/tmp/hacmp.out

[ereshatezer]

Hello,

I have the following question concerning a HACMP cluster.

in the /tmp/hacmp.out I found the following entry:

"config_too_long[91] echo WARNING: Cluster xxxxxx has been running recovery program '/usr/es/sbin/cluster/events/reconfig_resour
ce.rp' for 2573010 seconds.\n Please check cluster status."

I did clverify but found no problem.
Has anyone met this problem before?

The size of the hacmp.out file is over 100MB now.
I know that "config_too_long" event occures in case longer running of a script than 6minutes, but i found no problem in the cluster and it runs fine.

Could anyone help how to stop appearing that message in hacmp.out logfile and what can cause this?

Thank you in advance!

 

[stefanhei]

Hi

smitty hacmp -> Problem Determination Tools -> Recover From HACMP Script Failure

should stop the message.
Start and endtime for each event can be found in the log - you may take them as a starting point for further analysis.

Stefan

 

[DSMARWAY]

hi ,

firstly check why your start/stop scripts are taking along time to run . There could be a problem with one of the scripts , if possible try running your scripts manually.

The other possibility is that your application takes more than 6 minutes to stop/run in which case try increasing this parameter e.g by 10 minutes

chssys -s clstrmgr -a "-u 600000" ( thats miliseconds i.e. 600 seconds which is 10 minutes

You can kill the process config_too_long , becuase it has probably hung. Once other thing to check in your hacmp.out is event errors i.e. search for ERORR in upppercase and see
if any show up .

HTH

 

[stefanhei]

DSMARWAY gave you some valuable information, but I'd rather start analysis with your current hacmp.out.
Last time I've seen this message was due to powerpath sending powerresets to disks after the box was rebooted - no error or script involved (reason was inadequate microcode on FC adapters).
Locate in which section the message appears first. If it's the scripts section follow DSMARWAY's advice.

Stefan