Summary:ASTERISK-07639: asterisk crash - related to MixMonitor usage
Reporter:vortex_0_o (vortex_0_o)Labels:
Date Opened:2006-08-31 11:02:08Date Closed:2006-09-18 11:26:27
Versions:Frequency of
Environment:Attachments:( 0) 2-bt.txt
( 1) 2-messages.txt
( 2) 3-bt.txt
( 3) 3-messages.txt
( 4) 4-bt.txt
( 5) 4-messages.txt
( 6) bt-1.txt
( 7) debug-crash.txt
( 8) messages-crash.txt
Description:perhaps someone could reclassify as to cause - only thing to mention most calls are using mixmonitor.

This is my first crash report so please let me know if i have done anything wrong.
Comments:By: Serge Vecher (serge-v) 2006-08-31 11:13:10

1) Was this Asterisk compiled with 'make don-optimize'? If not, can you recompile and try to reproduce the crash?
2) Have you done modifications to the source code?
3) In the future, it's preferable that you post the console output. Just make sure the debug is enabled for it in logger.conf.

By: vortex_0_o (vortex_0_o) 2006-08-31 11:21:40

1. it was "make valgrind"  - is that the same?
(i was just following the wiki)
2. no
3. ok

By: Serge Vecher (serge-v) 2006-08-31 12:05:27

1) yes, it is, thanks.

By: vortex_0_o (vortex_0_o) 2006-09-01 04:32:19

attached another crash


By: vortex_0_o (vortex_0_o) 2006-09-01 05:08:53

and another one


By: Eduardo Gallo (edgallo) 2006-09-01 07:26:12

hey vortex ... had the same problem with 1.2.10 ... mixmonitor is broken. Just gave up and reverted back to 1.2.9 ... running OK for a full day now !! Will wait for a solid fix. Looks like there are a bunch of dramatic changes in mixmonitor, that still need some milage to be up to the task. Cheers :)

By: vortex_0_o (vortex_0_o) 2006-09-01 10:04:33

last one for good luck...


now going to deactivate the use of mixmmonitor as per suggestion

By: Eduardo Gallo (edgallo) 2006-09-04 10:11:10

Looks like bug 7411 is related to this situation. Since a solution appears to have been found for TRUNK, I wonder if we will see such correction applied to 1.2 as well. For now ... running for more than 3 days without a single hicup ... just a nasty mem leak that looks like is caused by some faulty frame duplication on the spying mechanism (a recent correction for channel.c in TRUNK).

By: Serge Vecher (serge-v) 2006-09-05 11:10:30

well, the commit that fixed 7411 will doubtly be made for 1.2.x branch, given the extensiveness of it. If anybody is willing to backport the patch and post it here for testing, this will be an acceptable scenario, though ...

By: Joshua C. Colp (jcolp) 2006-09-05 15:06:00

I just backported the trunk fixes I did this weekend into 1.2 - please update to latest 1.2 branch and give it a try!

By: Serge Vecher (serge-v) 2006-09-05 15:11:38

the revision is 42054 in 1.2 branch

By: Eduardo Gallo (edgallo) 2006-09-05 17:36:18

OK file... good work ! give a couple of days and I?ll give your patch a spin onto a very busy production machine :) and post the results afterwards

By: vortex_0_o (vortex_0_o) 2006-09-15 08:41:27

Seems to be working well with MixMonitor so far.

How are you getting on edgallo?

By: Eduardo Gallo (edgallo) 2006-09-18 11:17:38

OK ... running just fine with the patch applied . TKS for the great work !