events and logs /var/log/messages don't match

[haxhia]

Hi all,

(I use CentOS 4.2 and ZoneMInder 1.21.3.)
I have a strange problem. I have 10 cameras and for some reason they decide to skip recording events from seemingly random times.
For example:

9556 Event-9556 Nad_Desk Motion 12/03 14:39:13 0.71 24 4 15 3 5
9000 Event-9000 Nad_Desk Motion 12/03 11:14:08 7.70 52 28 147 5 10

Note: time jumps from 11:14 to 14:39 and like /var/log/message shows there is no way that an event didn't occur in between:

Dec 3 11:38:01 spyserver zma_m8[4169]: INF [Nad_Desk: 371447 - Gone into alarm state]
Dec 3 11:38:02 spyserver zma_m8[4169]: INF [Nad_Desk: 371448 - Gone into alert state]
Dec 3 11:38:03 spyserver zma_m8[4169]: INF [Nad_Desk: 371455 - Gone back into alarm state]
Dec 3 11:38:05 spyserver zma_m8[4169]: INF [Nad_Desk: 371461 - Gone into alert state]
Dec 3 11:38:05 spyserver zma_m8[4169]: INF [Nad_Desk: 371462 - Gone back into alarm state]
Dec 3 11:38:05 spyserver zma_m8[4169]: INF [Nad_Desk: 371463 - Gone into alert state]
Dec 3 11:38:06 spyserver zma_m8[4169]: INF [Nad_Desk: 371467 - Gone back into alarm state]
Dec 3 11:38:10 spyserver zma_m8[4169]: INF [Nad_Desk: 371483 - Gone into alert state]
Dec 3 11:38:13 spyserver zma_m8[4169]: INF [Nad_Desk: 371493 - Left alarm state (9027) - 56(24) images]

And an interesting part is that, Dec 2 is skipped altogether by all cameras.

The two things that could be causing this problem are:
1.) the filter "PurgeWhenFull" which now says:
Use _2_ filter expressions Use Filter: _PurgeWhenFull_
-----------------------------------------------------------------------
_Archive Status_ _Unarchived Only_
_and_ _Disk Percentage_ _greater than or equal to_ _85_
-----------------------------------------------------------------------
Sort by: _Date/Time_ _Desc_ Limit to first _5_ results only;

Is this right? I just want to delete old stuff when disk is 85% full.

2.) I have a raid-1 setup with hda and hdc and this error in /var/log/messages seems interesting (and it's repeated quite often):
Dec 3 20:55:16 spyserver smartd[2658]: Device: /dev/hdc, 5 Offline uncorrectable sectors

Thanks alot for you help,
-Haxhia

[zoneminder]

Sorting your purge filter by date/time desc will mean that the newest events get deleted first. This is probably not what you want and will result in a bunch of old events never getting removed while all those in between get deleted, which is what I think you are observing. Change it to asc to get the oldest events firsts. Run the filter manually to check it supplies the events you want, though you may have to temporarily change the disk space value if you are already over the threshold.

Phil

[jameswilson]

i though tht e disk vales was a greater than type so if set to 85 and if disk usage at 90 filter would still execute, is this wrong?

[zoneminder]

Yes. But if the broken filter had cleaned the disk down to 84% then test running the filter if it was set to 85% would not produce any results so it would be hard to check. I was just suggesting putting the threshold down to say 80% to test.

Phil

[jameswilson]

whoops
i see now sorry Phil

James