 we find that the container s sys and user time are almost the same as the zone s But overall the container s total consumed time is al most doubled which indicates that Solaris resource management is ef fectively giving the container a limited share of system resources We recommend that system administrators use resource management facilities to avoid some zones malicious or greedy resource usage and effect on the SOLARIS VIRTUALIZATION OPTIONS 15 overall performance of other zones Although the configurations may be a little complicated it is worth taking the time to get it right One successful experiment involves running a memory bomb a loop constantly allocating and touching additional memory in the zone and container The zone leaves the global zone dead but although the container suffers from running out of memory the global zone is still alive and works well For more details refer to our previous paper 2 The Performace Comparsion of the Zones w and w o Resource Limits 120 100 80 60 Wizone container 40 m 20 real user Time categrory w E E w 3 a w 2 2 E oa n w 2 oa E z w pe w g c w g 2 w a FIGURE 2 COMPARING A RESOURCE LIMITED CONTAINER TO A ZONE Conclusion Overall Solaris has offered us a variety of virtu
3. 1 capped memory physical 1G swap 1G locked 768M retl name zone cpu cap value priv privileged limit 50 action deny name zone max swap value priv privileged limit 1073741824 action deny name zone max locked memory value priv privileged limit 805306368 action deny name zone max lwps value priv privileged limit 200 action deny LISTING 8 THE RESOURCE LIMITATION FOR THE BENCHMARKED CONTAINER Figure 1 shows the relative overhead of the four virtual machines by a per centage of the baseline time to compile httpd Overall the zone has the least overhead compared to the baseline because it has full access to the whole global zone s resources Its performance is almost as good as the baseline The container experiences delay because it is limited to half of the overall CPU cycles The Xen guest domain is close to the overhead of the container but it consumes substantially more system time VirtualBox clearly has the highest overhead 250 Note that in our experiment on the same hard ware the overhead of Xen on Linux is less In general the overhead of Xen on Solaris is not necessarily the same as Xen on Linux The Performance Comparision by Compiling Httpd Test 300 250 200 Wizone Wi container Xen domU W vBox vm 0 ee real user sys 150 Oo Oo nn Oo w ao w 2 2 w n w T Z w pea o g c w 8 2 w a
4. Containers are an operating system level virtualization facility meaning that the OS itself provides the illusion of multiple independent systems each with its own IP address and file system all based on the same base kernel More recently support for paravirtu alization in the form of Xen called Sun xVM on Solaris has been added and now with the acquisi tion of VirtualBox full virtualization on Solaris is also an option Unlike OS level virtualization par avirtualization and full virtualization both offer the ability to run guest operating systems that are dif ferent from the underlying OS Full virtualization can run unmodified operating systems whereas paravirtualization requires targeted changes to the hardware interface and therefore correspond ing changes in the OS source code As a result proprietary operating systems such as Microsoft Windows can typically only be run on virtualiza tion systems that support full virtualization Some virtualization systems such as Xen require hard ware support for virtualization such as Intel VT or AMD V to support full virtualization Gett ing Started with Solaris Containers In Solaris Containers Zones a virtual machine is called a zone and a zone with resource limitations is called a container The basic command to operate a zone s configuration file is zonecfg z newzone This will bring up a shell in which you can issue a variety of commands to manipulate the spe
5. network installation for paravirtual guests but both ISO and network instal SOLARIS VIRTUALIZATION OPTIONS 11 12 lations for HVM guests Here we give an example for creating a paravirtual CentOS guest Virt manager first asks for the guest domain s name which later becomes the guest domain id and the type of the guest domain paravirtualized or fully virtualized Then it will request a network path address to the CentOS re pository such as http mirror clarkson edu pub centos 5 os i386 Later you will be asked to choose a disk partition or a file to be the guest file system If you have not set up a special disk partition for your new guest a file is the safest choice After allocating the memory size and the number of virtual CPUs it will go through the normal CentOS network installation process Our Web site 1 includes detailed screenshots of the entire process Getting Started with Sun xVM VirtualBox Since VirtualBox was acquired by Solaris only recently it is not yet automat ically installed in Solaris So the first step is to download VirtualBox from http virtualbox org wiki Downloads To determine which Solaris package to download 32 bit x86 or 64 bit AMD64 you can use the isainfo com mand on your base Solaris system The VirtualBox installation package includes two packages the kernel pack age used to install the Virtual Disk Image VDI kernel module and the VirtualBox package which will install t
6. remote access to the guest at all you can use the command svccfg s xvm xend delprop config vnc listen to remove that option and then refresh and restart Xend as we did earlier svccfg s xvm xend setprop config vnc listen astring 0 0 0 0 svccfg s xvm xend setprop config vncpasswd astring passwd LISTING 6 VNC SETUP IN XEND Next we can use virt install with the hvm argument to start an HVM So laris guest as for example virt install n SolarisHVM hvm r 1024 vnc f export home SolarisHVM img s 30 c export home newbie sol nv b87 iso Notice that we use almost the same options as we did when creating the paravirtual Solaris guest the only differences are hvm which specifies the guest works in HVM mode and vnc which specifies that the guest will have a VNC connection Also the argument c is used for specifying the virtual CD ROM rather than when installing the paravirtual guest In this case it indicates that the installation is from the ISO file export home sol ny b87 iso and from the guest s perspective the installation is from the guest CD ROM We could install guests based on any other operating sys tem including Windows in this same manner When you run the virt install command it will first display a window ask ing for a VNC password which is passwd as we have set up in Xend at the beginning of the HVM guest setup procedure The rest of the installation process will look exactl
7. B of memory and 512 MB of swap space It also guarantees that 64 MB of the zone s memory will be resident in physical memory at all times Finally the set max lwps command illustrates how we can place limits on things besides physical resources It limits the number of light weight processes in a running zone and is useful for preventing problems such as fork bombs from taking down the whole machine add capped cpu set ncpus 0 25 end add capped memory set physical 128M set swap 512M set locked 64M end set max lwps 175 LISTING 2 THE OPTIONS LIMITING THE CONTAINER S RESOURCES Once our new zone is configured we are ready to instantiate it with the command zoneadm z newzone install and then run it with the com mand zoneadm z newzone boot zoneadm can also be used for other zone administration functions such as listing the configured zones install ing or uninstalling the zones and booting pausing and halting installed zones Listing 3 shows the output of running zoneadm list after installing our new zone The parameter c will display all the configured zones infor mation v will display the detailed information of zones zoneadm list vc ID NAME STATUS PATH BRAND IP 0 global running native shared 1 newzone running export home newzone native shared LISTING 3 A ZONEADM LIST SHOWING ALL ZONES LOGIN VOL 33 NO 5 LOGIN OCTOBER 2008 Finally we are ready to log in to our new zone with the command zlogi
8. HE STEPS FOR CONFIGURING AND RESTARTING XEND With that we are ready to create a new guest domain In this section we will show three primary examples a PV Solaris guest domain a PV Linux domain and an HVM Solaris domain In Solaris virt install is a tool used to create the guest domain images regardless of whether it is PV or HVM For example a paravirtualized Solaris image can be created with the following command virt install nographics n newSolaris paravirt f export home newSolaris img r 1024 s 30 I export home sol nv b87 iso where n is for specifying a domain name to be newSolaris paravirt is for selecting the mode to be paravirtualized f is for specifying the domain image name newSolaris img in the path export home r is for assigning the domain memory size to be 1024 MB s is for creating the domain image size in gigabytes and I is for choosing the installation location Note that in Solaris xVM for paravirtual guests both Solaris and Linux the video card and CD ROM drivers are not yet fully ported will not have a graphical win dow and cannot use fully use the CD ROM However we were able to use an ISO file as the guest CD ROM during the install of the paravirtual Solaris guest by using the I option to specify the ISO location But for the HVM guest the guest can fully use a standard CD ROM driver within it and then have a CD ROM device We can also install the guest via a Solaris NFS share on Sola
9. Similar to VirtualBox in Solaris xVM each guest s file system is independently separated either by files or partitions or disks Theoretically we can dynamically add Domain0 s disk or partitions to the guest domain sharing with the guest domain But we do not recom mend attempting that because there is no way to maintain the consistency of shared files or file systems The preferred method of sharing files with xVM guests would be to use a network file server For the network topology in Solaris Zones all zones share the network in terface with the global zone s network interface in a bridged mode There are no other network topology choices Bridging is also currently the only op tion available for Solaris xVM guest domains The routing and NAT topolo gies that are available to Xen on Linux are still in development for Solaris VirtualBox supports only the NAT topology This means that there is no way for an outsider to directly access the VirtualBox VM through the network This is a crucial difference for running server VMs It is worth noting that with the Crossbow project changes in networking support should be coming for all virtualization systems Crossbow is a So laris network virtualization and flow control solution It provides universal network architecture to the virtualization systems described here Contain ers xVM and VirtualBox to manage the flow control of those virtual NICs such as bandwidth and packet types Cro
10. WENJIN HU TODD DESHANE AND JEANNA MATTHEWS Solaris virtualization options Wenjin Hu is a PhD student in Computer Science at Clarkson University He focuses on applying virtualization and file system techniques to provide security and reliability on the desktop huwj clarkson edu Todd Deshane is a PhD student in Engineering Science at Clarkson University His current research deals with using virtualization to provide security on the desktop deshantm clarkson edu Jeanna Neefe Matthews is an associate profes sor of Computer Science at Clarkson University in Potsdam NY where she leads an incredible team of students She is currently on sabbatical and work ing with VMware in Cambridge MA jnm clarkson edu LOGIN OCTOBER 2008 THE VIRTUALIZATION OPTIONS FOR Solaris have been expanding rapidly In this article we discuss three types of virtualiza tion systems available in OpenSolaris for x86 Solaris Containers Solaris xVM and Sun xVM VirtualBox We include instruc tions on how to deploy each solution and a detailed comparison of the three to help system administrators and virtualization fans alike choose the appropriate virtualiza tion technology for their needs Even if you don t use Solaris we do explain the differ ences among OS level virtualization para virtualization and full virtualization clearly Solaris has included Containers also called Zones since Solaris 10 was released in 2005
11. alization systems to use So laris Containers Solaris xVM and Sun xVM VirtualBox Each of these has its own unique advantages Sun xVM VirtualBox offers full virtualization is straightforward to use and has nice GUI windows but its performance overhead is also high and with an NAT only network running servers is difficult Still for easy to use desktop virtualization on Solaris VirtualBox is probably the best choice In contrast Solaris Containers Zones OS level virtualization is targeted at server level usage It achieves good performance but to make it work properly you need to master the resource management control tools which can be somewhat complicated Containers Zones also do not give you a choice of guest operating systems For fast Solaris servers they are likely the best choice However if you want a choice of guest op erating systems and good performance then Solaris xVM is likely to be the best choice Its performance is comparable with OS level virtualization and it is suitable for both desktop usage and server usage Solaris xVM can be a bit complicated to configure but there are a variety of configuration options from GUI to command line and the available tools continue to improve REFERENCES 1 Clarkson Web site for screenshots and documents http www clarkson edu projects virtualization solaris login08 2 Quantifying the Performance Isolation Properties of Virtualization Systems http people clarkson edu j
12. cified zone As shown in Listing 1 you create a zone add the attached devices such as add net set zone options such as set autoboot true display the configuration verify the configuration and finally commit a zone s resources which writes out a SOLARIS VIRTUALIZATION OPTIONS 7 final configuration file for the zone You can also use zonecfg to browse the characteristics of an existing zone and modify it as desired zonecfg z newzone newzone No such zone configured Use create to begin configuring a new zone zonecfg newzone gt create zonecfg newzone gt set zonepath export home newzone zonecfg newzone gt set autoboot true zonecfg newzone gt add net zonecfg newzone net gt set address 192 168 16 109 zonecfg newzone net gt set physical pcnO zonecfg newzone net gt end zonecfg newzone gt verify zonecfg newzone gt commit zonecfg newzone gt exit LISTING 1 THE STEPS SHOWING THE ZONE CONFIGURATION Solaris provides a number of options to manage the resources that a zone can own including the CPU the memory and the number of the pro cesses This is the heart of making a zone into a container with resource limitations In Listing 2 we illustrate how to add a variety of restrictions to our new zone These commands are also issued at the zonecfg prompt The capped cpu command limits the CPU cycles assigned to the zone to a fourth of one CPU Similarly our capped memory command assigns the zone 128 M
13. he VirtualBox application and GUI library Once the proper package is downloaded use the command pkgadd to first install the VirtualBox kernel package and then the VirtualBox pack age as illustrated in Listing 7 pkgadd G d VirtualBoxKern VERSION OS BIT pkg pkgadd G d VirtualBox VERSION OS BIT pkg VirtualBox kernel module unloaded VirtualBox kernel module loaded Creating links Done Installation of lt SUNWvbox gt was successful LISTING 7 THE STEPS FOR INSTALLING VIRTUALBOX Once the packages are installed simply issue the command VirtualBox so that the VirtualBox management window will pop up One important note for trying all the virtualization systems available on Solaris is that VirtualBox and Solaris xVM cannot currently operate at the same time If you have been running in Solaris xVM it is necessary to re boot the machine and switch to the first option Solaris in the GRUB menu Otherwise you will see the error VirtualBox Kernel Driver not Loaded when you do run the VirtualBox command because the VirtualBox kernel has not yet been ported to the Xen kernel To create a virtual machine when you choose the NEW button it will lead you to the VM installation wizard VirtualBox can run any guest operating system including Windows Linux and Solaris or any x86 OS running on it without any modifications to your guest OS All of their device drivers will work normally with no need to port to the guest OS
14. n newzone It will be running the same version of OpenSolaris as the base operating system as is required with OS level virtualization However the zone will behave like a separate system with its own IP address its own file system and its own set of installed applications Getting Started with Solaris xVM A virtual machine in xVM is called a domain just as it is in other Xen im plementations Domains are divided into two major categories paravirtual ized PV domains and Hardware assisted Virtual Machine HVM domains In a PV domain the hardware abstraction presented to the guest VM is not identical to the underlying physical hardware instead strategic changes are made to make the system easier to virtualize To run properly on the para virtualized hardware the guest operating system must be modified to be aware of these changes This requires source code level changes to the OS and is therefore rarely available for proprietary operating systems such as Windows Two common choices for paravirtualized guest domains are So laris and Linux In an HVM domain the hardware abstraction presented to the guest is iden tical to the underlying hardware so any operating system that runs on x86 hardware architecture can be installed To make this possible an HVM do main relies on special hardware support for virtualization such as Intel V T or AMD V If your system does not have this hardware support for virtual ization then paravi
15. nm publications isolation_ExpCS _FINALSUBMISSION pdf LOGIN VOL 33 NO 5 PRACTICAL RESOURCES 1 Wenjin s blog on Solaris Virtualization Tutorials http deepenintocs blogspot com 2 Xen introduction and tutorial http runningxen com 3 Solaris Containers online documentation http www sun com bigadmin content zones 4 Solaris xVM http opensolaris org os community xen 5 VirtualBox user manual http www virtualbox org wiki End user _documentation 6 Crossbow project http opensolaris org os project crossbow LOGIN OCTOBER 2008 SOLARIS VIRTUALIZATION OPTIONS 17
16. nt tools such as virt manager are extended and integrated into Solaris the ease of use will improve Of course ease of use is just one part of the story It is also important to consider the features of each system Containers can only run Solaris guests so some common applications of virtualization e g running alternate op erating systems simply won t work in Containers Solaris xVM requires the running guest to be a modified OS and generic virtual device drivers need to be ported If the unmodified OS is to run on Solaris xVM VT or AMD V hardware support will be needed on the CPU chip VirtualBox however can run any type of unmodified guest even without hardware support for virtu alization In terms of storage Solaris Zones can either share files with the global zone or have their own version of files from the global zone In other words zones can use the same library files as the global zone or have older or newer versions of libraries than the global zone You should also be aware that when you change the files in the global zone you may also affect other zones that are sharing them For VirtualBox the system files in a VM of Vir SOLARIS VIRTUALIZATION OPTIONS 13 14 tualBox are only used by that VM and will not affect other VMs Each VM is encapsulated in a separate vdi or vdmk files However if sharing is desired VirtualBox does have a shared folder option that can mount a base OS di rectory to share with the guest
17. ris xVM If we use ZFS volumes for the guest disk storage it should have better perfor mance and more reliability than using a file based image in the DomainO file system After running this command and finishing the normal Solaris Installation process you can use virsh list again to see the newSolaris guest domain running as illustrated in Listing 4 To access the running domain newSo laris we can use the command virsh console newSolaris To get out of the guest domain the combination key Ctrl is needed If we wish to save the guest domain s configuration file for later use we can use the virsh tool to write an XML format configuration file when the guest domain is running with virsh dumpxml newSolaris gt newSolaris xml With the guest domain s xml configuration file we can directly boot the LOGIN VOL 33 NO 5 LOGIN OCTOBER 2008 image with the command virsh create newSolaris xml Furthermore we can use the command virsh shutdown newSolaris to turn off the guest Next we will walk through an example of installing an HVM guest domain Before creating the Solaris HVM image we may need to enable VNC in Xend as Listing 6 shows In the svccfg command shown 0 0 0 0 indicates that the VNC server is listening to any IP and passwd should be replaced with the actual password If you don t explicitly set the password it will de fault to the empty string which is of course not secure If you don t want to use VNC for
18. rs on the DomainO drivers but the guest driver is unaware of this so the hypervisor must trap accesses to I O space or DMA operations in order to forward them on to the proper device driver in DomainO PV drivers have much lower overhead be cause they avoid this expensive process of trapping and forwarding Our first step in running Xen guests is to make sure that the hypervisor Domain0 and Xend are all running After Solaris xVM boots up you can use the command virsh list as shown in Listing 4 to check whether DomainO is running SOLARIS VIRTUALIZATION OPTIONS 9 10 virsh list 0 Domain 0 running 110 newSolaris blocked LISTING 4 VIRSH LIST SHOWING DOMAINO Next we check that Xend is running and specify the default network in terface in Xend first so that our guest domains are able to set up the virtual network based on that physical NIC Xend is wrapped as an application ser vice in Solaris Listing 5 illustrates the use of svccfg and svcadm facilities to configure and restart Xend The s parameter specifies the Xend service setprop specifies the Xend configuration options After making this or any change to the Xend configuration you can apply the change by refreshing and restarting Xend as shown The svcadm facility can be used to enable disable refresh or restart the Xend service at any time svccfg s xvm xend setprop config default nic bge0 svcadm refresh xvm xend svcadm restart xvm xend LISTING 5 T
19. rtualized domains are your only option Xen is a virtual machine monitor called a hypervisor which intercepts the guest domain s system calls It is necessary to boot the Xen enabled Solaris kernel on the physical machine rather than a normal Solaris kernel Since Solaris Nevada Version build 75 Xen has been developed and well inte grated into Solaris Express Community Edition through a variety of boot options in the GRUB menu For example the standard GRUB menu displays three choices Solaris Express Community Edition snv87 X86 Solaris xVM and Solaris Failsafe the second grub option Solaris xVM should be chosen Xen also relies on a special privileged domain called Domain and the Xen control daemon Xend for communication between the hypervisor and the guests Domain0 is granted the full privileges of managing the guest do mains and the physical devices by the Xen hypervisor similar to a normal Solaris OS instance Device drivers can also be fully virtualized or paravirtualized Even a system that does full virtualization of the CPU and memory can load paravirtual ized drivers to handle external devices such as a disk or network interface In a paravirtualized driver the driver running in the guest operating system is aware of the hypervisor and explicitly participates in communicating its requests to domainO where the real physical device drivers are running In a fully virtualized driver the real device access still occu
20. ssbow is not yet stable but is being tested as part of SNV build 91 Finally a critical aspect of the comparison is performance overhead from virtualization A full performance comparison on various types of hardware and running a wide variety of tests is beyond the scope of this article Here we present the results of some simple compilation tests on baseline Solaris and on each of the virtualization systems Specifically we report the time to compile the Apache Web server All our tests are run on Open Solaris Community Express Nevada build 87 SNV b87 running on a Dell Optiplex Intel VT dual core 2 4 GHz 6600 CPU with 4 GB memory a 250 GB disk and 1 Gb NIC The guest resource allocation can be seen in Table 1 Zone is a virtual machine in Solaris Xen domU is the virtual machine in Solaris xVM and VBox VM is the virtual machine running in VirtualBox Container is a zone with resource controls Here the container is assigned the limited CPU to be 1 ncpus If we use zonecfg z newzone info we can see the information in Listing 8 For more complicated resource configu ration you can look at the usage of project and task facilities CPU dual Memory Image size Network Zone Bridged Container l 1024M Bridged Xen domU l 1024M 30G Bridged Vbox VM l 1024M 30G NAT TABLE 1 VM RESOURCE ALLOCATIONS FOR EACH VIRTUALIZATION SYSTEM LOGIN VOL 33 NO 5 LOGIN OCTOBER 2008 zonecfg z newzone info capped cpu ncpus
21. the existing vdi files By default the VM virtual disk images are stored in the root VirtualBox VDI directory There is a virtual disk manager to manage them You can press the combination key Ctrl D to pop up the disk manager window If you already have a system installed vdi file you can press the Add button to add the existing VM image into the VirtualBox Then you can go through the previous process of creating a new VM proce dure to run a VM But if you want to remove a VM image from VirtualBox you need to first Release the image and then Remove it from the virtual disk manager because when a disk image is assigned to a VM VirtualBox automatically registers it and grants a unique uuid to that VM and image Comparing Containers xVM and VirtualBox Now that we have shown you the basics of getting started with three differ ent virtualization options on Solaris in this section we will present some comparisons among them One important point of comparison is ease of use In our opinion the easi est to use is VirtualBox It is fully GUI guided straightforward and simple Solaris Containers are also relatively easy to use especially because they are so well integrated into Solaris and have such a complete tool chain for con figuration and management However since they have no GUI interface they are better suited to server applications than desktop virtualization So laris xVM is the most complicated but as more manageme
22. y like installing Solaris on a real machine As with the PV guest we can use the virsh facility to create and shut down the guest domain We used virt install to create Solaris guests and it s also possible to use it to create Linux guests in a similar way However for Linux guests there are some additional options In this section we will illustrate the use of one of these virt manager to install a Linux guest Refer to our Web site 1 for additional information on options for Linux guests such as netinstall isoin stall and cdrom install Virt manager is a GUI installation and management tool for guest domains It provides a GUI tool for the creation of new domains an integrated VNC viewer for accessing domains and other useful tools for the management of domains and their resources Virt manager is available on newer releases of CentOS and Red Hat Enterprise It is also available in the latest version of Solaris now You can find its icon from menu gt All Applications gt System Tools gt Virtual Machine Manager You can also run it as root with the com mand virt manager It will first display a window asking you to connect to Xen hypervisor Single clicking the Connect button will show you the Vir tual Machine Manager GUI interface displaying the Domain0O and guest do main s running status and resource usage Click the New button and it will guide you step by step through the guest creation wizard It only supports
Download Pdf Manuals
Related Search