Page 1 of 1

purgewhenfull filter not working as expected

Posted: Wed Jul 15, 2009 6:19 am
by btrotter
I have the purgewhenfull filter set to run in the background and delete all events when the disk percentage is greater than or equal to 88%. It is currently sitting at 93% and staying there. If I run "df -h", I confirm the disk is at 93% full.

When I look at /var/log/messages, I see lots of messages like this:
Jul 15 02:14:37 t5cctv zmfilter[6540]: INF [Deleting event 2531]
Jul 15 02:14:37 t5cctv zmfilter[6540]: INF [Deleting event 2532]
Jul 15 02:14:37 t5cctv zmfilter[6540]: INF [Deleting event 2533]
Jul 15 02:14:37 t5cctv zmfilter[6540]: INF [Deleting event 2534]

These messages are continous. However, the disk percentage never goes down. It just sits there at 93%. I have had to go to the console a couple times lately and manually delete events to get the percentage to drop.

Any idea on where to start troubleshooting this one?

Posted: Fri Jul 17, 2009 6:01 pm
by rcswrc
How many cameras do you have? How big is your drive?

Do you have enabled the option OPT_FAST_DELETE ?

In my purgewhenfull filter, I select the option delete all matches. If you have a lot of cameras you should increase the limit to first <limit> results only. Also remember to adjust the time how often the filters are run FILTER_EXECUTE_INTERVAL and how often to check database and filesystem consistency AUDIT_CHECK_INTERVAL. You may customize your system with all these options.

For example I have the events to end every 600 seconds, so my FILTER_EXECUTE_INTERVAL is 660 seconds, I have four cameras so I limit the filter to the first 30 results and very important my AUDIT_CHECK_INTERVAL is set to 660 seconds.

So give it a try and then show your results.