[asterisk-bugs] [JIRA] (ASTERISK-28998) Segfault in pj/timer.c

[Gregory Massel (JIRA)]

    [ https://issues.asterisk.org/jira/browse/ASTERISK-28998?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=251477#comment-251477 ] 

Gregory Massel commented on ASTERISK-28998:
-------------------------------------------

The system has 24,215 PJSIP endpoints. Originally it was handling inbound registration, MWI and BLF subscriptions for 22,352 of these and qualifying them all every 60 seconds. A couple of days ago I offloaded all the inbound registration, OPTIONS (qualify) and BLF to OpenSIPS and changed all the PJSIP aors to statically point the the OpenSIPS proxy and disabled qualification (OPTIONS). With this, I also bypassed by far the bulk of the media (RTP) handling off Asterisk. Despite that dramatic reduction of workload on the Asterisk box, it still segfaulted again today in the exact same manner as it had done the previous five times. This would indicate to me that the issue is not load related and appears to be in the core of PJSIP (call handling) not in secondary functionality (e.g. qualification, presence, registration, etc.) MWI is still in Asterisk, however, looking at the backtrace, I don't see any indication that it relates to MWI.

Looking at the timing, it is possible - although pure speculation - that this particular problem has started since upgrading to 16.11.1, however, I do have multiple other Asterisk systems running 16.11.1 and have not experienced similar issues with them.

> Segfault in pj/timer.c
> ----------------------
>
>                 Key: ASTERISK-28998
>                 URL: https://issues.asterisk.org/jira/browse/ASTERISK-28998
>             Project: Asterisk
>          Issue Type: Bug
>      Security Level: None
>          Components: pjproject/pjsip
>    Affects Versions: 16.11.1
>         Environment: Ubuntu 18.04.4 LTS, kernel 5.3.0-62-generic, Asterisk 16.11.1
>            Reporter: Gregory Massel
>            Assignee: Gregory Massel
>         Attachments: backtraces.tgz
>
>
> Repeated, erratic unprovoked segfaults. 
> The backtraces all seem to show a similar pattern.
> Occurs on average daily, although it has sometimes gone days at a time without happening and sometimes multiple times within a single day.



--
This message was sent by Atlassian JIRA
(v6.2#6252)