Page 1 of 1

another PurgeWhenFull Issue

Posted: Sat Apr 03, 2010 7:12 am
by Pr0n
I know you guys are probably sick of troubleshooting these ones but before you yell RTFM I've scanned the forums and this seems to be a pretty unique issue...
PurgeWhenFull has been working perfectly fine keeping the drive at %90 full on a 200GB drive 4 cameras constantly recording at 3 FPS and 20 FPS during events for months and months and months. every thing has been perfectly stable, haven't even had to reboot in that time.

all of a sudden I was getting php errors on my login page that it was failing drive reads, I tried logging in directly to the machine and it wouldn't allow it saying the profile something or other couldn't be created because the drive was full, I booted to a live disc and manually deleted a few hundred events. booted back up, and it's working fine now.

Database audit is set to 30 minute intervals and no deletions of events where done during the last few months. I have never seen the drive percentage above %91. Fast delete is also disabled if that makes any difference.

What logs should I check or is this a known thing? Any ideas?

Posted: Sat Apr 03, 2010 12:26 pm
by mastertheknife
You can try this:
1) Set the audit to run every 15 minutes.
2) In the filter, change "Limit to first X results only" to something higher, such as 5 or 10.

Posted: Sun Apr 04, 2010 1:47 am
by kingofkya
you can also run zmaudit manually also check you database may have corrupt tables

did it again

Posted: Mon Apr 05, 2010 5:33 pm
by Pr0n
checked zmfilter.log and it has been deleting all the events i deleted manually just fine, no erroring out or anything although I can't see where it has ran the the full drive filter but then again the log reached it's max length (question for the creator: why would you put a max length on a log? Never seen that before) anyways I think I found the issue, somehow delete all matches got unchecked... weird since I'm the only person that accesses this thing and I haven't touched the filters. I'll test it and let you know

Posted: Tue Apr 06, 2010 3:31 am
by kingofkya
why would you put a max length on a log?

so you can open it sometime this decade. Windows server does this too.

linux app s(zm too) for the most part will do this zm.log zm.log0 zm.log1 and so on

log

Posted: Tue Apr 06, 2010 5:38 pm
by Pr0n
yeah it does and I've seen it but it takes a brute forcer and about 72 hours of guesses to get it there. not deleting a few files. anyways, sorry for off topic, issue fixed. If anyone complains of that checkbox magically getting unchecked (that's why I want a longer log so I can see that.) don't disbelieve them.