hastop -sysoffline fails to offline VCS service groups with VCS WARNING V-16-1-10483

Article:TECH181711  |  Created: 2012-02-18  |  Updated: 2012-03-01  |  Article URL http://www.symantec.com/docs/TECH181711
Article Type
Technical Solution

Product(s)

Environment

Issue



Veritas Cluster Server (VCS) stop script calls hastop -sysoffline when stopping the cluster on the local node. This indicates VCS that the system is being brought down intentionally for reboot / shutdown.

Running hastop -sysoffline would fail with the message V-16-1-10483 as mentioned below. None of the service groups running on the node are brought offline and VCS engine would continue to run on the node

 


Error



root@node2 # hastop -sysoffline
VCS WARNING V-16-1-10483 Offlining system node2 would result in group dependency being violated for one of the groups online on node2; offline the parent of such a group first
root@node2 #
 

Environment



This issue is applicable to systems running any supported operating system / platform on VCS 6.0.


Cause



Prior to VCS 6.0, when hastop -sysoffline is issued, VCS behaviour with service groups offline is either NONE or ALL, i.e., all groups are brought offline or none brought offline.

From VCS 6.0, when hastop -sysoffline is issued, VCS checks the value of cluster level attribute SystemRebootAction to determine what action it should take in regards to frozen service groups. If the SystemRebootAction value is IgnoreFrozenGroup, VCS ignores service groups that are frozen (TFrozen and Frozen) and takes the remaining service groups offline. If the frozen service groups have firm dependencies or hard dependencies on any other service groups which are not frozen, VCS gives an error as mentioned above and does not offline ANY service groups running on the node.

If the SystemRebootAction value is "", VCS tries to take all service groups offline. Because VCS cannot be gracefully stopped on a node where a frozen service group is online, applications on the node might get killed. 

In summary, the problem happens under the following scenario:

- VCS version is 6.0 or above

- VCS cluster level attribute SystemRebootAction is set to IgnoreFrozenAction

# haclus -value SystemRebootAction
IgnoreFrozenGroup
#

- A frozen service group has either a online (global / local) hard or firm dependency on an unfrozen service group.

 


Solution



To stop the cluster on the node, one of the following options could be used.

1) Unfreeze the frozen service group so that it could be brought offline and online.

2) Change the hard / firm dependency to a soft dependency.




Article URL http://www.symantec.com/docs/TECH181711


Terms of use for this information are found in Legal Notices