2014-01-31 10:39 AM
I've completed an upgrade/migration from OCUM 5.1 on Windows to OCUM5.2 on Linux - installed 5.2, backed up the DB on the 5.1 instance, restored it to the 5.2 instance, all seemed to work OK. Looks like all the settings came over, it seems to have picked up monitoring (mostly), all looks good.
I noticed that not all the events came over - the 5.1 instance shows 5 errors and 44 warnings, the 5.2 instance shows 1 and 26. At first I thought that old events from the 5.1 version that were no longer in effect (volumes full alerts that been resolved, for example) had been excluded, but that doesn't seem to be the case. The 5.1 version shows a volume that is 93% full, and shows the current warning alert of volume almost full (triggered 29 Jul 2013). The 5.2 instance shows the same volume at the same utilization, but doesn't list any alerts - so not only did the 5.1 alert not carry over in the migration, the 5.2 version has not noticed that the volume is above the warning threshhold and generated its own alert.
Checking the age of the oldest alerts, the 5.2 version goes to 12 Aug 2013. The 5.1 version goes to 31 Aug 2012, but before the 12 Aug 2013 alerts, the next oldest is 29 July - I restored the DB on 29 Jan, so did it's possible it just dropped everything older than 6 months? Even if that was the case, shouldn't it re-alert on the events that are still in effect?
2014-02-02 09:10 PM
As part of upgrade to OCUM 5.2, there is lot of database maintenance being done, I am sure you would have seen this messages during your restore operations.
You are right with the Events part, starting version 5.2 and later, all event older than 6 months or the value specified in eventspurgeInterval, which ever is lower is used to purge events table to do db maintenance.
Is there a reason why events older than 6months is still acted on ?
2014-02-03 08:45 AM
Thanks Adai - you've answered the event retention question. But shouldn't events that are still in error or warning state get re-alerted?
In answer to your question, the only reason the events have not been acted on is laziness...
Something else I noticed this morning. I still have both OCUM instances running. The 5.1 version shows an alert in the "information or worse" report. The event is "snapshot full", and it lists the severity as "error". When I pull up the "error or worse" report, though, the event doesn't show up. On the 5.2 instance the event doesn't show in either report, but if I pull up the volume details for the volume in question, it shows a current warning state for snapshot full.
Am I missing something?