Error in osspi_chktab file (not creating) - Alarmdef file not creating

[Brandy Reid]

I'm having this same issue on multiple HP-UX 11.11 systems running the OVO agent and OVPA. I'm using OVOW to monitor this system.

I keep getting error messages like the ones listed below:

MeasureWare: Recreation of alarms failed
Node:
Message group: OSSPI-Internal
Application: HP OSSPI
Object: osspi_chkalarm.sh
Severity: Critical
Text: MeasureWare: Recreation of alarms failed

and

syntax error in osspi_alarmdef file
Node:
Message group: OSSPI-Internal
Application: HP OSSPI
Object: osspi_alarm_mw.sh
Severity: Major
Text: syntax error in osspi_alarmdef file

The contents of the directory is:

[/var/opt/OV/conf/osspi]
#ll
total 288
fsmon.cfg.chktab
harddisk.cfg
netif.cfg
opcagtinfo
osspi.cfg
osspi_alarmdef.chk
osspi_fsmon-hash.cfg
osspi_fsmon.cfg
procmon.cfg
syntax_gp
syntax_gp.old

Any Ideas?

Thanks,
Brandy

P.S. This thread has been moved from Systems Management (OpenView-OP Mgmt) Support and News Forum to OV Performance Management Forum - Hp Forum Moderator

[Tony Cicone]

It looks like ovpa is working now. Just check the status to make sure.

/opt/perf/bin/ovpa status

If you removed the alarmdef_write templates, then you should be able to do a distribute and not get the original errors.

[Tony Cicone]

Page 71 explains how the osspi_alarmdef and alarmdef file work together.

http://ovweb.external.hp.com/ovnsmdps/pdf/UNIXOS-SPI_a35_AdminRef.pdf

I think you had 2 issues.

1. The fix for osspi_alarmdef was needed.

2. No /var/opt/perf/alarmdef file, so you weren't monitoring with OVPA, you were just collecting data.

In my environment, I have customized our alarmdef file and we are using that to monitor peformance metrics. I'm only using the OSSPI for discovery, and process monitoring (ie procmon.cfg, etc). I have removed the alarmdef_write template, so that osspi_alarmdef does not get recreated after each distribute.

[Brandy Reid]

The version of OVPA is 4.5

[Bill Costigan]

I think the problem was that data for
/mnt/winunix wasn't being collected by ovpa unitl you restarted ovpa.

Therefore the metric
/mnt/winunix:FS_SPACE_UTIL was not valid when the alarmdef file was verified.

[Brandy Reid]

On the HP-UX node, I went through the following steps to get the alarmdef file to update correctly:

1. Killed the OVO agent. /opt/OV/bin/OpC/opcagt -kill

2. Stopped the MWA agent. /opt/perf/bin/mwa -stop

3. Deleted the osspi_alarmdef file. /var/opt/OV/conf/osspi/osspi_alarmdef

4. Deleted the osspi_alarmdef.chk file. /var/opt/OV/conf/osspi/osspi_alarmdef.chk

5. Removed all policies from the node in the management console except for
OSSPI-alarmdef_write
OSSPI-alarmwriteMsg
OvSvcDiscErrorLog
UNIX_OSSPI-AutoDiscovery
UNIX_OSSPI-opcmsg

6. Ran a opcagt -start. /opt/OV/bin/OpC/opcagt -start

7. Ran a mwa -start. /opt/perf/bin/mwa -start

8. Ran a opcagt -status (all processes are running)

9. Ran a mwa -status (all process are running)

10. Deployed OSSPI-MWA_Cpu_Load to the node through the management console

11. Ran a cat osspi_alarmdef and the file had been updated correctly.

12. Deployed other OSSPI-MWA policies and the osspi_alarmdef file updated correctly.

13. Deployed the OSSPI-MWA_Filesystem policy and a major alert popped up in the management console:

syntax error in osspi_alarmdef file

Node:
Message group: OSSPI-Internal
Application: HP OSSPI
Object: osspi_alarm_mw.sh
Severity: Major
Text: syntax error in osspi_alarmdef file

I'm starting a new thread to work on the OSSPI-MWA_Filesystem policy issue.

Thanks for all your help,
Brandy

[Jeena Me]

Use a small tool available at http://pathtoodeep.com ,it will resolve this error,you cannot do that manually