Page 1 of 1

Trouble with epochtime on one cam out of four

Posted: Wed May 12, 2010 12:55 pm
by polarit
Hi all

I am running zoneminder with four cams, on a dual 3GHz system, with plenty of ram.

Three of the cams are running great, capturing events, converting to video, and other nice things, like upload video to a secure location, webcam function, etc.

three of the cameras are running great, zoneminder captures events, converts to video, etc.
the fourth does not - most of the time - produce any events

all cameras are captured with zmu, three of them are combined to a panorama, all four cameras are captured fine, so the actual capture from the capture card is working fine.

copy from camera window :

Code: Select all

115175	e-115175	2010-05-12 10:09:42	19.68	91/1	100/100	X
115174	e-115174	2010-05-12 10:09:42	6.01	31/1	100/100	X
114990	e-114990	1970-01-01 01:00:00	-14976815.98	91/1	100/100	X
114935	e-114935	1970-01-01 01:00:00	-15001400.22	91/1	85/85	X
114862	e-114862	1970-01-01 01:00:00	-15005655.36	91/1	7/7	X
114831	e-114831	1970-01-01 01:00:00	-15008618.93	91/1	92/92	X
versions:
Linux version 2.6.27-7-server (buildd@palmer) (gcc version 4.3.2 (Ubuntu 4.3.2-1ubuntu11) ) #1 SMP Tue Nov 4 20:18:35 UTC 2008 (Ubuntu 2.6.27-7.16-server)

Zoneminder v1.24.2

Apache/2.2.9 (Ubuntu)

mysql> status
--------------
mysql Ver 14.12 Distrib 5.0.67, for debian-linux-gnu (i486) using readline 5.2

Connection id: 41890
Current database:
Current user: root@localhost
SSL: Not in use
Current pager: stdout
Using outfile: ''
Using delimiter: ;
Server version: 5.0.67-0ubuntu6 (Ubuntu)
Protocol version: 10
Connection: Localhost via UNIX socket
Server characterset: latin1
Db characterset: latin1
Client characterset: latin1
Conn. characterset: latin1
UNIX socket: /var/run/mysqld/mysqld.sock
Uptime: 2 hours 40 min 12 sec

Threads: 26 Questions: 127792 Slow queries: 1 Opens: 45 Flush tables: 1 Open tables: 39 Queries per second avg: 13.295
--------------


nothing looking suspicious in any logfiles

i have seen in other threads that this may be produced by a overloaded sqlserver, but i don't think this is the case with this one, as the other three works fine.

Posted: Thu May 13, 2010 1:59 am
by bb99
Have you tried swapping one of the known working cameras for the fourth? This is easiest done by changing device and channel settings within Zoneminder or simply do a physical swap. Either way you eliminate the possibility of the camera as the cause.

Posted: Thu May 13, 2010 4:19 am
by cordel
Are you using IPCS or Mapped memory?

Posted: Thu May 13, 2010 6:22 am
by polarit
@bb99
the cameras are fine, and likewise the grabbercard
i am able to extract excellent pictures with zmu.

@cordel
I belive i am using mapped memory for zoneminder to communicate
(how do i check ?)

Posted: Thu May 13, 2010 8:16 am
by cordel

Code: Select all

ipcs -m
If you see segments created by the apache user while ZM is running, then you are using ipcs shared memory. If not then you are using mapped memory.

Posted: Thu May 13, 2010 9:49 am
by polarit

Code: Select all

------ Shared Memory Segments --------
key        shmid      owner      perms      bytes      nattch     status      
0x7a6d0001 0          www-data  700        55297352   3                       
0x7a6d0002 32769      www-data  700        55297352   3                       
0x7a6d0003 65538      www-data  700        55297352   3                       
0x7a6d0005 98307      www-data  700        92161672   3                       
shared ... ?

Posted: Thu May 13, 2010 10:07 am
by cordel
Indeed!

Let me backup from my original direction though...
Your 4th cam is using more memory than the first three so I would guess you have the buffer settings higher on this cam or it is set to a larger image size?

Have you tuned your zone for this cam?
If the field of view is different from the other cams it is often the case that the preset zone is not sufficient and need to be adjusted. Since there is no way to know the field of view and lens used for every cam, the preset zone are very generic.

Otherwise you can check the following which I was initially thinking:

If you turn debug on and set to level 5, see if a "Can't shmget out of memory" error pops up, and if it does look in the WIKI-FAQ for setting shared memory.

OR

You can try reducing the buffer settings or image size for that cam. Make sure that your pre-buffer is lower than your overall ring buffer.

Posted: Fri May 14, 2010 9:26 am
by polarit
none"Can't shmget out of memory"
i think i would have noticed the memory-thing, usually the 'source' column turns red, when no 'modetect' is selected - or not enough memory allocated.

Code: Select all

05/14/10 11:25:02.911658 zmu[23557].DB1-zm_monitor.cpp/2262 [Got 1 monitors]
05/14/10 11:25:02.915610 zmu[23557].DB1-zm_monitor.cpp/338 [monitor purpose=0]
05/14/10 11:25:02.915668 zmu[23557].DB1-zm_monitor.cpp/345 [mem.size=92161672]
05/14/10 11:25:02.916009 zmu[23557].DB1-zm_zone.cpp/50 [Initialised zone 0/All - 1 - 640x480 - Rgb:ff0000, CM:3, MnAT:15, MxAT:0, MnAP:50, MxAP:75000, FB:3x3, MnFP:50, MxFP:50000, MnBS:10, MxBS:0, MnB:0, MxB:0, OF: 0]
05/14/10 11:25:02.938867 zmu[23557].DB1-zm_monitor.cpp/458 [Monitor e has function 3]
05/14/10 11:25:02.938927 zmu[23557].DB1-zm_monitor.cpp/459 [Monitor e LBF = '%N - %y/%m/%d %H:%M:%S', LBX = 0, LBY = 0]
05/14/10 11:25:02.938952 zmu[23557].DB1-zm_monitor.cpp/460 [Monitor e IBC = 100, WUC = 10, pEC = 40, PEC = 50, EAF = 1, FRI = 1000, RBP = 7, FM = 0]
05/14/10 11:25:02.938980 zmu[23557].DB1-zm_monitor.cpp/2459 [Loaded monitor 5(e), 0 zones]
why is monitor 5 loaded with 0 zones ?

Posted: Mon May 17, 2010 9:39 am
by cordel
Do you have a zone configured and enabled for that MonitorID?
The only other reason maybe that it's failing to load the zone(s) because of an unexpected zone setting.

Posted: Tue May 18, 2010 8:32 am
by polarit
The camera...monitor... has 4 zones.
i created two zones with 'best, high sensitivity' and two with 'fast low sensitivity' presets, in total about 40% of the image.

the only thing not default are the buffer settings, because i want to record longer periods, without breaks.

this installation tends to record multiple 20sec slices right after each other ?