Video Screencast Help

Freezing a node and service group

Created: 29 Dec 2010 • Updated: 03 Jan 2011 | 1 comment
This issue has been solved. See solution.

Hi All ,

Could you please explain what is dfference between freezing a node and freezing a service group.



Comments 1 CommentJump to latest comment

g_lee's picture


From VCS 5.1 Administrator's Guide:
Administering Groups -> Freezing and unfreezing service groups
Freeze a service group to prevent it from failing over to another system. This freezing process stops all online and offline procedures on the service group.
Unfreeze a frozen service group to perform online or offline operations on the service group

Administering systems -> Freezing a system
Freeze a system to prevent service groups from coming online on the system.

Additional information from hasys manpage (link for VCS 5.1 Solaris - ):
-freeze [-persistent] [-evacuate] system
Freeze a system. No group configured on the frozen system can come online, whether manually, by failover, or by switching until the system is thawed with the -unfreeze option. Using the -persistent option specifies that the system remains frozen after VCS has been stopped and restarted. Note that the configuration must be writable in order to use -persistent option. Groups online on the frozen system can be taken offline manually, by failover, or by being switched to another system. Using the -evacuate option specifies that all groups are switched before the system is frozen; if no other system is available for a service group, it is taken offline. Groups running on other systems do not fail over to a frozen system.

So - if you freeze a service group, it cannot be switched to another node in VCS until it is unfrozen. If any of the resources in the service group fault, VCS will not take any action (eg: even if that fault would normally trigger a failover if the group was unfrozen).

If you freeze a node/system, this just prevents service groups from being brought online on the node. Existing service groups will continue to run on the system, however if a critical resource faults, this will trigger failover to another system.


- two node cluster, consisting of nodeA, nodeB
- two service groups, sgA (running on nodeA), sgB (running on nodeB)
- say both sgA and sgB have Critical Network resources NetworkA, NetworkB respectively.

Say you need to perform maintenance on the network interface on nodeA that requires you to take the interface down. Under normal conditions (ie: without freezing anything), this would cause the NetworkA resource to fault, and would trigger failover (sgA would failover to nodeB)

If you freeze service group sgA, and then take the network interface down on nodeA, although VCS will still recognise the NetworkA resource as faulted, it won't failover/bring up the group on nodeB, as the freeze tells VCS not to perform any offline/online actions for that service group.

However, if you freeze the system nodeA, then take down the interface, VCS will see the NetworkA resource as faulted on nodeA, and so will try to failover to nodeB, as you've only told it not to bring anything online on nodeA (ie: it will still register fault, and take groups offline on nodeA if required).

Also, if you performed network maintenance on nodeB while nodeA was still frozen, VCS would not failover sgB to nodeA, as freezing the node means VCS won't bring groups online on the frozen nodeA.

Hope that helps. If you need further clarification, please advise which details are not clear.



If this post has helped you, please vote or mark as solution