[ClusterLabs] Example of somewhat confusing log message from crmd

Ulrich Windl Ulrich.Windl at rz.uni-regensburg.de
Thu Aug 9 10:50:12 UTC 2018


Hi!

While setting up a new clustered VG in SLES11 SP4 (create it, then configure it in the cluster), I noticed this warning message, which is a good example how log messages could be improved:

Aug  9 12:19:41 h10 crmd[20946]:  warning: status_from_rc: Action 58 (prm_LVM_VMD:0_monitor_0) on h05 failed (target: 7 vs. rc: 0): Error

1) In the front the message says "warning:", and in the back the message says ": Error"; what now?
2) "Action 58 ... failed": what was the action? "start", "monitor"? (from the rc it's obvious to the gurus that it must have been "monitor". OK, one could also deduce from the ID "_monitor_0"...)
3) "7 vs. 0": Only few people starting with the cluster know the magic of these values...

So in summary, the message I'd like to see is like:
"Aug  9 12:19:41 h10 crmd[20946]:  warning: the monitor status of prm_LVM_VMD on h05 indicates that the resource is unexpectedly running (before having started it)"

or:

"Aug  9 12:19:41 h10 crmd[20946]:  notice: prm_LVM_VMD on h05 was expected to be stopped, but it was found to be running"
(Note that it's less than a warning now)

Maybe some "info" message saying "accepting status as started for resource prm_LVM_VMD on h05" would be OK also, because that is what the cluster does:

Aug  9 12:19:41 h10 crmd[20946]:   notice: te_rsc_command: Initiating action 98: monitor prm_LVM_VMD_monitor_300000 on h05

One of the problems is that log messages should be useful by themselves, i.e.: not having to read the ten preceeding messages, or so.

I know it's hard to make everybody happy, but maybe in the next 30 years... ;-)

Regards,
Ulrich




More information about the Users mailing list