You are on page 1of 4

HACMP Failover Checklist

Checks before Testing

Manual testing ( using smit panels )

7
The installed lpps( PowerHA) on both nodes must be exactely
the same and at the same level.

Cluster network configuration: the IP @ of the interfaces used


by HACMP must be known to the two nodes and pingable
from the two nodes.

Cluster Resource and Topology Verification ( clverify )

Using smit panals stop clsuter services on node 1, observe


hacmp.out file for events completed state
Using smit panals stop clsuter services on node 2, observe
hacmp.out file for events completed state
Using smit panals start clsuter services on node 1, observe
hacmp.out file for events completed state
Using smit panals start clsuter services on node 2, observe
hacmp.out file for events completed state
Using smit panels stop cluster with takeover option on node 1
to failover resources to node 2
Using smit panels again start cluster on node 1 to takover
prority resources from node 2
Using smit panels stop cluster with takeover option on node 2
to failover resources to node 1
Using smit panels again start cluster on node 2 to takover
prority resources from node 1
1) for adapter in `cllsif | awk '{print $1}' | grep -v
Adapter | grep -v hdiskpower`
do
grep -q $adapter /etc/hosts
if [[ $? != 0 ]]
then
print "Interface $adapter is not defined in /etc/hosts on
`hostname`"
fi
lslpp -l "cluster*" ; lslpp -l "cluster*" | wc -l ; lppchkdone
-v
2) for
adapter in `cllsif | awk '{print $1}' | grep -v Adapter |
grep -v hdiskpower` do
ping
-q -c 1 $adapter
done
clsmuxpdES, clstrmgrES and clinfoES at the active
state. 2) Execute the
following on any node: clfindres
Normally,
you'll find two resource groups: CI_disk_fs and
DB_disk_fs. Check that each resource group is online
on one location and offline on the other location which
means that each resource group is activated in its
home node.
3) execute clstat -o
You must have the following:
-State: must be up

smit clstop with graceful shutdown (node1)

smit clstop with graceful shutdown(node 2)

smity clstart (node 1)

smity clstart (node 2)

smitty clstop ( node 1)

smitty clstart ( node 1 )

smitty clstop ( node 2)

smitty clstop ( node 2)


Remarks

Observed hacmp.out for events taken place and


completed
Observed hacmp.out for events taken place and
completed
Observed hacmp.out for events taken place and
completed
Observed hacmp.out for events taken place and
completed
Observed hacmp.out for events taken place and
completed
Observed hacmp.out for events taken place and
completed
Observed hacmp.out for events taken place and
completed
Observed hacmp.out for events taken place and
completed

You might also like