by Duncan Hardie
Published July 2014
Oracle Solaris 11 is a complete, integrated, and open platform engineered for large-scale enterprise environments. Its built-in Oracle Solaris Native Zones technology provides application virtualization through isolated, encapsulated, and highly secure environments that run on top of a single, common kernel. As a result, native zones provide a highly efficient, scalable, zero-overhead virtualization solution that sits at the core of the platform.
With the inclusion of the Kernel Zones feature, Oracle Solaris 11.2 provides a flexible, cost-efficient, cloud-ready solution that is perfect for the data center.
This article describes how to create a kernel zone in Oracle Solaris 11.2, as well as how to configure the kernel zone to your requirements, install it, and boot it.
You will learn about the two main methods of installing a kernel zone: direct installation and installation via an ISO image. In addition, you will learn about a third installation method that enables you to convert a native zone to a kernel zone. You will learn how to update a kernel zone so that it uses a different Oracle Solaris release than the release that is running in the host machine's kernel.
The examples in this article will leave you familiar with the basic procedures for installing, configuring, and managing kernel zones in Oracle Solaris 11.2.
Note: This article demonstrates how to update a kernel zone from Oracle Solaris 11.2 to a later release through examples that mention "Oracle Solaris 11.3" and "Oracle Solaris Next." These examples are purely hypothetical and are for demonstration purposes only; no release later than Oracle Solaris 11.2 is currently available.
Oracle Solaris Zones let you isolate one application from others on the same operating system (OS), allowing you to create a user-, security-, and resource-controlled environment suitable to that particular application. Each Oracle Solaris Zone can contain a complete environment and also allows you to control different resources such as CPU, memory, networking, and storage.
The system administrator who owns the host system can choose to closely manage all the Oracle Solaris Zones on the system. Alternatively, the system administrator can assign rights to other system administrators for specific Oracle Solaris Zones. This flexibility lets you tailor an entire computing environment to the needs of a particular application.
Kernel zones, the newest type of Oracle Solaris Zones, provide all the flexibility, scalability, and efficiency of Oracle Solaris Zones while adding the capability to have zones with independent kernels. This capability is highly useful when you are trying to coordinate the updating of multiple zone environments belonging to different owners.
With kernel zones, the updates can be done at the level of an individual kernel zone at a time that is convenient for each owner. In addition, applications that have specific version requirements can run side by side on the same system and benefit from the high consolidation ratios that Oracle Solaris Zones provide.
In this article, we will create three kernel zones using different methods:
Figure 1 summarizes what we will do:
Figure 1. Illustration of the three methods for creating kernel zones
There are a couple of tasks that need to be completed before we create our first kernel zone. We need to check that the hardware is capable of running kernel zones, and we also need to provide a hint to the system about application memory usage.
Kernel zones will run only on certain types of hardware, as follows:
You can easily check that the system is capable of running kernel zones by using the
virtinfo command, as shown in Listing 1:
root@global:~# virtinfo NAME CLASS non-global-zone supported kernel-zone supported
You can see from the output in Listing 1 that kernel zones are supported.
There are some other hardware prerequisites; for a full list, see the Oracle Solaris Kernel Zones documentation.
When using kernel zones, is it necessary to provide a hint to the system about application memory usage. This information is used to limit the growth of the ZFS Adaptive Replacement Cache (ARC) so that more memory stays available for applications and, in this case, for the kernel zones themselves.
Providing this hint is achieved by setting the
user_reserve_hint_pct parameter. A script is provided for doing this, and the current recommendation is to set the value to 80.
root@global:~# ./set_user_reserve.sh -f 80 Adjusting user_reserve_hint_pct from 0 to 80 Adjustment of user_reserve_hint_pct to 80 successful.
You can find this script and more information by visiting the My Oracle Support website and then accessing Doc ID 1663862.1.
For a full discussion on all the steps involved in creating a kernel zone and configuring all its attributes, please see Creating and Using Oracle Solaris Kernel Zones. This article will concentrate on a subset of the steps to demonstrate how to quickly get a kernel zone instance up and running.
First, check the status of the ZFS file system and the network, as shown in Listing 2:
demo@global:~$ zfs list | grep zones rpool/VARSHARE/zones 16.5G 348G 32K /system/zones demo@global:~$ dladm show-link LINK CLASS MTU STATE OVER net1 phys 1500 unknown -- net2 phys 1500 unknown -- net0 phys 1500 up -- net3 phys 1500 unknown --
Note: In Listing 2, there are no ZFS datasets associated with any specific zones. We will see later how these are created for you as you install zones. Also note that there are no virtual network interface card (VNIC) devices.
Let's also check the Oracle Solaris version of the global zone, as shown in Listing 3, because we will use this information later:
root@global:~# uname -a SunOS global 5.11 11.2 i86pc i386 i86pc
In Listing 3, we can see the version is Oracle Solaris 11.2.
Note: In this article, we will use
uname as a quick way of showing the kernel version of the system. However, that is not the recommended way to check the system version. The recommended way is to query the
entire package, as shown in Listing 4, which also indicates that the version is Oracle Solaris 11.2. (See "Understanding Oracle Solaris 11 Package Versioning" for an explanation about how to decipher the output when you query the entire package.)
demo@dcsw-79-168:~$ pkg list entire NAME (PUBLISHER) VERSION IFO entire 0.5.11-0.175.2.0.0.41.0 i—
We can also see that the system has a publisher set up:
root@global:~# pkg publisher PUBLISHER TYPE STATUS P LOCATION solaris origin online F http://ipkg.us.oracle.com/solaris11/dev/
Let's start by creating our first kernel zone using the command line, as shown in Listing 5:
root@global:~# zonecfg -z myfirstkz create -t SYSsolaris-kz
In Listing 5, note that all we need to supply is the zone name (
myfirstkz) and the kernel zone brand (
By default, all Oracle Solaris Zones are configured to have an automatic VNIC called
anet, which gives us a network device automatically. We cannot see this network device, but it is automatically created upon booting the zone (and also automatically destroyed upon shutdown). We can check this by using the
root@global:~# dladm show-link LINK CLASS MTU STATE OVER net1 phys 1500 unknown -- net2 phys 1500 unknown -- net0 phys 1500 up -- net3 phys 1500 unknown --
We can also see that, as of yet, no storage has been created for our kernel zone:
root@global:~# zfs list | grep zones rpool/VARSHARE/zones 16.5G 348G 32K /system/zones
We can verify that the kernel zone is now in the configured state by using the
root@global:~# zoneadm list -cv ID NAME STATUS PATH BRAND IP 0 global running / solaris shared - myfirstkz configured - solaris-kz excl
Let's take a look at the default settings for the kernel zone that we have created. We can do this by passing the
info option to the
zonecfg command, as shown in Listing 6:
root@global~# zonecfg -z myfirstkz info zonename: myfirstkz brand: solaris-kz autoboot: false autoshutdown: shutdown bootargs: pool: scheduling-class: hostid: 0x3888f5a3 tenant: anet: lower-link: auto allowed-address not specified configure-allowed-address: true defrouter not specified allowed-dhcp-cids not specified link-protection: mac-nospoof mac-address: auto mac-prefix not specified mac-slot not specified vlan-id not specified priority not specified rxrings not specified txrings not specified mtu not specified maxbw not specified rxfanout not specified vsi-typeid not specified vsi-vers not specified vsi-mgrid not specified etsbw-lcl not specified cos not specified evs not specified vport not specified id: 0 device: match not specified storage: dev:/dev/zvol/dsk/rpool/VARSHARE/zones/myfirstkz/disk0 id: 0 bootpri: 0 capped-memory: physical: 2G
From the output in Listing 6, we can see that the zone is called
myfirstkz, that it is a kernel zone (
brand: solaris-kz), that we have a boot disk (and its location is
dev:/dev/zvol/dsk/rpool/VARSHARE/zones/myfirstkz/disk0) and, finally, that we have 2 GB of physical memory assigned to this kernel zone.
What we don't see is the amount of CPU resources we have for this kernel zone. When nothing is specified, the default is to have one virtual CPU assigned. We'll see how to verify this later when we boot the kernel zone.
Now that the kernel zone has been created, we need to install it.
For this first installation, we are going to use what is called a direct installation. With a direct installation, the installer runs on the host. It will create and format the kernel zone's boot disk and install Oracle Solaris packages on that disk, using the host's package publishers. Since the installer is running on the host, the installer can install only the exact version of Oracle Solaris that it is actively running on the host.
This installation method makes use of the Oracle Solaris 11 Image Packaging System. You will need to make sure you have access to your Image Packaging System repository; in this case, we have network access to our repository. For more details on the Image Packaging System, see "Introducing the Basics of Image Packaging System (IPS) on Oracle Solaris 11."
Run the following command to install the
myfirstkz kernel zone:
root@global:~# zoneadm -z myfirstkz install Progress being logged to /var/log/zones/zoneadm.20140724T124406Z.myfirstkz.install pkg cache: Using /var/pkg/publisher. Install Log: /system/volatile/install.7395/install_log AI Manifest: /tmp/zoneadm6814.Voa43n/devel-ai-manifest.xml SC Profile: /usr/share/auto_install/sc_profiles/enable_sci.xml Installation: Starting ... Creating IPS image Installing packages from: solaris origin: http://ipkg.us.oracle.com/solaris11/dev/ The following licenses have been accepted and not displayed. Please review the licenses for the following packages post-install: consolidation/osnet/osnet-incorporation Package licenses may be viewed using the command: pkg info --license <pkg_fmri> DOWNLOAD PKGS FILES XFER (MB) SPEED Completed 483/483 64276/64276 543.8/543.8 11.6M/s PHASE ITEMS Installing new actions 87529/87529 Updating package state database Done Updating package cache 0/0 Updating image state Done Creating fast lookup database Done Installation: Succeeded Done: Installation completed in 538.018 seconds.
We can check on the status of the
myfirstkz kernel zone using the
root@global:~# zoneadm list -cv ID NAME STATUS PATH BRAND IP 0 global running / solaris shared - myfirstkz installed - solaris-kz excl
Note: A kernel zone needs a boot disk on which it is installed; by using the command shown in Listing 7, we can see that this boot disk has been created for us:
root@global:~# zfs list | grep zones rpool/VARSHARE/zones 16.5G 348G 32K /system/zones rpool/VARSHARE/zones/myfirstkz 16.5G 348G 31K /system/zones/myfirstkz rpool/VARSHARE/zones/myfirstkz/disk0 16.5G 361G 2.92G -
You can see in Listing 7 that the
/myfirstkz/disk0 dataset has been created automatically for you.
The final step in getting
myfirstkz up and running is to boot it and set up the system configuration. We will boot the zone and then access its console using one command at the command line, as shown in Listing 8, so the majority of the console output can be seen:
root@global:~# zoneadm -z myfirstkz boot; zlogin -C myfirstkz [Connected to zone 'myfirstkz' console] Boot device: disk0 File and args: reading module /platform/i86pc/amd64/boot_archive...done. reading kernel file /platform/i86pc/kernel/amd64/unix...done. SunOS Release 5.11 Version 11.2 64-bit Copyright (c) 1983, 2014, Oracle and/or its affiliates. All rights reserved. Loading smf(5) service descriptions: 183/183 Configuring devices.
-C option to
zlogin shown in Listing 8 lets us access the zone console; the command will bring us into the zone and let us work within the zone.
Because no system configuration files are available, the System Configuration Tool starts up, as shown in Figure 2.
Figure 2. Initial screen of the System Configuration Tool
Press F2 to continue.
In the System Identity screen (shown in Figure 3), enter
myfirstkz as the computer name, and then press F2 to continue.
Figure 3. System Identity screen
In the Network screen (shown in Figure 4), Enter the network settings appropriate for your network and then press F2. Here we will select Automatically.
Figure 4. Network screen
In the Time Zone: Regions screen (shown in Figure 5), choose the time zone region appropriate for your location. In this example, we chose Europe. Then press F2.
Figure 5. Time Zone: Regions screen
In the Time Zone: Locations screen (shown in Figure 6), choose the time zone location appropriate for your location, and then press F2.
Figure 6. Time Zone: Locations screen
In the Time Zone screen (shown in Figure 7), choose the time zone appropriate for your location, and then press F2.
Figure 7. Time Zone screen
In the Locale: Language screen (shown in Figure 8), choose the language appropriate for your location, and then press F2.
Figure 8. Locale: Language screen
In the Locale: Territory screen (shown in Figure 9), choose the language territory appropriate for your location, and then press F2.
Figure 9. Locale: Territory screen
In the Date and Time screen (shown in Figure 10), set the date and time, and then press F2.
Figure 10. Date and Time screen
In the Keyboard screen (shown in Figure 11), select the appropriate keyboard, and then press F2.
Figure 11. Keyboard screen
In the Users screen (shown in Figure 12), choose a root password and enter information for a user account. Then press F2.
Figure 12. Users screen
In the Support — Registration screen (shown in Figure 13), enter your My Oracle Support credentials. Then press F2.
Figure 13. Support — Registration screen
In the Support — Network Configuration screen (shown in Figure 14), choose how you will send configuration data to Oracle. Then press F2.
Figure 14. Support — Network Configuration screen
In the System Configuration Summary screen (shown in Figure 15), verify that the configuration you have chosen is correct and apply the settings by pressing F2.
Figure 15. System Configuration Summary screen
The zone will continue booting, and soon you will see the console login:
SC profile successfully generated as: /etc/svc/profile/sysconfig/sysconfig-20140724-130314/sc_profile.xml Exiting System Configuration Tool. Log is available at: /system/volatile/sysconfig/sysconfig.log.287 Hostname: myfirstkz Jul 24 15:20:29 myfirstkz sendmail: My unqualified host name (myfirstkz) unknown; sleeping for retry Jul 24 15:20:29 myfirstkz sendmail: My unqualified host name (myfirstkz) unknown; sleeping for retry myfirstkz console login:
The zone is now ready to be logged in to. For this example, we will now exit the console using the "
~." escape sequence.
You can check that your zone is booted and running using the
root@global:~# zoneadm list -cv ID NAME STATUS PATH BRAND IP 0 global running / solaris shared 2 myfirstkz running - solaris-kz excl
As promised, a VNIC was automatically created for us when the zone was booted. We can verify this by using the
dladm command shown in Listing 9:
root@global:~# dladm show-link LINK CLASS MTU STATE OVER net1 phys 1500 unknown -- net2 phys 1500 unknown -- net0 phys 1500 up -- net3 phys 1500 unknown -- myfirstkz/net0 vnic 1500 up net0
In Listing 9, we can see the VNIC is listed as
The last step is to log in to your zone and have a look. You can do this from the global zone using the
zlogin command, as shown in Listing 10:
root@global:~# zlogin myfirstkz [Connected to zone 'myfirstkz' pts/1] Oracle Corporation SunOS 5.11 11.2 June 2014 root@myfirstkz:~# uname -a SunOS myfirstkz 5.11 11.2 i86pc i386 i86pc root@myfirstkz:~# ipadm show-addr ADDROBJ TYPE STATE ADDR lo0/v4 static ok 127.0.0.1/8 net0/v4 dhcp ok 10.134.79.210/24 lo0/v6 static ok ::1/128 net0/v6 addrconf ok fe80::8:20ff:fe47:ca30/10 root@myfirstkz:~# dladm show-link LINK CLASS MTU STATE OVER net0 phys 1500 up -- root@myfirstkz:~# zfs list NAME USED AVAIL REFER MOUNTPOINT rpool 4.65G 10.7G 32.5K /rpool rpool/ROOT 2.58G 10.7G 31K legacy rpool/ROOT/solaris-5 2.58G 10.7G 2.08G / rpool/ROOT/solaris-5/var 510M 10.7G 508M /var rpool/VARSHARE 2.52M 10.7G 2.43M /var/share rpool/VARSHARE/pkg 63K 10.7G 32K /var/share/pkg rpool/VARSHARE/pkg/repositories 31K 10.7G 31K /var/share/pkg/repositories rpool/VARSHARE/zones 31K 10.7G 31K /system/zones rpool/dump 1.03G 10.8G 1.00G - rpool/export 96.5K 10.7G 32K /export rpool/export/home 64.5K 10.7G 32K /export/home rpool/export/home/demo 32.5K 10.7G 32.5K /export/home/demo rpool/swap 1.03G 10.8G 1.00G - root@myfirstkz:~# zonename global root@myfirstkz:~# exit logout [Connection to zone 'myfirstkz' pts/1 closed]
Note: In Listing 10, we did not use the
-C option for the
zlogin command, which means we are not accessing the zone via its console. This is why we can simply exit the shell at the end to leave the zone.
Let's look at the output shown in Listing 10 to see what we have:
unamecommand shows that we are running on Oracle Solaris 11.2—the same kernel version used in the global zone in which our
myfirstkzkernel zone is running.
ipadmcommand shows the IP address for
myfirstkz. There are four entries: two loopback devices (IPv4 and IPv6), our IPv4
net0device with an IP address of
10.134.79.210and, finally, an IPv6
dladmcommand shows our automatically created
zfs listcommand shows our ZFS dataset.
zonenamecommand shows that our zone name is
global. With native zones, this would be the actual zone name. However, a kernel zone actually runs a full kernel instance, so users running inside the kernel zone have their own instance of a global zone.
If you want to determine the zone name from inside the kernel zone, you can use the
root@global:~# zlogin myfirstkz [Connected to zone 'myfirstkz' pts/1] Oracle Corporation SunOS 5.11 11.2 June 2014 root@myfirstkz:~# virtinfo -c current get zonename NAME CLASS PROPERTY VALUE kernel-zone current zonename myfirstkz root@myfirstkz:~# exit logout
Note: From within
myfirstkz, we cannot see any information about the global zone; we can see only the attributes of our own zone.
You have now verified that
myfirstkz is up and running. You can give the login information to your users to allow them to complete the setup of their team's kernel zone as if it were a single system.
One of the main features of kernel zones is the ability to run your kernel zone with a different kernel version from that of the host global zone.
Starting with Oracle Solaris 11.2, kernel zones support both backwards and forwards compatibility. What that means in practice is that you can not only have a kernel zone running Oracle Solaris 11.2 on a host running a later Oracle Solaris version, say Oracle Solaris 11.3 (when it is available), but you can also have a kernel zone running a later Oracle Solaris version, say Oracle Solaris 11.3, on a host running Oracle Solaris 11.2. Figure 16 illustrates this capability.
Figure 16. Example of forward and backward compatibility of kernel zones
Let's update our kernel zone to use a later Oracle Solaris release—a hypothetical "Oracle Solaris Next" release—rather than the release running on the host (Oracle Solaris 11.2).
First, let's use the command shown in Listing 11 to look at what boot environments we have from the host global zone:
root@global:~# beadm list BE Active Mountpoint Space Policy Created -- ------ ---------- ----- ------ ------- solaris - - 44.67M static 2012-01-26 18:59 solaris-1 - - 47.78M static 2014-06-25 08:12 solaris-2 - - 46.71M static 2014-06-25 08:40 solaris-3 - - 1.03G static 2014-06-25 23:30 solaris-3-backup-1 - - 221.0K static 2014-07-10 14:23 solaris-4 NR / 52.77G static 2014-07-20 18:41 solaris-backup-1 - - 144.0K static 2012-01-26 19:28
In Listing 11, we could see a list of native zone boot environments, if there were any. However, we will not see kernel zone boot environments listed, because a kernel zone has its own boot disk.
Let's check what our current publisher is and point the kernel zone to a publisher that has a newer kernel. We start by logging in to
myfirstkz, as shown in Listing 12:
root@global:~# zlogin myfirstkz [Connected to zone 'myfirstkz' pts/2] Oracle Corporation SunOS 5.11 11.2 June 2014 root@myfirstkz:~# pkg publisher PUBLISHER TYPE STATUS P LOCATION solaris origin online F http://ipkg.us.oracle.com/solaris11/dev/ root@myfirstkz:~# pkg set-publisher -G '*' -g http://ipkg.us.oracle.com/solaris-n/dev/ solaris root@myfirstkz:~# pkg publisher PUBLISHER TYPE STATUS P LOCATION solaris origin online F http://ipkg.us.oracle.com/solaris-n/dev/
Note that in this example, we will use an internally created repository. You will be able to reproduce this for yourself as later releases of Oracle Solaris become available. In Listing 12, we can see that we are running Oracle Solaris 11.2, and we have set our publisher to point to the
Before we update, let's look at what the kernel zone sees as its boot environment:
root@myfirstkz:~# beadm list BE Active Mountpoint Space Policy Created -- ------ ---------- ----- ------ ------- solaris-5 NR / 7.91M static 2014-07-24 13:44
Now, let's update our kernel zone, as shown in Listing 13:
root@myfirstkz:~# pkg update --accept ------------------------------------------------------------ Package: pkg://email@example.com,5.12- 126.96.36.199.0.52.0:20140714T022826Z License: lic_OTN You acknowledge that your use of this Oracle Solaris software product is subject to (i) the license terms that you accepted when you obtained the right to use Oracle Solaris software; or (ii) the license terms that you agreed to when you placed your Oracle Solaris software order with Oracle; or (iii) the Oracle Solaris software license terms included with the hardware that you acquired from Oracle; or, if (i), (ii) or (iii) are not applicable, then, (iv) the OTN License Agreement for Oracle Solaris (which you acknowledge you have read and agree to) available at http://www.oracle.com/technetwork/licenses/solaris-cluster-express-license- 167852.html. Note: Software downloaded for trial use or downloaded as replacement media may not be used to update any unsupported software. Packages to remove: 37 Packages to install: 57 Packages to update: 432 Mediators to change: 4 Create boot environment: Yes Create backup boot environment: No DOWNLOAD PKGS FILES XFER (MB) SPEED Completed 526/526 23157/23157 363.4/363.4 16.7M/s PHASE ITEMS Removing old actions 6580/6580 Installing new actions 9594/9594 Updating modified actions 24807/24807 Updating package state database Done Updating package cache 469/469 Updating image state Done Creating fast lookup database Done Updating package cache 1/1 A clone of solaris-5 exists and has been updated and activated. On the next boot the Boot Environment solaris-6 will be mounted on '/'. Reboot when ready to switch to this updated BE. Updating package cache 1/1 --------------------------------------------------------------------------- NOTE: Please review release notes posted at: http://www.oracle.com/pls/topic/lookup?ctx=solaris11&id=SERNS ---------------------------------------------------------------------------
In the command shown in Listing 13, we use the
--accept option to automatically accept any licenses. We can see in the output that a boot environment has been created. Let's look at what that is, as shown in Listing 14:
root@myfirstkz:~# beadm list BE Active Mountpoint Space Policy Created -- ------ ---------- ----- ------ ------- solaris-5 N / 7.91M static 2014-07-24 13:44 solaris-6 R - 7.39G static 2014-07-24 17:04
In Listing 14, we can see from the
R next to the
solaris-6 boot environment that after a reboot, we will select this new environment.
Finally, let's reboot the zone.
root@myfirstkz:~# reboot [Connection to zone 'myfirstkz' pts/2 closed] root@global:~#
We are now back in the host global zone, and we can use
zoneadm to check the status of our kernel zone, as shown in Listing 15:
root@global:~# zoneadm list -cv ID NAME STATUS PATH BRAND IP 0 global running / solaris shared 3 myfirstkz running - solaris-kz excl
As shown in Listing 15, our kernel zone has already rebooted and is running again.
Let's log back in, as shown in Listing 16, and check what kernel version we are running:
root@global:~# zlogin myfirstkz [Connected to zone 'myfirstkz' pts/2] Oracle Corporation SunOS 5.n sn_52 June 2014
Listing 16 shows we are running a completely different kernel: the hypothetical "Oracle Solaris Next."
Let's run the command shown in Listing 17 to take a final look at the boot environments before we leave this kernel zone:
root@myfirstkz:~# beadm list BE Active Mountpoint Space Policy Created -- ------ ---------- ----- ------ ------- solaris-5 - - 12.35M static 2014-07-24 13:44 solaris-6 NR / 7.50G static 2014-07-24 17:04 root@myfirstkz:~# exit logout [Connection to zone 'myfirstkz' pts/2 closed]
In Listing 17, we can see that we are running in the new boot environment.
Sometimes you might not want to do a direct installation with a kernel zone; you might want to install from an ISO image instead. This is supported for kernel zones, and this section will show how to do that.
We will also use this opportunity to explore how to allocate some dedicated CPU resources to the kernel zone, as well as how to add some extra memory and increase the size of its boot disk.
Let's create a new kernel zone similar to what we did before, but this time we will use the
zonecfg command to add some dedicated CPU resources.
Let's start by checking how many CPU resources we have:
root@global:~# psrinfo -t socket: 0 core: 0 cpus: 0,8 core: 1 cpus: 1,9 core: 2 cpus: 2,10 core: 3 cpus: 3,11 socket: 1 core: 8 cpus: 4,12 core: 9 cpus: 5,13 core: 10 cpus: 6,14 core: 11 cpus: 7,15
Now, let's create a new kernel zone called
iso-kz and then add four CPU's worth of dedicated CPU resources to it:
root@global:~# zonecfg -z iso-kz create -t SYSsolaris-kz root@global:~# zonecfg -z iso-kz zonecfg:iso-kz> add dedicated-cpu zonecfg:iso-kz:dedicated-cpu> set ncpus=4 zonecfg:iso-kz:dedicated-cpu> end zonecfg:iso-kz> verify zonecfg:iso-kz> commit zonecfg:iso-kz> exit
We can check that the zone creation and resource configuration worked by using the
root@global:~# zonecfg -z iso-kz info dedicated-cpu dedicated-cpu: ncpus: 4 cpus not specified cores not specified sockets not specified
You can set a kernel zone to have either virtual CPUs or dedicated CPUs. The difference between the two types is basically about sharing.
We can also use the
zonecfg command to add some extra memory to the kernel zone:
root@global:~# zonecfg -z iso-kz zonecfg:iso-kz> select capped-memory zonecfg:iso-kz:capped-memory> set physical=3g zonecfg:iso-kz:capped-memory> end zonecfg:iso-kz> verify zonecfg:iso-kz> commit zonecfg:iso-kz> exit root@global:~# zonecfg -z iso-kz info capped-memory capped-memory: physical: 3G
It's now time to install our zone. We will use a hypothetical Oracle Solaris 11.3 ISO image to do this and we will also increase the size of the install disk. The default is a 16 GB disk, so let's increase that to 24 GB. Listing 18 shows how you do this at installation time:
root@global:~# zoneadm -z iso-kz install -b /root/sol-11_3-42-text-x86.iso -x install-size=24g
In Listing 18, you can see that this time, the image we used is using the text installer.
Once we have answered the usual installation questions, we can log in to our zone, as shown in Listing 19:
root@global:~# zoneadm list -cv ID NAME STATUS PATH BRAND IP 0 global running / solaris shared 3 myfirstkz running - solaris-kz excl 5 iso-kz running - solaris-kz excl root@global:~# zlogin iso-kz [Connected to zone 'iso-kz' pts/2] Oracle Corporation SunOS 5.11 11.3 June 2014 root@:~# root@solarisiso-kz:~# psrinfo -t socket: 0 core: 0 cpu: 0 core: 1 cpu: 1 core: 2 cpu: 2 core: 3 cpu: 3 root@:~# exit logout [Connection to zone 'iso-kz' pts/2 closed]
In Listing 19, we can see the four dedicated CPUs we assigned and we can see that we are running a release different than that of the host global zone.
Before we move on, let's shut down our two kernel zones:
root@global:~# zoneadm -z myfirstkz shutdown root@global:~# zoneadm -z iso-kz shutdown
The final operation to try is converting a native zone to a kernel zone, which is made especially easy through the use of Oracle Solaris Unified Archives.
In this example, we will use a native zone that has already been created. If you are not sure how to create a native zone, see "How to Get Started Creating Oracle Solaris Zones in Oracle Solaris 11."
Let's start by having a look at the native zone we are going to convert, as shown in Listing 20:
root@global:~# zoneadm list -cv ID NAME STATUS PATH BRAND IP 0 global running / solaris shared 6 native-zone running /system/zones/native-zone solaris excl - myfirstkz installed - solaris-kz excl - iso-kz installed - solaris-kz excl root@global:~# zlogin native-zone [Connected to zone 'native-zone' pts/2] Oracle Corporation SunOS 5.11 11.2 June 2014 root@native-zone:~# touch my_special_files root@native-zone:~# zonename native-zone root@native-zone:~# exit logout [Connection to zone 'native-zone' pts/2 closed] root@global:~# zoneadm list -cv ID NAME STATUS PATH BRAND IP 0 global running / solaris shared - myfirstkz installed - solaris-kz excl - iso-kz installed - solaris-kz excl - native-zone installed /system/zones/native-zone solaris excl
In Listing 20, we can see our native-zone is already up and running, and we have logged in and created a file called
my_special_files. This example is just to reflect any configuration that we might have done when taking a zone from a real environment. Finally, we checked the zone name, logged out, and shut down the native zone.
Note: One of the big advantages of using a Unified Archive to capture a zone is that you can do the capture on a running zone, which means you can avoid outages to end users. In this case, because we want to convert our native zone to a kernel zone (rather than clone the native zone), we shut down the native zone.
Now let's create a Unified Archive of our native zone:
root@global:~# archiveadm create -z native-zone ./native-zone-archive.uar Initializing Unified Archive creation resources... Unified Archive initialized: /root/native-zone-archive.uar Logging to: /system/volatile/archive_log.26165 Executing dataset discovery... Dataset discovery complete Preparing archive system image... Beginning archive stream creation... Archive stream creation complete Beginning final archive assembly... Archive creation complete
Once we have created the archive, we can examine it to see what it contains:
root@global:~# ls -l total 5602777 -rw-r--r-- 1 root root 901992448 Jul 1 06:18 0.175.2_ai_i386.iso -rw-r--r-- 1 root root 1308958720 Jul 24 17:27 native-zone-archive.uar -rw-r--r-- 1 demo staff 675102720 Jul 24 07:37 sol-11_2-42-text-x86.iso root@global:~# archiveadm info -v ./native-zone-archive.uar Archive Information Creation Time: 2014-07-24T21:54:08Z Source Host: global Architecture: i386 Operating System: Oracle Solaris 11.3 X86 Recovery Archive: No Unique ID: e1bf0d42-338b-e879-fec4-ab78290ef55c Archive Version: 1.0 Deployable Systems 'native-zone' OS Version: 0.5.11 OS Branch: 0.175.3.0.0.1.0 Active BE: solaris Brand: solaris Size Needed: 978MB Unique ID: f488ea7c-ab1e-6cc4-d407-c60fce1e3818 AI Media: 0.175.3_ai_i386.iso Root-only: Yes
Next, let's configure a new kernel zone and when we are ready to install it, we will pass in the archive, as shown in Listing 21:
root@global:~# zonecfg -z converted-zone-kz create -t SYSsolaris-kz root@global:~# zoneadm -z converted-zone-kz install -a ./native-zone-archive.uar Progress being logged to /var/log/zones/zoneadm.20140724T233807Z.converted-zone-kz.install [Connected to zone 'converted-zone-kz' console] Boot device: cdrom1 File and args: -B install=true,auto-shutdown=true -B aimanifest=/system/shared/ai.xml reading module /platform/i86pc/amd64/boot_archive...done. reading kernel file /platform/i86pc/kernel/amd64/unix...done. SunOS Release 5.11 Version 11.2 64-bit Copyright (c) 1983, 2014, Oracle and/or its affiliates. All rights reserved. Remounting root read/write Probing for device nodes ... Preparing image for use Done mounting image Configuring devices. Hostname: solaris Using specified install manifest : /system/shared/ai.xml solaris console login: Automated Installation started The progress of the Automated Installation will be output to the console Detailed logging is in the logfile at /system/volatile/install_log Press RETURN to get a login prompt at any time. 23:40:15 Install Log: /system/volatile/install_log 23:40:15 Using XML Manifest: /system/volatile/ai.xml 23:40:15 Using profile specification: /system/volatile/profile 23:40:15 Starting installation. 23:40:15 0% Preparing for Installation 23:40:15 100% manifest-parser completed. 23:40:15 100% None 23:40:15 0% Preparing for Installation 23:40:18 1% Preparing for Installation 23:40:18 2% Preparing for Installation 23:40:19 3% Preparing for Installation 23:40:19 4% Preparing for Installation 23:40:19 5% archive-1 completed. 23:40:21 8% target-discovery completed. 23:40:23 Pre-validating manifest targets before actual target selection 23:40:23 Selected Disk(s) : c1d0 23:40:24 Pre-validation of manifest targets completed 23:40:24 Validating combined manifest and archive origin targets 23:40:24 Selected Disk(s) : c1d0 23:40:24 9% target-selection completed. 23:40:24 10% ai-configuration completed. 23:40:24 9% var-share-dataset completed. 23:40:29 10% target-instantiation completed. 23:40:29 10% Beginning archive transfer 23:40:29 Commencing transfer of stream: ce6d4b69-ad85-e7e1-aaf7-fdbfdc17f001-0.zfs to rpool 23:40:35 30% Transferring contents 23:40:39 50% Transferring contents 23:40:43 70% Transferring contents 23:40:54 86% Transferring contents 23:41:09 Completed transfer of stream: 'ce6d4b69-ad85-e7e1-aaf7-fdbfdc17f001-0.zfs' from file:///system/shared/uafs/OVA 23:41:12 Archive transfer completed 23:41:31 90% generated-transfer-965-1 completed. 23:41:31 90% Beginning IPS transfer 23:41:31 Setting post-install publishers to: 23:41:31 solaris 23:41:31 origin: http://ipkg.us.oracle.com/solaris11/dev/ 23:41:32 90% generated-transfer-965-2 completed. 23:41:32 Changing target pkg variant. This operation may take a while 23:51:17 90% apply-pkg-variant completed. 23:51:21 Setting boot devices in firmware 23:51:21 91% boot-configuration completed. 23:51:21 91% update-dump-adm completed. 23:51:21 92% setup-swap completed. 23:51:22 92% device-config completed. 23:51:23 92% apply-sysconfig completed. 23:51:23 93% transfer-zpool-cache completed. 23:51:36 98% boot-archive completed. 23:51:36 98% transfer-ai-files completed. 23:51:37 98% cleanup-archive-install completed. 23:51:38 100% create-snapshot completed. 23:51:39 100% None 23:51:39 Automated Installation succeeded. 23:51:39 You may wish to reboot the system at this time. Automated Installation finished successfully Shutdown requested. Shutting down the system Log files will be available in /var/log/install/ after reboot svc.startd: The system is coming down. Please wait. root@global:~# zoneadm list -cv ID NAME STATUS PATH BRAND IP 0 global running / solaris shared - myfirstkz installed - solaris-kz excl - iso-kz installed - solaris-kz excl - native-zone installed /system/zones/native-zone solaris excl - converted-zone-kz installed - solaris-kz excl
As we can see in Listing 21, the install process completed successfully and we have an installed kernel zone.
Let's boot up our newly converted zone and have a look at it, as shown in Listing 22:
root@global:~# zoneadm -z converted-zone-kz boot root@global:~# zoneadm list -cv ID NAME STATUS PATH BRAND IP 0 global running / solaris shared 3 converted-zone-kz running - solaris-kz excl - duckstack unavailable - solaris-kz excl - native-zone installed /system/zones/native-zone solaris excl root@global:~# zlogin converted-zone-kz [Connected to zone 'converted-zone-kz' pts/1] Oracle Corporation SunOS 5.11 11.2 June 2014 root@unknown:~# ls my_special_files
In Listing 22, we can see that the contents of our native zone have been preserved.
In this article, we explored how to create, install, boot, and configure a kernel zone. You learned that Oracle Solaris Kernel Zones have the ability to run kernel versions that are different from the kernel version running on the host. We also saw how to do a direct installation and an installation based on an ISO image. Finally, we saw how to convert a native zone to a kernel zone using Unified Archives.
Also see these additional resources:
Duncan Hardie is an Oracle Solaris Product Manager with responsibility for Oracle Solaris cloud and virtualization technologies. Joining Oracle as part of the Sun Microsystems acquisition, he started in engineering working on device drivers for fault-tolerant products and moved into customer-facing roles related to monitoring, high-performance computing, and virtualization. In his current role, Duncan works to help define, deliver, and position Oracle Solaris offerings.
|Revision 1.1, 12/10/2014|
|Revision 1.0, 07/28/2014|