Error Message

vcs1: # hagrp -online ClusterService -sys vcs1
VCS WARNING V-16-1-10165 Failover group ClusterService is not completely offline in cluster

CurrentCount shows number of nodes having online the service group, and it can be incorrectly set due to some offline service group failure.  Once this attribute is set, further group online will be failed.  You can verify the issue by looking below output:

# hagrp -display ClusterService |grep CurrentCount

ClusterService                          CurrentCount          global      1

If the service group is OFFLINE on all nodes, it should be 0.

Note that "hastop -local -force" will not work for multiple nodes as CurrentCount will remain non ZERO

5.1SP1RP4 and onwards changed code, and 'hagrp -flush' can clear the CurrentCount.  Old version doesn't have any command to clear the value.  So, only way to fix the issue restart cluster on all nodes with following steps:

1) stop the cluster on all nodes

# hastop -all -force

2) Make sure none of nodes are running HAD

# ps -ef|grep had

<< No 'had' process running >>

# gabconfig -a

<< port 'h' should not be appeared >>

3) restart cluster one by one

# hastart

4) check the CurrentCount

# hagrp -display ClusterService |grep CurrentCount

<< it should be 0 >>

In case, the issue happened due to Concurrency Violation, below HF will modify the CurrentCount correctly:

vcs-aix-6.0.5.001

Applies To

AIX6.1 / VCS5.1RP2