| Summary: | ASTERISK-18156: SIP messages stop being processed with res_timing_dahdi | ||||
| Reporter: | ppower (ppower) | Labels: | |||
| Date Opened: | 2011-07-19 15:01:00 | Date Closed: | 2011-08-26 14:31:36 | ||
| Priority: | Major | Regression? | |||
| Status: | Closed/Complete | Components: | |||
| Versions: | 1.8.2 | Frequency of Occurrence | |||
| Related Issues: |
| ||||
| Environment: | gentoo 2.6.36 dual quad core hyper threaded 2.67GHz intel processors 8GB RAM | Attachments: | ( 0) backtrace.txt ( 1) backtrace2.txt ( 2) locks.txt | ||
| Description: | This issue is very similar to ASTERISK-18074. There are differences: CPU usage was nil, DUNDi traffic seems to have stopped its processing as well. The code was not set for showing locks, but I will attach a back trace. There were 729 DUNDi threads! I expect this failure to occur rarely, so bear with me in the feedback department. | ||||
| Comments: | By: ppower (ppower) 2011-07-19 15:02:38.090-0500 The promised back trace. By: ppower (ppower) 2011-07-26 15:40:35.208-0500 Happened again today. uploading the back trace and core show locks. What i did notice prior to the lock up was very erratic CPU usage. Many times the CPU usage spiked at of 50%. A sister machine with more SIP end points has spent the day processing at a steady 1-2% (compiled with out thread debugging). By: Richard Mudgett (rmudgett) 2011-07-26 18:52:22.148-0500 The [^locks.txt] file shows a deadlock that was fixed with ASTERISK-17760. By: ppower (ppower) 2011-08-01 14:39:54.025-0500 FWIW - this lock up does occur with out reloading the dial plan. By: Paul Belanger (pabelanger) 2011-08-26 14:31:36.942-0500 We believe this issue is fixed in 1.8.6.0. | ||||