Perform the following procedure to upgrade the second subcluster (node03 and node04).
To upgrade the second subcluster
Take the service groups offline on node03 and node04.
Warning: |
Failure to perform this step can cause data corruption if you try to bring the failover groups (sg3 and sg4) online on the upgraded subcluster (node01 and node02) |
# hagrp -offline sg1 -sys node03 # hagrp -offline sg1 -sys node04 # hagrp -offline sg2 -sys node03 # hagrp -offline sg2 -sys node04 # hagrp -offline sg3 -sys node03 # hagrp -offline sg4 -sys node04
Stop VCS and its components on node03 and node04.
# /sbin/init.d/vcs stop # /sbin/init.d/vxfen stop # /sbin/gabconfig -U # /sbin/lltconfig -Uo
Edit the following files on node03 and node04 to make the values of LLT_START, GAB_START and VXFEN_START equal to zero. (By default it is 1).
# /etc/rc.config.d/lltconf
# /etc/rc.config.d/gabconf
# /etc/rc.config.d/vxfenconf
Performing this step prevents LLT, GAB, and VxFEN from restarting if the nodes reboot accidentally. If node03 and node04 reboot accidentally, it may cause node01 and node02 to panic.
On node01 and node02, start VCS and its components.
# /sbin/init.d/llt start # /sbin/init.d/gab start # /sbin/gabconfig -cx # /sbin/init.d/vxfen start # /sbin/init.d/vcs start
Make the VCS configuration writable on node01 or node02. Note that you need to make the configuration writable because the installer froze the service groups during the upgrade.
# haconf -makerw
Unfreeze all the service groups node01 and node02.
# hagrp -unfreeze sg1 -persistent # hagrp -unfreeze sg2 -persistent # hagrp -unfreeze sg3 -persistent # hagrp -unfreeze sg4 -persistent
Dump the configuration and make it read-only.
# haconf -dump -makero
On node01 and node02, bring the service groups online.
# hagrp -online sg1 -sys node01 # hagrp -online sg1 -sys node02 # hagrp -online sg2 -sys node01 # hagrp -online sg2 -sys node02 # hagrp -online sg3 -sys node01 # hagrp -online sg4 -sys node02
Verify the state of the service groups.
# hagrp -state
The output resembles the following:
#Group Attribute System Value sg1 State node01 |ONLINE| sg1 State node02 |ONLINE| sg1 State node03 |OFFLINE| sg1 State node04 |OFFLINE| sg2 State node01 |ONLINE| sg2 State node02 |ONLINE| sg2 State node03 |OFFLINE| sg2 State node04 |OFFLINE| sg3 State node01 |ONLINE| sg3 State node02 |OFFLINE| sg3 State node03 |OFFLINE| sg3 State node04 |OFFLINE| sg4 State node01 |OFFLINE| sg4 State node02 |ONLINE| sg4 State node03 |OFFLINE| sg4 State node04 |OFFLINE|
Upgrade VCS on node03 and node04. Follow the procedure mentioned in the section for upgrading the first subcluster.
After the upgrade completes, reboot node03 and node04.
After you have rebooted the nodes, all the four nodes now run the latest version of VCS, which is 5.0.1.
In this example, you have performed a phased upgrade of VCS because the service groups were down only from the point when you took them offline on node01 and node02, to the time VCS brought them online on node03 or node04 as appropriate.