| Summary: | ASTERISK-19594: app_meetme unable to write frame (stuck channel) | ||||
| Reporter: | Mark Murawski (kobaz) | Labels: | |||
| Date Opened: | 2012-03-26 09:57:29 | Date Closed: | 2012-06-04 08:57:05 | ||
| Priority: | Major | Regression? | |||
| Status: | Closed/Complete | Components: | Applications/app_meetme | ||
| Versions: | SVN 1.8.10.1 | Frequency of Occurrence | Constant | ||
| Related Issues: |
| ||||
| Environment: | Attachments: | ( 0) meetme ( 1) meetme-hangup-trunk-360459.patch | |||
| Description: | Test setup: Asterisk 1.8 server (pbxA) Asterisk 1.8 server (pbxB) Make two calls from pbxA over IAX to pbxB, which both join the same meetme Hang up one call from pbxA pbxB console: [2012-03-25 19:04:48] WARNING[29025] app_meetme.c: Unable to write frame to channel IAX2/co01-144 [2012-03-25 19:04:48] WARNING[29025] app_meetme.c: Unable to write frame to channel IAX2/co01-144 [2012-03-25 19:04:48] WARNING[29025] app_meetme.c: Unable to write frame to channel IAX2/co01-144 [2012-03-25 19:04:48] WARNING[29025] app_meetme.c: Unable to write frame to channel IAX2/co01-144 [2012-03-25 19:04:48] WARNING[29025] app_meetme.c: Unable to write frame to channel IAX2/co01-144 [2012-03-25 19:04:48] WARNING[29025] app_meetme.c: Unable to write frame to channel IAX2/co01-144 [2012-03-25 19:04:48] WARNING[29025] app_meetme.c: Unable to write frame to channel IAX2/co01-144 [2012-03-25 19:04:48] WARNING[29025] app_meetme.c: Unable to write frame to channel IAX2/co01-144 [2012-03-25 19:04:48] WARNING[29025] app_meetme.c: Unable to write frame to channel IAX2/co01-144 [2012-03-25 19:04:48] WARNING[29025] app_meetme.c: Unable to write frame to channel IAX2/co01-144 [2012-03-25 19:04:48] WARNING[29025] app_meetme.c: Unable to write frame to channel IAX2/co01-144 [2012-03-25 19:04:48] WARNING[29025] app_meetme.c: Unable to write frame to channel IAX2/co01-144 [2012-03-25 19:04:48] WARNING[29025] app_meetme.c: Unable to write frame to channel IAX2/co01-144 ....etc Attached patch fixes the problem for me | ||||
| Comments: | By: Matt Jordan (mjordan) 2012-04-02 12:34:01.530-0500 Mark: I remember discussing this in #asterisk-dev when you first found this issue. I'm curious to know where in the loop we missed the hang-up in the first place - do you mind attaching a DEBUG log, just to help clarify what went wrong in the first place? Thanks Matt By: Mark Murawski (kobaz) 2012-04-05 23:45:00.332-0500 Will do. By: Mark Murawski (kobaz) 2012-04-11 12:02:59.220-0500 core set debug 5 This particular instance happened when two callers joined, and then one caller left, and then the last caller left and became a stuck channel. By: Kinsey Moore (kmoore) 2012-04-13 08:41:10.855-0500 This issue is a symptom of a deeper problem that probably should not be worked around. I am unable to reproduce this on my end, but I'd be interested in seeing whether the channel's readq still has frames in it. The other option is to find the revision that introduced this. By: Johan Wilfer (johan) 2012-04-13 08:54:06.707-0500 I've seen this in asterisk 1.4 for quite some time. Exactly the same error, but only once at a time using SIP. My guess is for 1-2 years... It doesn't seems to affect quality in any way so I havn't payed very much attention to it.. By: Mark Murawski (kobaz) 2012-04-13 09:56:44.254-0500 It doesn't seem to affect performance or quality but it does leave a stuck channel that is infinitely spewing lines to the console and log files. I'll see about dumping the channel readq By: Kinsey Moore (kmoore) 2012-04-24 13:41:46.781-0500 Mark, Have you had a chance to take a look at the readq to see if there are any frames sitting in it? By: Kinsey Moore (kmoore) 2012-05-22 10:08:50.885-0500 Mark, This may have been fixed in revision 357761 which was not in 1.8.10.1 and contains some fixes for missed frames. By: Mark Murawski (kobaz) 2012-06-01 15:11:39.736-0500 Seems to be fixed using 1.8 svn branch r368266 | ||||