This How to Guide instructs users, system administrators, and developers unfamiliar with Solaris 10 OS on how to consolidate applications onto a single server using Solaris Containers technology. The guide starts with a brief overview of Solaris Containers and follows with an example of using Solaris Containers to consolidate two Web server applications and an email server application onto a single server.
Users are guided step-by-step through the consolidation process, with code examples and illustrations. After using this guide, a user should be able to create Solaris Containers by: Creating a resource pool Defining Solaris Zones Assigning CPU usage with the Fair Share Scheduler (FSS) Installing and booting a zone Configuring access to raw devices from the zone PDF[514K] Table of Contents 1. Solaris Containers: Overview 2. Solaris Containers: An Example
Solaris Containers:
Overview Solaris Containers—Sun's operating system virtualization technique—consists of several technologies that work together to foster improved resource management and isolate the environment from the underlying OS. With Solaris Containers, organizations can determine how to allocate resources among applications and services, and ensure they do not interfere with one another. This guide uses an example in which an organization wants to consolidate three applications—an email server application and two Web servers—onto a single server with four CPUs.
Solaris Zones As part of the consolidation effort, it is important to evaluate the needs of the applications that will share the consolidated system. In this example, the email server and Web server applications need to run in isolated environments that make it appear as if they are running on physically separate machines. This is made possible by a Solaris Container technology called Solaris Zones, which provides separate environments on a machine and logically isolates applications from one another. Each application receives a dedicated namespace in which to run, and cannot see, monitor, or affect applications running in another zone. See Figure 2.
Dynamic Resource Pools In this example there are two types of applications, one that needs a fully dedicated CPU—an email server, and another that is more flexible and can share CPUs—the two Web servers. To accomplish these different levels of isolation we use a Solaris Container technology called Dynamic Resource Poolsthat enables CPU resources to be dedicated to specific applications. In this example, the email server needs a separate resource pool, while the Web servers can share another. As can be seen in Figure 3, we give one CPU to resource pool 1 and three CPUs to resource pool 2.
Fair Share Scheduler While the two Web servers are capable of sharing the remaining CPUs on the system, they each need a minimum guarantee of CPU resources that will be available to them. This is made possible by another Solaris Container technology called the Fair Share Scheduler(FSS). This software enables CPU resources to be allocated proportionally to applications. That is, each application gets assigned a number of the available shares of the total CPU. See Figure 4. Using all these Solaris Container technologies on this new server creates an environment like that shown in Figure 5. Now the system includes three Containers: one with dedicated CPU resources, and two others sharing CPU resources.
Two Types of Zones All systems that run Solaris 10 contain a master zone, called the global zone. The global zone is the original Solaris OS instance. It has access to the physical hardware and can control all processes. It also has the authority to create and control new zones, called non-global zones, in which applications run. Non-global zones do not run inside the global zone—they run along side it—yet the global zone can look inside non-global zones to see how they are configured, monitor, and control them. Like the other zones, this global zone is also associated with a resource pool. In our example this is resource pool 2. In this case, it shares CPU resources with the Web servers. When the Fair Share Scheduler is enabled, the global zone defaults to one share. So Figure 5 needs to be amended to include the global zone as shown in Figure 6. Back To Top
Solaris Containers: An Example This section is a step-by-step guide that shows how to create the email and Web server example described above. It does this by going through the following steps: Creating a new resource pool Creating the email zone on the new resource pool Enabling the Fair Share Scheduling on the Web server resource pool Creating the first Web server Container Creating the second Web server Container When creating the last two Containers we will also show some extra configuration options available to Solaris Zones. As the starting point, Figure 7 shows how the system looks before any of these steps are taken. The only existing objects are the global zone, and one resource pool. The global zone is associated with this one resource pool. This first resource pool is also called the default pool, and before any action is taken, it has all the CPUs in the system.
Creating a New Resource Pool In the Solaris OS, a resource pool is a logical entity that owns a subset of the system resources, like CPU and Memory. These subsets are known as resource sets. Currently the Solaris OS only has one type of resource set—a processor set. Every resource pool is associated with a processor set—so if you want to give this pool its own unique CPUs, you will need to define the processor set, the number of processors it contains, and associate it with a pool. Every additional resource pool created is taken out of the default pool. So, the email server's CPU comes out of the four CPUs initially assigned to the default pool (leaving three). Resource pools can be dynamically created and removed on a running system, but you must leave one CPU for the default pool.
To create a new resource pool: 1. Enable the resource pools feature, using the pooladm(1M) command.
global# pooladm -e
The use of the "global#" prompt throughout this example indicates that a command is run in the global zone, this is where you'll be by default. 2. Save the current configuration to a file by using the pooladm(1M) command.
global# pooladm -s 3. See if any pools already exist on the system, using the pooladm(1M) command.
global# pooladm
system my_system
string system.comment
int system.version 1
boolean system.bind-default true
int system.poold.pid 638
pool pool_default
int pool.sys_id 0
boolean pool.active true
boolean pool.default true
int pool.importance 1
string pool.comment
pset pset_default
pset pset_default
int pset.sys_id -1
boolean pset.default true
uint pset.min 1
uint pset.max 65536
string pset.units population
uint pset.load 7
uint pset.size 8
string pset.comment
cpu
int cpu.sys_id 1
string cpu.comment
string cpu.status on-line
cpu
int cpu.sys_id 0
string cpu.comment
string cpu.status on-line
cpu
int cpu.sys_id 3
string cpu.comment
string cpu.status on-line
cpu
int cpu.sys_id 2
string cpu.comment
string cpu.status on-line
You are starting on a system on which no pools were configured, you should only see one pool entry, "pool pool_default" 4. Create a processor set containing one CPU, using the poolcfg(1M) command.
global# poolcfg -c 'create pset email-pset (uint pset.min=1; uint pset.max=1)'
This command says change the pool configuration to create a processor set (pset) called email-pset with a minimum of 1 CPU and maximum of 1 CPU. 5. Create a resource pool for the processor set.
global# poolcfg -c 'create pool email-pool' 6. Link the pool to the processor set. This resource pool will be used by the email server.
global# poolcfg -c 'associate pool email-pool (pset email-pset)' 7. Activate the configuration.
global# pooladm -c 8. Verify the existence of the resource pool, using pooladm(1M).
global# pooladm
system my_system
string system.comment
int system.version 1
boolean system.bind-default true
int system.poold.pid 638
pool email-pool
int pool.sys_id 1
boolean pool.active true
boolean pool.default false
int pool.importance 1
string pool.comment
pset email
pool pool_default
int pool.sys_id 0
boolean pool.active true
boolean pool.default true
int pool.importance 1
string pool.comment
pset pset_default
pset email-pset
int pset.sys_id 1
boolean pset.default false
uint pset.min 1
uint pset.max 1
string pset.units population
uint pset.load 0
uint pset.size 1
string pset.comment
cpu
int cpu.sys_id 0
string cpu.comment
string cpu.status on-line
pset pset_default
int pset.sys_id -1
boolean pset.default true
uint pset.min 1
uint pset.max 65536
string pset.units population
uint pset.load 7
uint pset.size 7
string pset.comment
cpu
int cpu.sys_id 1
string cpu.comment
string cpu.status on-line
cpu
int cpu.sys_id 3
string cpu.comment
string cpu.status on-line
cpu
int cpu.sys_id 2
string cpu.comment
string cpu.status on-line Note "pool email-pool" and "pset email-pset" sections in the output. The "pset email-pset" section also displays that there is one CPU allocated to this pool "pset.size 1"
Creating the Email Zone on the New Resource Pool Once the processor sets and resource pools are created, Solaris Zones technology can be used to create isolated environments for the email server and Web server applications.
Creating a zone requires the following steps: Configuration—Define the zone properties, such as the required file systems and network interfaces Installation—Create the zone on the system by installing and populating the part of the file system hierarchy reserved for the zone Virtual platform management—Use zone tools to boot, halt, or reboot the zone Zone login—Move in and out of the zone to perform administrative tasks The zone we're creating for the email server is a very standard zone—the other two zones will show more options. The following steps outline the process for creating a container with a zone on the new resource pool. This zone will be called email-zone, and will have the IPv4 address 10.0.0.1.
Configuration
In order to configure networking parameters in the following section, you will need: a. an IP address b. the name of the network interface
Note: A quick way of finding a physical network interface on the system is by using the "ifconfig -a" command in the global zone. c. You should also confirm that the file system you use has sufficient disk space to accommodate the zone.
Note: A default configuration requires approximately 100 MB of free disk space plus the amount of space needed by applications. To configure and define a new zone: 1. Enter the zone configuration tool by using the zonecfg(1M) command.
global# zonecfg -z email-zone
This will return the message No such zone configured before prompting you to begin configuring a new zone, you are now in the zonecfg shell that is identified by its prompt:"zonecfg:email-zone>". 2. Create a new zone definition with the create command.
zonecfg:email-zone> create 3. Assign the zone to a file system, using the set zonepath command.
zonecfg:email-zone> set zonepath=/export/home/zones/email-zone
Note: It is important to select a file system with sufficient disk space to accommodate the zone. A default configuration requires approximately 100 MB of free disk space plus the amount of space needed by applications. 4. Decide if the zone should boot automatically at system boot time. Use the set autoboot command to specify this behavior.
zonecfg:my-zone> set autoboot=true
"True" indicates the zone should be booted at system boot time, while "no" indicates it should not be booted automatically. 5. Configure networking parameters, using the add net command and its subcommands.
zonecfg:email-zone> add net
zonecfg:email-zone:net> set address=10.0.0.1
zonecfg:email-zone:net> set physical=eri0
zonecfg:email-zone:net> end
In this example, the virtual network interface with IP address 10.0.0.1 is configured on the eri0 interface. 6. Assign the zone to the email pool.
zonecfg:email-zone> set pool=email-pool 7. Verify the configuration is syntactically correct, using the verify command.
zonecfg:email-zone> verify
This will only return a message if any errors in the configuration were found. 8. Write the in-memory configuration to stable memory, using the commit command, and then exit the shell
zonecfg:email-zone> commit zonecfg:email-zone> exit (or ^D [Ctrl d])
A standard zone automatically shares the /usr, /lib, /platform, and /sbin file systems with the global zone. It is important to note that a standard zone configuration mounts all global file systems as read-only. As a result, an attempt to install an application to any of these directories will fail. See the section Creating the First Web Server Container(page 9) to learn how to mount a global zone file system with write permissions in the directory in which the application is installed.
Installation
The verify command in the zone configuration process is only concerned with the syntactic correctness of the configuration. It determines if the configuration could be created on a system, but not necessarily this particular system. The true verification for the target system is performed automatically when the zone is installed. The zoneadm(1M) command checks to see if all resources, such as the physical network interface specified in the configuration, are available. It then installs the files needed for the zone's root file system in the correct location under the zonepath, and creates the mount points for additional file systems specified in the configuration. Install the zone, using the zoneadm(1M) command.
global# zoneadm -z email-zone install
Preparing to install zone email-zone
Creating list of files to copy from the global zone.
[Some output was omitted here for brevity]
Zone email-zone is initialized.
This can take a few minutes.
Virtual Platform Management When the installation is complete, the zone is ready to be booted. While the zone is now installed, the system identification internal to the zone has not run. At this point the administrator can configure things like the root password to the zone and the name server with which it should connect. The first time the zone is booted, the system automatically interacts with the user to configure this system identification. Get the zone ready to run: 10. Boot the zone, using the zoneadm(1M) boot command.
global# zoneadm -z email-zone boot
Since this is the first time the zone is booted after installation, standard system identification questions must be answered. You must log in to the zone's console to answer these questions.
Zone Login 11. Log on to the zone Console using the zlogin(1M) command. Once the system boots, the normal system identification process for a newly installed Solaris OS instance is started. The output of this process is omitted here for brevity, but the configuration questions concerning the name service, time zone, and other system parameters should be answered as appropriate for the site.
global# zlogin -C email-zone
[Connected to zone email-zone console]
[This will now show the same type of output as when a normal system boots.
You will also be asked for the regular system identification questions.
Here are some suggestions]
Terminal=(12)X Terminal Emulator (xterms)
Hostname for eri0:1 = email-zone
No Kerberos
Name service = None
Time Zone = my-time-zone
root passwd = (Your choice)
After system identification is complete and the root password is set, the zone will reboot and is then ready for use. 12. Disconnect from the console using ~. (tilde dot) as in tip(1). The zone can now be accessed over the network using the telnet(1), rlogin(1) or ssh(1) commands, like a standard Solaris OS system. Figure 9 now shows the current state of the system with the new zone and its resource pool. It also shows the remaining CPUs in the default pool with the global zone. Note that each zone created on the system must be installed, configured, and booted. In addition, a sysidcfg(4) file can be used to automate the identification process. See the http://docs.sun.com site for details.
Enabling the Fair Share Scheduling on the Web Server Resource Pool Once the Container for the email server application is created, installed, and booted, you will create another Container for the first Web server. While this new Container is similar to the one created for the email server application, it also utilizes the Fair Share Scheduler to set CPU usage guarantees. To set the Fair Share Scheduler: 1. Set the scheduler for the default pool to the Fair Share Scheduler, use the poolcfg(1M) command.
global# poolcfg -c 'modify pool pool_default (string pool.scheduler="FSS")' 2. Create an instance of the configuration, using the pooladm(1M) command.
global# pooladm -c 3. Move all the processes in the default pool and its associated zones under the FSS.
global# priocntl -s -c FSS -i class TS
global# priocntl -s -c FSS -i pid 1
If you don't want to reboot the system you can use priocntl(1). This step could also be done by rebooting the system.
Creating the First Web Server Container Installing this zone will be slightly more sophisticated. You will assign three Fair Share shares to it as well as provide readwrite access to the /usr/local file system. To create the zone: 1. Define the zone for the first Web server using the same process used to define and install the email-zone zone (Page 6). Be sure to change the name of the zone, its location, the name of the pool used, and the IP address.
global# zonecfg -z Web1-zone
Web1-zone: No such zone configured
Use 'create' to begin configuring a new zone.
zonecfg:Web1-zone> create
zonecfg:Web1-zone> set zonepath=/export/home/zones/Web1-zone
zonecfg:Web1-zone:net> set address=10.0.0.2
zonecfg:Web1-zone:net> set physical=eri0
zonecfg:Web1-zone:net> end
zonecfg:Web1-zone> set pool=pool_default Remember,
the two Web servers share the CPU resources of the default pool with each other as well as the global zone, so you need to specify how those resources should be shared using the Fair Share Scheduler (FSS). With FSS, the relative importance of applications is expressed by allocating CPU resources based on shares—a portion of the system's CPU resources assigned to an application. The larger the number of shares assigned to an application, the more CPU resources it receives from the FSS software relative to other applications. The number of shares an application receives is not absolute—what is important is how many shares it has relative to other applications, and whether they will compete with it for CPU resources. 2. Assign three shares to this zone, using the following commands.
zonecfg:Web1-zone> add rctl
zonecfg:Web1-zone:rctl> set name=zone.cpu-shares
zonecfg:Web1-zone:rctl> add value (priv=privileged,limit=3,action=none)
zonecfg:Web1-zone:rctl> end
zonecfg:Web1-zone> exit In the case of a standard zone install—like the email server—the /usr directory is configured to be read-only. In some cases an application may need to be installed into a sub-directory under /usr like /usr/local (i.e., open source software often installs here). A standard zone install will not allow this. However, this can be done by changing the zone configuration so that it mounts an additional directory on the /usr/local directory in the zone, as read-write. In this example, the first Web server is installed in /usr/local/bin, which means we need to configure the zone to support this. To configure a read-write /usr/local directory: 3. In the global zone, create the directory to be exported to the zone, using the mkdir(1) command.
global# mkdir -p /export/home/zones/Web1-zone/local 4. Set the permissions such that only root in the global zone can enter this directory, using the chmod(1) command.
global# chmod 700 /export/home/zones/Web1-zone 5. Create the directory on which the file system is to be mounted if it doesn't already exist (otherwise skip this step).
global# mkdir /usr/local 6. Enter the zone configuration tool for this zone, using zonecfg(1M)
global# zonecfg -z Web1-zone 7. Add a file system to the zone, using the add fs command.
zonecfg:Web1-zone> add fs 8. Specify a directory in the zone on which the file system can be mounted.
zonecfg:Web1-zone:fs> set dir=/usr/local 9. Export the directory from the global zone to the new zone.
zonecfg:Web1-zone:fs> set special=/export/home/zones/Web1-zone/local 10. Set the file system type to the loopback file system.
zonecfg:Web1-zone:fs> set type=lofs 11. Set the directory to have read and write permissions.
zonecfg:Web1-zone:fs> set options=[rw,nodevices] 12. End the configuration.
zonecfg:Web1-zone:fs> end 13. Be sure to verify and commit the configuration, and then install and boot the zone. Perform a system configuration as shown in the Zone Login section, above (Page 8, Step 11).
zonecfg:Web1-zone> verify
zonecfg:Web1-zone> commit
zonecfg:Web1-zone> exit
global# zoneadm -z email-zone install
global# [output omitted here for brevity]
global# zoneadm -z Web1-zone boot
global# zlogin -C Web1-zone
Note: zoneadm list -cv is a quick way to see what state the zone is in. This results in the configuration depicted in Figure 10. You now have two containers created; one with a fixed amount of CPU, and one dynamically sharing CPU with the global zone.
Creating the Second Web Server Container Once the Container for the first Web server is created, installed, and booted, a Container can be created for the second Web server. This Container is similar to the one just created, but will be assigned a different amount of FSS shares, and also includes access to a CD-ROM device and a raw disk partition. To create the second Container: 1. Create the zone for the second Web site using the same process used to create the Web1-zone zone. Be sure to change the name of the zone, its location, the name of the pool used, and the IP address.
zonecfg -z Web2-zone
Web2-zone: No such zone configured
Use 'create' to begin configuring a new zone.
zonecfg:Web2-zone> create
zonecfg:Web2-zone> set zonepath=/export/home/zones/Web2-zone
zonecfg:Web2-zone> add net
zonecfg:Web2-zone:net> set address=10.0.0.3
zonecfg:Web2-zone:net> set physical=eri0
zonecfg:Web2-zone:net> end...
zonecfg:Web2-zone> set pool=pool_default 2. Specify the use of the Fair Share Scheduler, and assign two shares to the zone.
zonecfg:Web2-zone> add rctl
zonecfg:Web2-zone:rctl> set name=zone.cpu-shares
zonecfg:Web2-zone:rctl> add value (priv=privileged,limit=2,action=none)
zonecfg:Web2-zone:rtcl> end To give the users of the Container access to the CD-ROM device:
Add a file system to the zone, using add fs.
zonecfg:Web2-zone> add fs 4. Specify the CD-ROM directory for the zone.
zonecfg:Web2-zone:fs> set dir=/cdrom 5. Export the directory from the global zone to the new zone.
zonecfg:Web2-zone:fs> set special=/cdrom 6. Use the loopback file system.
zonecfg:Web2-zone:fs> set type=l 7. Set the directory to have read only permission because it is a read only CD device.
zonecfg:Web2-zone:fs> set options=[nodevices] 8. End the configuration.
zonecfg:Web2-zone:fs> end
Note: If CD-ROM access is granted to a zone, other zones should not be configured to access the same CD-ROM drive. To configure the zone to access a raw device (raw disk partition) perform the following steps: 9. Add the block device for the raw partition to the zone.
zonecfg:Web2-zone> add device
zonecfg:Web2-zone:device> set match=/dev/dsk/c0t0d0s6
zonecfg:Web2-zone:device> end 10. Add the character device for the raw partition to the zone.
zonecfg:Web2-zone> add device
zonecfg:Web2-zone:device> set match=/dev/rdsk/c0t0d0s6
zonecfg:Web2-zone:device> end
zonecfg:Web2-zone> verify
zonecfg:Web2-zone> commit
zonecfg:Web2-zone> exit
The global zone administrator must ensure the disk partition is not exported to other zones for the duration of this process. Failure to do so may result in data corruption. 11. Install , boot, and configure. Figure 11 illustrates the resulting system configuration from our example. The email server will run on its own guaranteed CPU, protected from the other applications on this system, while the Webservers share the remaining three CPUs. To clarify the FSS share usage, the first Web server application has three out of the total six shares, entitling it to 1.5 CPUs worth of the three CPUs (3*3/6=1.5); the second has two of the six shares, giving it one CPUs worth; and the global zone gets the remaining 0.5 CPUs worth. As a final comment, Oracle acknowledges these types of Solaris Containers as a valid license boundary. In their terminology they are known as Capped Containers, and are made by a combination of Dynamic Resource Pools and Solaris Zones—where the amount of CPUs in the pool determines the size of the license. Back To Top
For More Information While this Solaris How To Guide provides the steps needed to get Solaris Container technology up and running quickly, more sophisticated configurations are possible.
For more information regarding Solaris Containers and Solaris Zones, visit sun.com/solaris.
Manuals System Administration Guide:
Solaris ContainersResource Management and Solaris Zones docs.sun.com/app/docs/doc/817-1592/ Sun BluePrint Articles Solaris Containers—What They Are and How to Use Them www.sun.com/blueprints/0505/819-2679.pdf Web Consolidation on the Sun Fire T1000 using Solaris Containers www.sun.com/blueprints/1205/819-5149.pdf Slicing and Dicing Servers: A Guide to Virtualization and Containment Technologies www.sun.com/blueprints/1005/819-3734.pdf Creating Self-Balancing Solutions with Solaris Containers www.sun.com/blueprints/0605/819-2888.pdf White Papers Solaris Patch Management—Recommended Strategy docs-pdf.sun.com/817-0574/817-0574.pdf Solaris Containers—Server Virtualization and Manageability www.sun.com/software/whitepapers/solaris10/grid_containers.pdf Solaris 9 Resource Manager www.sun.com/software/whitepapers/solaris9/srm.pdf Frequently Asked Questions (FAQ) Open Solaris FAQ opensolaris.org/os/community/zones/faq/ Graphical User Interface Tools Solaris Container Manager GUI www.sun.com/software/products/sunmanagementcenter/ Oracle Licensing Information on Solaris Containers and Oracle www.sun.com/third-party/global/oracle/consolidation/solaris10.html
Thanks