Closed gdha closed 4 years ago
Noticed in the log the following:
2019-07-10 05:34:02 Including prep/hp/43_check_perfalarm.sh
2019-07-10 05:34:08 WARNING: /var/opt/perf/alarmdef contains errors - see below
2019-07-10 05:34:08 Including prep/hp/45_check_serviceguard.sh
2019-07-10 05:34:09 Cluster CLRAUX004 (version A.11.20.00.01 - latest patch PHSS_44549) is up
The alarmdef.err file was not cat-ted in the log, albeit the file exists.
in script scripts/prep/hp/43_check_perfalarm.sh:
cat $VAR_DIR/$DS/alarmdef.err >> $LOGFILE
should become
cat $VAR_DIR/$DS/alarmdef.err >&2
Nice command overview of OV related stuff: http://margaux99.blogspot.com/2012/01/hp-ux-mwa-perfalarm.html
/opt/perf/bin/ovpa restart
perfalarm was still stopped.
From "/var/opt/perf/status.perfalarm" saw the issue: Metric 'PROC_DISK_PHYS_IO' is undefined. ("/var/opt/perf/alarmdef", near line - 170)
/opt/perf/bin/perfalarm : 05/21/19 04:21:16 ERROR: Errors found in the alarmdef file. (AG1-53)
/opt/perf/bin/perfalarm : 05/21/19 04:21:16 ERROR: The above error may indicate that there is a communication problem between the perfalarm process and the coda process respresenting the data source. (AG1-80) The OVO Message Command is not configured on this system. Contact your OVO Administrator to configure it. (OpC30-913)
Perhaps we should parse the alarmdef as a pre-task before patching, so we cannot be blamed that the perfalarm daemon is not started after the reboot (due to the patching).