| Summary: | ASTERISK-24360: Asterisk deadlock | ||||
| Reporter: | VoIPCamp (voipcamp) | Labels: | |||
| Date Opened: | 2014-09-26 13:18:19 | Date Closed: | 2014-09-26 14:01:29 | ||
| Priority: | Major | Regression? | No | ||
| Status: | Closed/Complete | Components: | Applications/app_mixmonitor Channels/chan_agent Core/Channels Core/ManagerInterface | ||
| Versions: | 11.11.0 11.12.0 11.12.1 11.13.0 | Frequency of Occurrence | Frequent | ||
| Related Issues: |
| ||||
| Environment: | Centos release 6.5 GNU/Linux 2.6.32-431.29.2.el6.x86_64 Intel(R) Xeon(R) CPU E5506 @ 2.13GHz 8Gb RAM | Attachments: | ( 0) core.ast1113.26sep-00.gdb.txt ( 1) core.ast1113.26sep-00.log.txt | ||
| Description: | We originate calls using AMI with about 130 agents logged in (using AgentLogin) and recording with MixMonitor before entering any of the about 30 Queues. This is easily reproducible and happens twice a day in two different machines with similar specs.
When in deadlock it's not possible to originate or receive more calls and it's confirmed by a "core show channels" in CLI where there is not output. | ||||
| Comments: | By: VoIPCamp (voipcamp) 2014-09-26 13:19:46.702-0500 GDB output of: (gdb) bt (gdb) bt full (gdb) thread apply all bt (gdb) info threads By: VoIPCamp (voipcamp) 2014-09-26 13:29:44.624-0500 Asterisk log By: Matt Jordan (mjordan) 2014-09-26 14:01:15.745-0500 This is a known bug with {{pbx_realtime}}. {{pbx_realtime}} has a deadlock potential, as it puts a channel into autoservice while holding the {{pbx}} lock. {{pbx_realtime}} is an extended support module. Response times may reflect that. You may want to consider moving away from {{pbx_realtime}}, and instead use {{func_odbc}} to integrate databases with your dialplan. By: Matt Jordan (mjordan) 2014-09-26 14:01:30.045-0500 Duplicate of ASTERISK-21228 (and others) | ||||