| Summary: | ASTERISK-17925: CPU spike if hang up during getting message "You are currently the only person in this conference" | ||
| Reporter: | Jacob LIpstein (jacobli) | Labels: | |
| Date Opened: | 2011-05-25 15:27:46 | Date Closed: | 2011-06-10 13:04:58 |
| Priority: | Major | Regression? | No |
| Status: | Closed/Complete | Components: | Applications/app_meetme |
| Versions: | 1.8.4 | Frequency of Occurrence | |
| Related Issues: | |||
| Environment: | Attachments: | ||
| Description: | Asterisk CPU usage jumps to 100% if a first call comes to the MeetMe application and the caller hangs up when getting the "Only person" message. When the call comes from other asterisk via IAX protocol then the receiving asterisk has to be restarted in order to restore a normal functionality. ****** ADDITIONAL INFORMATION ****** I'm pretty sure the problem persists from older versions of asterisk because I have run into it several times during last 2-3 months (upgraded to 1.8.4 two weeks ago) but only now realised how to reproduce the issue. In order to avoid the problem I have disabled the "Only person" message. | ||
| Comments: | By: Leif Madsen (lmadsen) 2011-05-27 10:58:40 You'll have to provide some debugging information such as a 'core show locks' and a backtrace from the running process. Documentation on how to do this correctly is available on the Asterisk wiki: https://wiki.asterisk.org/wiki/display/AST/Getting+a+Backtrace By: Jacob LIpstein (jacobli) 2011-05-27 11:09:18 Sorry, but everything happened at my Production asterisk and as I mentioned I disabled the "Only person" message. I don't like to enable it back and try to reproduce the issue at the Production. If at some point I do have other instance of asterisk I'll do that. By: Jacob LIpstein (jacobli) 2011-06-09 11:03:06.114-0500 Close the issue, please. The problem was related with timing_timerfd taking priority over timing_dahdi when timing_timerfd became available after the Linux kernel upgrade. By: Jacob LIpstein (jacobli) 2011-06-10 13:04:58.425-0500 The problem was related with timing_timerfd taking priority over timing_dahdi when timing_timerfd became available after the Linux kernel upgrade. | ||