zoneminder mixing up my sources

jameswilson · Post by **jameswilson** » Mon Sep 19, 2005 8:04 pm

er.... good luck

James

mdaughtrey · Post by **mdaughtrey** » Tue Sep 20, 2005 1:02 am

Ok, looking into the log dump a little further...
Phil/Cordel, does this request to zms look OK?

GET /zm.php?view=watchfeed&mode=&mid=9&scale=100&control= HTTP/1.1

"mode" is missing a parameter and "control" has a space.

Post by **zoneminder** » Tue Sep 20, 2005 9:53 am

That's not actually the request to zms, but it will be in that page generated. An empty mode I think will just cause it to use the default and I think control is just blank, the space separates the url from the protocol. If you can get the contents of the page generated by your request it might tell us a bit more.

Phil

jameswilson · Post by **jameswilson** » Tue Sep 20, 2005 11:16 am

My earlier input suggested this was a client issue and i thought it was a power issue. I noticed this morning that my home machine (windows fx53 1Gb Raptor Raid 0) was doing it to. A close and open fised it but obviously i was wrong

James

mdaughtrey · Post by **mdaughtrey** » Tue Sep 20, 2005 12:05 pm

Phil: a strings-filtered Ethereal capture of the session is at http://salem.dyndns.org/zmlogs/zm.dmp.txt

Post by **zoneminder** » Tue Sep 20, 2005 12:53 pm

It's a bit of a long shot but have you tried changing your zms path to nph-zms rather than just zms? I'm a little concerned to see chunked encoding in your transfer log and I'm wondering if this messes things up at all. Using an nph script should prevent your httpd from interfering with the stream.

Phil

jameswilson · Post by **jameswilson** » Tue Sep 20, 2005 2:34 pm

I have changed mine (in options - paths
And its less of a problem ie its done it once since and cleared itself which normally it doesnt. It was just after a restart mind of zm if this has anything to do with it

James

mdaughtrey · Post by **mdaughtrey** » Tue Sep 20, 2005 5:12 pm

Ok, I'll give it a shot tonight.

jameswilson · Post by **jameswilson** » Tue Sep 20, 2005 5:22 pm

Scap that still doing it
Damn
Sorry

James

mdaughtrey · Post by **mdaughtrey** » Tue Sep 20, 2005 6:50 pm

I'll do another Ethereal capture of it anyway, see if it shows up any clues.

I'll also enable debugging on zms.

mdaughtrey · Post by **mdaughtrey** » Wed Sep 21, 2005 1:48 am

It hasn't changed anything - I still see the alternating jpegs on Ethereal. I've run out of time for this evening; I'll enable debug output on zms tomorrow, suggestions appreciated in the meantime.

mdaughtrey · Post by **mdaughtrey** » Fri Sep 23, 2005 3:13 am

Started looking through the zms debug output, see ***:

Sep 22 23:05:00 salem last message repeated 11 times
Sep 22 23:05:00 salem zms[667]: INF [16/20: ]
Sep 22 23:05:00 salem zms[667]: INF [Image size 19185]
Sep 22 23:05:00 salem zms[667]: INF [shared_data 4098e000]
Sep 22 23:05:01 salem last message repeated 13 times
Sep 22 23:05:01 salem zms[667]: INF [17/20: ]
Sep 22 23:05:01 salem zms[667]: INF [Image size 19185]
Sep 22 23:05:01 salem zms[667]: INF [shared_data 4098e000]
Sep 22 23:05:01 salem zms[667]: INF [19/20: ]
*** Sep 22 23:05:01 salem zms[667]: INF [Image size 21770] ***
Sep 22 23:05:01 salem zms[667]: INF [shared_data 4098e000]

Post by **cordel** » Fri Sep 23, 2005 6:49 am

I don't think this is a zms issue but really zmc. I had turned on Record and mocord on your machine for 20 seconds each. The captured images in the events directories were alternating as well. What ever zmc does zms will follow. So the problem is not in the streaming but in the capture.

mdaughtrey · Post by **mdaughtrey** » Fri Sep 23, 2005 2:25 pm

Ok, noted. I'll go back to looking at zmc.

mdaughtrey · Post by **mdaughtrey** » Fri Sep 23, 2005 8:06 pm

You know what though...earlier logs indicated that zmc was sticking with the program as far as capture goes - the same image size was getting picked up every time. This makes me think that there's some problem with the way the shared memory is being used.
In my last posted log, the "x/y" indicates the current last written index (x) and the total number of buffers for that source (y). It's odd that the index jumps around and back again, and that every time it jumps, zms gets the wrong image.