Technical Blog Post
Abstract
Find out RecoveryRM Master Node
Body
In TSA cluster environment, RecoveryRM daemon decides the automated action to take.
You sometimes need to read its trace to diagnose issues. While RecoveryRM daemon runs
on all hosts in the cluster, it is the ‘Master’ RecoveryRM that serves as the "brain" or the
decision engine for TSA. Therefore before you start investigating the trace, you first need to
determine which node is the master node among the participating nodes in pureScale cluster.
Scenario 1 - Find out RecoveryRM master node in a live system
lssrc -ls IBM.RecoveryRM | grep Master
Master Node Name : hostD (node number = 4)
On hostD:
rpttr -o dtic /var/ct/<domain_name>/log/mc/IBM.RecoveryRM/trace_summary*
Note that root access is required to view these traces
Scenario 2 - Find out RecoveryRM master node from db2support.zip collected
1. Move to a working directory and extract db2support.zip
2. Search for master node in the trace collected
$ cat /home/hbwang/help/rmm
for i in `find . -name "RecoveryRM.*.trace_summary*" -print`; do
echo $i
grep "Master daemon" $i|tail -1
echo
done
$ rmm
./PURESCALE/CM/sa_data/101616_013015-node01-1/traces/RecoveryRM.db2domain_20160101174930.trace_summary.sp.out
[07] 10/13/16 20:41:35.346634 T(708) _RCD CHARM::processMembership() Master daemon node name - node02, NodeNum = 2
./PURESCALE/CM/101616_013015-node01-1/traces/RecoveryRM.db2domain_20160101174930.trace_summary.sp.out
[07] 10/13/16 20:41:35.346634 T(708) _RCD CHARM::processMembership() Master daemon node name - node02, NodeNum = 2
The above tells the master node is node02. Therefore you need to review RecoveryRM trace files on node02.
UID
ibm11140646