Alarm Frame Count time continuity issue [SOLVED]

[erebus]

Hi

I was having issues on a cloudy day with a massive number of 1-10 alarm frame count 2 second videos being recorded (around 1000 an hour). I searched around and found someone who said changing the Alarm Frame Count to around 16 would get rid of this issue, so I did that and it fixed the problem.

Now however I have a time continuity issue. The server records all the events fine but they jump back in time. This causes them not to play back correctly in the Event Viewer window. I can export them fine however, when I watch the event I see someone walking in the frame, then they jump back and walk the same path they just did and then the video finishes correctly. I set the alarm frame count to different values and still have this issue unless it is set to 1 (setting it to 1 causes it to work perfectly fine).

This also causes an error in my httpd log when I try to view the event. I am also either forced to close my browser or restart the httpd service to be able to open any ZM webpage (all pages outside the ZM directory on the server work fine with out restarting browser or service).

I wonder if it is an issue with my buffer settings, but I could not find anything on the forum about this issue.

Here is my setup:

Code: Select all

CentOS 5.5
ZoneMinder v1.24.2
Source Type: Ffmpeg
Image Buffer Size: 40
Warmup Frames: 25
Pre Event Image Count: 35
Post Event Image Count: 35
Stream Replay Image Buffer: 1000
Alarm Frame Count: 16

As I said, if i change Alarm Frame Count to 1, it works but I get a billion events. If i leave it above 1 I get a time discontinuity.

Here is my Apache error:

Code: Select all

[Tue Jul 27 09:09:24 2010] [error] [client client] array (\n  0 => \n  array (\n    'file' => '/var/www/html/zm/ajax/stream.php',\n    'line' => 55,\n    'function' => 'ajaxError',\n    'args' => \n    array (\n      0 => 'socket_sendto( /tmp/zms-796558s.sock ) failed: No such file or directory',\n    ),\n  ),\n  1 => \n  array (\n    'file' => '/var/www/html/zm/index.php',\n    'line' => 116,\n    'args' => \n    array (\n      0 => '/var/www/html/zm/ajax/stream.php',\n    ),\n    'function' => 'require_once',\n  ),\n), referer: http://server/zm/index.php?view=event&eid=20069&filter[terms][0][attr]=MonitorId&filter[terms][0][op]=%3D&filter[terms][0][val]=2&sort_field=StartTime&sort_asc=1&page=1

Thanks (and apologies if this has already been answered)

[bb99]

For your original problem did you try setting the "Overload Frame Ignore Count" in the zones settings to a higher setting? This zone setting is similar in result to alarm frame count. For your current problem I can't help, haven't seen this before. What happens if you change the setting to 16 and reboot? Something else: for 2 second clips with pre and post set to 35 (71 minimum frames per event) either these cameras are smokin' along or there could be another issue...

[erebus]

humm, never saw anything that documented that setting and just left it alone. Changed the setting and will see what happens tomorrow when it is again light outside.

In reality by fixing the first problem the second problem will probably be fixed also, if the "Overload Frame Ignore Count" works, then I should be able to set the "Alarm Frame Count" to 1 and the problem _should_ vanish!

as far as them cooking, they sure are! right now I am testing the system so they run at 30FPS all the time and capture at 30FPS. It is only 5 cameras (one in a dark server room 95% of the time) and currently there are only 3 people in the building. When it opens in a month or so and there are ~200 people a day I will probably cut back quite a bit on my settings, or cause the server to actually catch fire...which would look nice on the camera recording it happen =)

[bb99]

If this fixes your problem please edit your title to include "solved". This is a common reoccurring post so searches become dead end reading but with a result list with the word solved in the title searches become much easier.

[erebus]

So with the new settings overnight I still had over 125 events that had 1-3 alarm frames. I looked and I only had about 10 events over the same time period of time that had 4-15 alarm frames. I changed the settings in buffer "Alarm Frame Count" to 3 and I am still having the same issue where the video will jump back in time.

When I look at the stills this is what I see. If Alarm Frame Count is set to 4, then I will see 1 frame then it will jump back to the beginning of the video. If i have Alarm Frame Count set at 15, I see 12 frames and then it jumps back to the beginning of the video (start of the buffer with no one in the frame).

I would rather use the Alarm Frame Count feature than writing a script to delete all the 1-4 alarm frame events from the database and having cron execute it a few times a day. It almost looks like it is appending the alarm frames, then adding the buffer and the rest of the video to the event folder.

[erebus]

Think I found the problem. I changed my pre even image count to below my warmup frame count and then increased the alarm frame count. problem vanished. Not sure why this fixed the problem, but it did.

[whatboy]

>>Pre Event Image Count: 35

WHY???
Pre Event Image Count should not past half of the ring buffer... it says in the documentations!!!