topic Re: Monitor cmviewcl in Operating System - HP-UX

Monitor cmviewcl

Fernando Boza — Mon, 09 Aug 2010 20:10:38 GMT

To check the status of the cluster, just run cmviewcl.
But in certain situation, the cluster command stopped responding.
commands are left hanging without showing output.

as I can monitor with a cron these commands?

Re: Monitor cmviewcl

Kranti Mahmud — Tue, 10 Aug 2010 09:41:41 GMT

Hi Fernando,

Your question is not clear. BTW, you can check the below link: http://www.unix.com/man-page/all/1m/cmviewcl/

Rgds-Kranti

Re: Monitor cmviewcl

rariasn — Tue, 10 Aug 2010 10:04:07 GMT

Hi Fernando:

Please, put you HPUX and SGuard version.

Verify the latest patch for SGuard.

Verify network connections timeout.

Rgs,

Re: Monitor cmviewcl

Matti_Kurkela — Tue, 10 Aug 2010 10:32:38 GMT

When one or more nodes have lost contact with the cluster (e.g. suddenly disconnected from the cluster in an abrupt way), the cmviewcl may take a long time to run (about one minute for each missing node), but it should eventually complete anyway if you wait long enough.

This delay does not happen if a node leaves the cluster in a controlled fashion (using the "cmhaltnode" command).

MK

Re: Monitor cmviewcl

Deeos — Tue, 10 Aug 2010 10:55:06 GMT

Hi,

You should check syslog.log and cluster pkg log to know the root cause of problem occurance.however, you can check /etc/inetd.conf file to see hacl lines for udp/tcp is available or not?

Re: Monitor cmviewcl

nightwich — Tue, 10 Aug 2010 13:07:26 GMT

Hi Fernando

I'am not shore what you need.

If the cmviewcl "block" you must have some problem with your configuration/SG.

Check the syslog and cluster logs and see you can find anything.

Best regards.

Re: Monitor cmviewcl

Fernando Boza — Tue, 10 Aug 2010 14:22:54 GMT

I want to create a script to check every now commands the cluster if they are responding adequately .... and to report only when the cluster commands do not work.

Re: Monitor cmviewcl

Viktor Balogh — Tue, 10 Aug 2010 22:51:16 GMT

Hi Fernando,

>I want to create a script to check every now commands the cluster if they are responding adequately .... and to report only when the cluster commands do not work.

For what is it good? You should rather check your services, which are the critical part in the business. I think you want to check if the packages are running, but there are many other ways to check this, like a remote ssh to the package IP, and checking with ps if the application behind it runs, and so on...