Solaris 10's zone partitioning technology can be used to create local
zones that behave like virtual servers. All local zones are controlled
from the system's global zone. Processes running in a zone are
completely isolated from the rest of the system. This isolation
prevents processes that are running in one zone from monitoring or
affecting processes that are running in other zones. Note that
processes running in a local zone can be monitored from global zone; but
the pro
cesses running in a global zone or even in another local zone
cannot be monitored from a local zone.
lets make it very simple, a zone is a virtual operating system
environment created within a single instance of the Solaris operating
system. Efficient resource utilization is the main goal of this
technology.
As of now, the upper limit for the number of zones that can be
created/run on a system is 8192; of course, depending on the resource
availability, a single system may or may not run all the configured
zones effectively.
Global Zone
When we install Solaris 10, a global zone gets installed automatically; and the core operating system runs under global zone. To list all the configured zones, we can usezoneadm command:# zoneadm list -v ID NAME STATUS PATH 0 global running /Global zone is the only one:
- bootable from the system hardware
- to be used for system-wide administrative control, such as physical devices, routing, or dynamic reconfiguration (DR). ie., global zone is the only zone that is aware of all devices and all file systems
- from which a non-global zone can be configured, installed, managed, or uninstalled. ie., global zone is the only zone that is aware of the existence of non-global (local) zones and their configurations. It is not possible to create local zones, within a local zone
Steps to create a Local Zone
Prerequisites:- Plenty of disk space to hold the newly installed zone. It needs at least 2G space to copy the essential files to the local zone, and of course the disk space needed by the application(s) you are planning to run, in this zone; and
- A dedicated IP for network connectivity
- Check the disk space & network configuration
% df -h / Filesystem size used avail capacity Mounted on /dev/dsk/c1t1d0s0 29G 22G 7.1G 76% / % ifconfig -a lo0: flags=2001000849 mtu 8232 index 1 inet 127.0.0.1 netmask ff000000 eri0: flags=1000843 mtu 1500 index 2 inet 192.168.74.217 netmask fffffe00 broadcast 192.168.75.255
- Since there is more than 5G free space, I've decided to install a local zone under
/zones.% mkdir /zones
- Next step is to define/create the zone root. This is the path
to zone's root directory that is relative to the global zone's root
directory. Zone root must be owned by root user with the mode 700. This will be used in setting the
zonepathproperty, during the zone creation process% cd /zones % mkdir appserver % chmod 700 appserver % ls -l total 2 drwx------ 2 root root 512 Feb 17 12:46 appserver
- Create & configure a new 'sparse root' local zone, with root privileges
% zonecfg -z appserv appserv: No such zone configured Use 'create' to begin configuring a new zone. zonecfg:appserv> create zonecfg:appserv> set zonepath=/zones/appserver zonecfg:appserv> set autoboot=true zonecfg:appserv> add net zonecfg:appserv:net> set physical=eri0 zonecfg:appserv:net> set address=192.168.175.126 zonecfg:appserv:net> end zonecfg:appserv> add fs zonecfg:appserv:fs> set dir=/repo2 zonecfg:appserv:fs> set special=/dev/dsk/c2t40d1s6 zonecfg:appserv:fs> set raw=/dev/rdsk/c2t40d1s6 zonecfg:appserv:fs> set type=ufs zonecfg:appserv:fs> set options noforcedirectio zonecfg:appserv:fs> end zonecfg:appserv> add inherit-pkg-dir zonecfg:appserv:inherit-pkg-dir> set dir=/opt/csw zonecfg:appserv:inherit-pkg-dir> end zonecfg:appserv> info zonepath: /zones/appserver autoboot: true pool: inherit-pkg-dir: dir: /lib inherit-pkg-dir: dir: /platform inherit-pkg-dir: dir: /sbin inherit-pkg-dir: dir: /usr inherit-pkg-dir: dir: /opt/csw net: address: 192.168.175.126 physical: eri0 zonecfg:appserv> verify zonecfg:appserv> commit zonecfg:appserv> exit
Sparse Root Zone Vs Whole Root Zone
In a Sparse Root Zone, the directories/usr, /sbin, /lib and /platformwill be mounted as loopback file systems. That is, although all those directories appear as normal directories under the sparse root zone, they will be mounted as read-only file systems. Any change to those directories in the global zone can be seen from the sparse root zone.
However if you need the ability to write into any of those directories listed above, you may need to configure a Whole Root Zone. For example, softwares like ClearCase need write permissions to/usrdirectory. In that case configuring a Whole Root Zone is the way to go. The steps for creating and configuring a new 'Whole Root' local zone are as follows:
% zonecfg -z appserv appserv: No such zone configured Use 'create' to begin configuring a new zone. zonecfg:appserv> create zonecfg:appserv> set zonepath=/zones/appserver zonecfg:appserv> set autoboot=true zonecfg:appserv> add net zonecfg:appserv:net> set physical=eri0 zonecfg:appserv:net> set address=192.168.175.126 zonecfg:appserv:net> end zonecfg:appserv> add inherit-pkg-dir zonecfg:appserv:inherit-pkg-dir> set dir=/opt/csw zonecfg:appserv:inherit-pkg-dir> end zonecfg:appserv> remove inherit-pkg-dir dir=/usr zonecfg:appserv> remove inherit-pkg-dir dir=/sbin zonecfg:appserv> remove inherit-pkg-dir dir=/lib zonecfg:appserv> remove inherit-pkg-dir dir=/platform zonecfg:appserv> info zonepath: /zones/appserver autoboot: true pool: inherit-pkg-dir: dir: /opt/csw net: address: 192.168.175.126 physical: eri0 zonecfg:appserv> verify zonecfg:appserv> commit zonecfg:appserv> exit
Brief explanation of the properties that I added:
\*zonepath=/zones/appserver
Local zone's root directory, relative to global zone's root directory. ie., local zone will have all thebin, lib, usr, dev, net, etc, var, optetc., directories physically under /zones/appserver directory
\*autoboot=true
boot this zone automatically when the global zone is booted
\*physical=eri0
eri0card is used for the physical interface
\*address=192.168.175.126
192.168.175.126 is the IP address. It must have all necessary DNS entries
[Added 08/25/08] The wholeadd fssection adds the file system to the zone. In this example, the file system that is being exported to the zone is an existing UFS file system.
\*set dir=/repo2
/repo2is the mount point in the local zone
\*set special=/dev/dsk/c2t40d1s6 set raw=/dev/rdsk/c2t40d1s6
Grant access to the block (/dev/dsk/c2t40d1s6) and raw (/dev/rdsk/c2t40d1s6) devices so the file system can be mounted in the non-global zone. Make sure the block device is not mounted anywhere right before installing the non-global zone. Otherwise, the zone installation may fail withERROR: file system check
zoneadm -z uninstall ) then install the zone from the scratch (no need to re-configure the zone, just do a re-install).
\* set type=ufsThe file system is of type UFS \*
set options noforcedirectioMount the file system with the option
noforcedirectio[/Added 08/25/08]\*
dir=/opt/cswread-only path, will be lofs'd (loop back mounted) from global zone. Note: it works for sparse root zone only -- whole root zone cannot have any shared file systems
zonecfg commands verify and commit,
verifies and commits the zone configuration for the zone,
respectively. Note that it is not necessary to commit the zone
configuration; it will be done automatically when we exit from zonecfg tool. info displays information about the current configuration% zoneadm list -cv ID NAME STATUS PATH 0 global running / - appserv configured /zones/appserver
% zoneadm -z appserv install /zones must not be group writable. could not verify zonepath /zones/appserver because of the above errors. zoneadm: zone appserv failed to verify % ls -ld /zones drwxrwxr-x 3 root root 512 Feb 17 12:46 /zonesSince /zones must not be group writable, let's change the mode to 700.
% chmod 700 /zones % ls -ld /zones drwx------ 3 root root 512 Feb 17 12:46 /zones % zoneadm -z appserv install Preparing to install zone . Creating list of files to copy from the global zone. Copying <2658> files to the zone. Initializing zone product registry. Determining zone package initialization order. Preparing to initialize <1128> packages on the zone. Initialized <1128> packages on zone. Zone is initialized. Installation of these packages generated errors: Installation of <2> packages was skipped. Installation of these packages generated warnings:The file contains a log of the zone installation.
appserv zone, one more time
% zoneadm list -cv ID NAME STATUS PATH 0 global running / - appserv installed /zones/appserver
appserv zone. Let's note down the
ifconfig output to see how it changes after the local zone boots up.
Also observe that there is no answer from the server yet, since it is
not up
% ping 192.168.175.126 no answer from 192.168.175.126 % ifconfig -a lo0: flags=2001000849 mtu 8232 index 1 inet 127.0.0.1 netmask ff000000 eri0: flags=1000843 mtu 1500 index 2 inet 192.168.74.217 netmask fffffe00 broadcast 192.168.75.255 ether 0:3:ba:2d:0:84 % zoneadm -z appserv boot zoneadm: zone 'appserv': WARNING: eri0:1: no matching subnet found in netmasks(4) for 192.168.175.126; using default of 255.255.0.0. % zoneadm list -cv ID NAME STATUS PATH 0 global running / 1 appserv running /zones/appserver % ping 192.168.175.126 192.168.175.126 is alive % ifconfig -a lo0: flags=2001000849 mtu 8232 index 1 inet 127.0.0.1 netmask ff000000 lo0:1: flags=2001000849 mtu 8232 index 1 zone appserv inet 127.0.0.1 netmask ff000000 eri0: flags=1000843 mtu 1500 index 2 inet 192.168.74.217 netmask fffffe00 broadcast 192.168.75.255 ether 0:3:ba:2d:0:84 eri0:1: flags=1000843 mtu 1500 index 2 zone appserv inet 192.168.175.126 netmask ffff0000 broadcast 192.168.255.255Observe that the zone
appserv has it's own virtual instance of lo0, the system's loopback interface and the zone's IP address is also being served by the eri0 network interfacezlogin
utility can be used to enter a zone. The first time we log in to the
console, we get a chance to answer a series of questions for the
desired zone configuraton. -C option of zlogin can be used to log in to the Zone console.
% zlogin -C -e [ appserv
[Connected to zone 'appserv' console]
Select a Language
0. English
1. es
2. fr
Please make a choice (0 - 2), or press h or ? for help: 0
Select a Locale
0. English (C - 7-bit ASCII)
1. Canada (English) (UTF-8)
2. Canada-English (ISO8859-1)
3. U.S.A. (UTF-8)
4. U.S.A. (en_US.ISO8859-1)
5. U.S.A. (en_US.ISO8859-15)
6. Go Back to Previous Screen
Please make a choice (0 - 6), or press h or ? for help: 0
...
Enter the host name which identifies this system on the network. The name
must be unique within your domain; creating a duplicate host name will cause
problems on the network after you install Solaris.
A host name must have at least one character; it can contain letters,
digits, and minus signs (-).
Host name for eri0:1 appserv v440appserv
...
...
System identification is completed.
...
rebooting system due to change(s) in /etc/default/init
[NOTICE: Zone rebooting]
SunOS Release 5.11 Version snv_23 64-bit
Copyright 1983-2005 Sun Microsystems, Inc. All rights reserved.
Use is subject to license terms.
Hostname: v440appserv
v440appserv console login: root
Password:
Feb 17 15:15:30 v440appserv login: ROOT LOGIN /dev/console
Sun Microsystems Inc. SunOS 5.11 snv_23 October 2007
%
Mounting file systems in a non-global zone
Sometimes it might be necessary to export file systems or create new file systems when the zone is already running. This section's focus is on exporting block devices and the raw devices in such situations i.e., when the local zone is already configured.
Exporting the Raw Device(s) to a non-global zone
If the file system does not exist on the device, raw devices can be exported as they are, so the file system can be created inside the non-global zone using the normal
newfs command.The following example shows how to export the raw device to a non-global zone when the zone is already configured.
# zonecfg -z appserv zonecfg:appserv> add device zonecfg:appserv:device> set match=/dev/rdsk/c5t0d0s6 zonecfg:appserv:device> end zonecfg:appserv> verify zonecfg:appserv> commit zonecfg:appserv> exitIn this example
/dev/rdsk/c5t0d0s6 is being exported.After the zonecfg step, reboot the non-global zone to make the raw device visible inside the non-global zone. After the reboot, check the existence of the raw device.
# hostname v440appserv # ls -l /dev/rdsk/c5t0d0s6 crw-r----- 1 root sys 118, 126 Aug 27 14:33 /dev/rdsk/c5t0d0s6Now that the raw device is accessible within the non-global zone, we can use the regular Solaris commands to create any file system like UFS.
eg.,
# newfs -v c5t0d0s6 newfs: construct a new file system /dev/rdsk/c5t0d0s6: (y/n)? y mkfs -F ufs /dev/rdsk/c5t0d0s6 1140260864 -1 -1 8192 1024 251 1 120 8192 t 0 -1 8 128 n Warning: 4096 sector(s) in last cylinder unallocated /dev/rdsk/c5t0d0s6: 1140260864 sectors in 185590 cylinders of 48 tracks, 128 sectors 556768.0MB in 11600 cyl groups (16 c/g, 48.00MB/g, 5824 i/g) super-block backups (for fsck -F ufs -o b=#) at: 32, 98464, 196896, 295328, 393760, 492192, 590624, 689056, 787488, 885920, Initializing cylinder groups: ............................................................................... ............................................................................... ......................................................................... super-block backups for last 10 cylinder groups at: 1139344160, 1139442592, 1139541024, 1139639456, 1139737888, 1139836320, 1139934752, 1140033184, 1140131616, 1140230048Exporting the Block Device(s) to a non-global zone
If the file system exists on the device, block devices can be exported as they are, so the file system can be mounted inside the non-global zone using the normal Solaris command,
mount.The following example shows how to export the block device to a non-global zone when the zone is already configured.
# zonecfg -z appserv zonecfg:appserv> add device zonecfg:appserv:device> set match=/dev/dsk/c5t0d0s6 zonecfg:appserv:device> end zonecfg:appserv> verify zonecfg:appserv> commit zonecfg:appserv> exitIn this example
/dev/dsk/c5t0d0s6 is being exported.After the zonecfg step, reboot the non-global zone to make the block device visible inside the non-global zone. After the reboot, check the existence of the block device; and mount the file system within the non-global zone.
# hostname v440appserv # ls -l /dev/dsk/c5t0d0s6 brw-r----- 1 root sys 118, 126 Aug 27 14:40 /dev/dsk/c5t0d0s6 # fstyp /dev/dsk/c5t0d0s6 ufs # mount /dev/dsk/c5t0d0s6 /mnt # df -h /mnt Filesystem size used avail capacity Mounted on /dev/dsk/c5t0d0s6 535G 64M 530G 1% /mntMounting a file system from the global zone into the non-global zone
Sometimes it is desirable to have the flexibility of mounting a file system in the global zone or non-global zone on-demand. In such situations, rather than exporting the file systems or block devices into the non-global zone, create the file system in the global zone and mount the file system directly from the global zone into the non-global zone. Make sure to unmount that file system in the global zone if mounted, before attempting to mount it in the non-global zone.
eg., In the non-global zone:
# mkdir /repo1In the global zone:
# df -h /repo1 /dev/dsk/c2t40d0s6 134G 64M 133G 1% /repo1 # umount /repo1 # ls -ld /zones/appserv/root/repo1 drwxr-xr-x 2 root root 512 Aug 27 14:45 /zones/appserv/root/repo1 # mount /dev/dsk/c2t40d0s6 /zones/appserv/root/repo1Now go back to the non-global zone and check the mounted file systems.
# hostname v440appserv # df -h /repo1 Filesystem size used avail capacity Mounted on /repo1 134G 64M 133G 1% /repo1To unmount the file system from the non-global zone, run the following command from the global zone.
# umount /zones/appserv/root/repo1Removing the file system from the non-global zone
eg., Earlier in the zone creation step, the block device
/dev/dsk/c2t40d1s6 was exported and mounted on the mount point /repo2
inside the non-global zone. To remove the file system completely from
the non-global zone, run the following in the global zone.# zonecfg -z appserv
zonecfg:appserv> remove fs dir=/repo2
zonecfg:appserv> verify
zonecfg:appserv> commit
zonecfg:appserv> exit
Reboot the non-global zone for this setting to take effect.Shutting down and booting up the local zones (Updated 01/15/2008)
- To bring down the local zone:
% zlogin appserv shutdown -i 0
- To boot up the local zone:
% zoneadm -z appserv boot
Steps to delete a Local Zone
- Shutdown the local zone
% zoneadm -z appserv halt % zoneadm list -cv ID NAME STATUS PATH 0 global running / - appserv installed /zones/appserver
- Uninstall the local zone -- remove the root file system
% zoneadm -z appserv uninstall Are you sure you want to uninstall zone appserv (y/[n])? y zoneadm list -cv ID NAME STATUS PATH 0 global running / - appserv configured /zones/appserver
- Delete the configured local zone
% zonecfg -z appserv delete Are you sure you want to delete zone appserv (y/[n])? y zoneadm list -cv ID NAME STATUS PATH 0 global running /
The following instructions are for cloning a non-global zone on the same system. The example shown below clones the
siebeldb zone. After the cloning process, a brand new zone oraclebi emerges as a replica of siebeldb zone.eg.,
# zoneadm list -cv ID NAME STATUS PATH BRAND IP 0 global running / native shared - siebeldb installed /zones/dbserver native excl
- Export the configuration of the zone that you want to clone/copy
# zonecfg -z siebeldb export > /tmp/siebeldb.config.cfg
- Change the configuration of the new zone that differ from
the existing one -- for example, IP address, data set names, network
interface etc. To make these changes, edit
/tmp/siebeldb.config.cfg
- Create the zone root directory for the new zone being created
# mkdir /zones3/oraclebi # chmod 700 /zones3/oraclebi # ls -ld /zones3/oraclebi drwx------ 2 root root 512 Mar 12 15:41 /zones3/oraclebi
- Create a new (empty, non-configured) zone in the usual manner with the edited configuration file as an input
# zonecfg -z oraclebi -f /tmp/siebeldb.config.cfg # zoneadm list -cv ID NAME STATUS PATH BRAND IP 0 global running / native shared - siebeldb installed /zones/dbserver native excl - oraclebi configured /zones3/oraclebi native excl
- Ensure that the zone you intend to clone/copy is not running
# zoneadm -z siebeldb halt
- Clone the existing zone
# zoneadm -z oraclebi clone siebeldb Cloning zonepath /zones/dbserver...
This step takes at least 5 minutes to clone the whole zone. Larger zones may take longer to complete the cloning process.
- Boot the newly created zone
# zoneadm -z oraclebi boot
Bring up the halted zone (the source zone) as well, if wish.
- Login to the console of the new zone to configure IP, networking, etc., and you are done.
# zlogin -C oraclebi
Keywords: Solaris, Non-Global Zone, Migration, Attach, Detach
The following instructions demonstrate how to migrate the non-global zone,
orabi to another server with examples.# zoneadm list -cv ID NAME STATUS PATH BRAND IP 0 global running / native shared 4 siebeldb running /zones/dbserver native excl - orabi installed /zones3/orabi native shared
- Halt the zone to be migrated, if running
# zoneadm -z orabi halt
- Detach the zone. Once detached, it will be in the configured state
# zoneadm -z orabi detach # zoneadm list -cv ID NAME STATUS PATH BRAND IP 0 global running / native shared 4 siebeldb running /zones/dbserver native excl - orabi configured /zones3/orabi native shared
- Move the
zonepathfor the zone to be migrated from the old host to the new host. Do the following on the old host:
# cd /zones3 # tar -Ecf orabi.tar orabi # compress orabi.tar # sftp newhost Connecting to newhost... sftp> cd /zones3 sftp> put orabi.tar.Z Uploading orabi.tar.Z to /zones3/orabi.tar.Z sftp> quit
On the newhost:
# cd /zones3 # uncompress orabi.tar.Z # tar xf orabi.tar
- On the new host, configure the zone.
Create the equivalent zone
orabion the new host -- use thezonecfgcommand with the-aoption and thezonepathon the new host. Make any required adjustments to the configuration and commit the configuration.
# zonecfg -z orabi orabi: No such zone configured Use 'create' to begin configuring a new zone. zonecfg:orabi> create -a /zones3/orabi zonecfg:orabi> info zonename: orabi zonepath: /zones3/orabi brand: native autoboot: false bootargs: pool: limitpriv: all,!sys_suser_compat,!sys_res_config,!sys_net_config,!sys_linkdir,!sys_devices,!sys_config,!proc_zone,!dtrace_kernel,!sys_ip_config scheduling-class: ip-type: shared inherit-pkg-dir: dir: /lib inherit-pkg-dir: dir: /platform inherit-pkg-dir: dir: /sbin inherit-pkg-dir: dir: /usr net: address: IPaddress physical: nxge1 defrouter not specified zonecfg:orabi> set capped-memory zonecfg:orabi:capped-memory> set physical=8G zonecfg:orabi:capped-memory> end zonecfg:orabi> commit zonecfg:orabi> exit
- Attach the zone on the new host with a validation check and
update the zone to match a host running later versions of the dependent
packages
# ls -ld /zones3 drwxrwxrwx 5 root root 512 Jul 15 12:30 /zones3 # chmod g-w,o-w /zones3 # ls -ld /zones3 drwxr-xr-x 5 root root 512 Jul 15 12:30 /zones3 # zoneadm -z orabi attach -u Getting the list of files to remove Removing 1740 files Remove 607 of 607 packages Installing 1878 files Add 627 of 627 packages Updating editable files The file within the zone contains a log of the zone update. # zoneadm list -cv ID NAME STATUS PATH BRAND IP 0 global running / native shared - orabi installed /zones3/orabi native shared
Note:
It is possible to force the attach operation without performing the validation. You can do so with the help of-Foption
# zoneadm -z orabi attach -F
Be careful when using this option because it could lead to an incorrect configuration; and an incorrect configuration could result in undefined behavior
If the command
zonename returns global, then you are connected to the OS instance that was booted from the physical hardware. If you see any string other than global, you might have connected to the virtual OS instance.Alternatively try running
prstat -Z or zoneadm list -cv commands. If you see exactly one non-zero Zone ID, it is an indication that you are connected to a non-global zone.
No comments: