Nagios Support Forum

Hi. I'm writing to see if anyone has worked out a way to monitor the duration of DB2 deadlocks. I will test the check_db2_health script on the exchange, but it monitors for deadlocks per second, rather than the duration.

Thanks!

Putting a monitor on the duration of deadlocks sounds like a great idea. The check_db2_health plugin in the Exchange has a flag for lock-waiting, but lock waits seem like they may be somewhat different from deadlocks...?

I did stumble across this post on Stack Overflow.
https://stackoverflow.com/questions/490 ... ock-in-db2

Specifically, this message.

Yes that means no deadlock. If there was a deadlock and the event monitor was active at the time, then you should see a "Deadlock Event" stanza in the db2evmon output. You can also use the monitoring functions to confirm if a deadlock happened (but not how it happened). The following query reports cumulative values since database-activation:

Code: Select all

SELECT deadlocks, lock_timeouts FROM TABLE (MON_GET_DATABASE(-2));

So it looks like there may be ways to monitor the deadlock event. Maybe they either include duration information, or a way to determine the duration. At which point, you could create a plugin.

Thank you very much! I'll try the method mentioned in the article.

Cool! Do you want to leave this thread open, or should we close it?

We can close it. Thanks

awilson wrote:We can close it. Thanks

Great!

Locking