Signal-Events

essup · Post by **essup** » Wed Jan 10, 2007 8:06 am

Hi
Each day i check the ZM console i get ~15 events and more which are caused by a 'Signal' instead of a 'Motion'.

Conspicuously are those events with >8000000 sec. which are like 92 days duration, but they aren't. Even the date is wrong and each day i delete those events i get new events caused from 'Signal'. Some of these events contain just 1 Picture, others somes seconds and even others run until a filter-set stops them from recording.
So what does this mean? I am runnung ZM 1.22.3

Post by **cordel** » Wed Jan 10, 2007 8:10 am

Do you have any scripts, devices, sockets, etc. configured that trigger zm?
What OS are you using?

essup · Post by **essup** » Wed Jan 10, 2007 11:07 am

The OS is Linux Slackware 11. I've got the filter "PurgeWhenFull*" configured with the SQL Syntax/Statement:

( MonitorID = 1 OR MonitorID = 2 OR MonitorID = 3 OR MonitorID = 4 OR MonitorID = 5 OR MonitorID = 6 ) AND ( DiskPercentage >= 90 )
SORT ASC ID only FIRST 50 HITS and DELETE HITS

And there is a cron which changes the running mode

0 8 * * 1-5 /usr/local/bin/zmpkg.pl day
0 17 * * 1-5 /usr/local/bin/zmpkg.pl night

so that after 8 am the cameras are idle (monitor mode) and start recording after 5 pm.

Post by **cordel** » Wed Jan 10, 2007 9:49 pm

Okay but mysql has nothing to do with the problem you suggested.
The caused by a 'Signal' suggests that you enabled ZM_OPT_TRIGGERS and you are externally triggering ZM whether intentional or not.

essup · Post by **essup** » Thu Jan 11, 2007 7:46 am

No, i don't think MySQL has anything to do with it, but you asked me about 'triggers' and this sql statement (for me) is a trigger and the only filter btw.
Nevertheless, ZM_OPT_TRIGGERS isn't set.
Yesterday i deleted all signal-events

But new events have just been recorded this night.
Most of these signal-events (i just checked, if not all of them) have black screen(s) as first or/and last picture(s), so i guess it's because of that. Actually this is all fine if at least the date and duration were correct.
Could there be a problem regarding ntpd request? There is a cron daemon which runs a 'ntpdate' command each morning at 7am.

jameswilson · Post by **jameswilson** » Thu Jan 11, 2007 8:22 am

shot in the dark, not an amd x2 running the box is it?

essup · Post by **essup** » Thu Jan 11, 2007 9:28 am

No.

Intel Celeron 1000MHz (P3)

Post by **cordel** » Thu Jan 11, 2007 9:59 am

Another shot in the dark....
could it be reporting that video has been lost???
Since ZM looks for a color for video loss detection (I'll have to test as I have never lost video to see if I get the same thing).

essup · Post by **essup** » Thu Jan 11, 2007 10:49 am

hm...

root@ZoneMinder:/tmp# tail zmdc.log
01/11/07 11:48:31.396671 zmdc[10468].INF ['zmc -d /dev/video0' starting at 07/01/11 11:48:31, pid = 30467]
01/11/07 11:48:41.902890 zmdc[10468].INF ['zmc -d /dev/video0' stopping at 07/01/11 11:48:41]
01/11/07 11:48:41.921132 zmdc[10468].ERR ['zmc -d /dev/video0' exited abnormally, exit status 255]
01/11/07 11:48:42.241379 zmdc[10468].INF ['zmc -d /dev/video1' starting at 07/01/11 11:48:42, pid = 30476]
01/11/07 11:48:42.758111 zmdc[10468].INF ['zmc -d /dev/video1' stopping at 07/01/11 11:48:42]
01/11/07 11:48:42.763379 zmdc[10468].ERR ['zmc -d /dev/video1' exited abnormally, exit status 255]
01/11/07 11:48:43.088921 zmdc[10468].INF ['zmc -d /dev/video0' starting at 07/01/11 11:48:43, pid = 30482]
01/11/07 11:48:43.601409 zmdc[10468].INF ['zmc -d /dev/video1' starting at 07/01/11 11:48:43, pid = 30486]
01/11/07 11:48:44.177482 zmdc[10468].INF ['zmc -d /dev/video0' stopping at 07/01/11 11:48:44]
01/11/07 11:48:44.242082 zmdc[10468].ERR ['zmc -d /dev/video0' exited abnormally, exit status 255]
root@ZoneMinder:/tmp# du -hs zmdc.log
252M zmdc.log

jameswilson · Post by **jameswilson** » Thu Jan 11, 2007 10:29 pm

shot 3
overloaded what the 5 min load

essup · Post by **essup** » Wed Jan 17, 2007 10:32 am

jameswilson wrote:shot 3
overloaded what the 5 min load

What do you mean? Should i post the load average when ZM is in 'modect' mode?

Edit:
I put ZM into modect mode. After 5min the load average varies from 0.6 up to 0.9. The color of the video-source path also changes between red and green.

Cell · Post by **Cell** » Thu Jan 18, 2007 12:39 pm

Using Slackware 11 myself, and having also had color bands (I hope thats what you are referring to) found that adding "gbufsize=16 card=x,x " to my modprobe.conf resolved this for me.

My thinking is that the "Signal" event is a false "camera failure detection". Don't have ZM open to check the options. But you might only be seeing Red & Green. Blue, I think the default triggers the event.

Also upgrading to kernel 2.6.19 patch 1, made my system more stable.

Perhaps another shot in that void called darkness?