clstat displays multiple clusters, with a fake ID 0

You might encounter this little problem when executing clstat -o, which says that you have multiple clusters configured and you need to execute clstat -i, to select the cluster you want to display:

# clstat

There is more than one cluster accessible.
Either provide the cluster to look for, or run
the program interactively (or try again).

usage: clstat [-c cluster ID | -n cluster_name] [-i] [-r seconds] [-a|-o] [-s]

-c cluster ID > run in automatic (non-interactive) mode for the
specified cluster.
-n name       > run in automatic (non-interactive) mode for the
specified cluster name.
-i            > run in interactive mode
-r seconds    > number of seconds between each check of the
cluster status
-a            > run ascii version.
-o            > run once and exit.
-s            > display both up and down service labels.

# clstat -i
clstat - HACMP Cluster Status Monitor
-------------------------------------

Number of clusters active: 2
ID      Name            State

1589772197 cluster1              UP
         0 cluster1              UP

Here, we have a cluster with an ID=0, which is not to be found everywhere else.

At first I thought it was a cluster misconfiguration.
Actually it’s no big deal (unless you have 2 clusters with the same ID, which is not my case here), it seems to be a small bug with the clinfo daemon, you just need to restart it to erase the problem.

# stopsrc -s clinfoES && startsrc -s clinfoES
0513-044 The clinfoES Subsystem was requested to stop.
0513-059 The clinfoES Subsystem has been started. Subsystem PID is 27394194.

Now we only see one cluster, everything is in order :

# clstat -i
clstat - HACMP Cluster Status Monitor
-------------------------------------

Number of clusters active: 1
ID      Name            State

1589772197 cluster1              UP

Same here, but with real **actual** clusters displayed

This is more tricky,  I had 2 clusters that « saw » another cluster state… Like this one :

 # clstat -i

clstat - HACMP Cluster Status Monitor
-------------------------------------

Number of clusters active: 3

ID      Name            State

1636674608 cluster_1                UP
1590729639 other_cluster            UP
1213723326 another_unknown_cluster  UP

Select an option:

Wow… this one is seeing 2 more clusters, which aren’t affiliated to it… It has to be a snmp-related problem.

Let’s follow this post and try to spot and fix the errors:

# netstat -Aan |grep -w 199

#

–> it seems that the smux daemon is down, we’ll have to restart the snmpd related services,in order to get it all back to normal :

#  stopsrc -s clinfoES
 #  stopsrc -s aixmibd
 #  stopsrc -s snmpmibd
 #  stopsrc -s hostmibd
 #  stopsrc -s snmpd
 #  startsrc -s snmpd
 #  startsrc -s snmpmibd
 #  startsrc -s clinfoES
 #  refresh -s clstrmgrES

Links :

IV03412: CLSTAT -I INCORRECTLY SHOWS MULTIPLE CLUSTERS WITH ID 0

PowerHA : clstat,cldump don’t work ? Check your SNMP configuration!

Multi-cluster ASCII display mode

Share Button
PowerHA : clstat displays multiple clusters when you only configured one…
Taggé sur :        

Laisser un commentaire