[asterisk-bugs] [JIRA] (ASTERISK-28689) res_pjsip: Crash when locking group lock when sending stateful response

[Joshua Elson (JIRA)]

    [ https://issues.asterisk.org/jira/browse/ASTERISK-28689?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=253907#comment-253907 ] 

Joshua Elson commented on ASTERISK-28689:
-----------------------------------------

So we've been working on tracking this down, as this one is pretty disruptive to the fleet, and we're approaching being certain that this condition is only triggered when realtime is used. Looking at the three related tickets on this, it appears that everyone is using realtime, although not everyone is using WebRTC. We initially enabled realtime as a part of another upgrade and assumed that 18.2 was the trigger for this condition, but thus far, we can observe crashes on all 18.x releases but only when realtime is enabled, and not thus far when we switch back to a static config.

Like the others, this occurs under a high rate of registration or failed/hung registration attempts. Still working on seeing if we can get a full lab repro that we can share, but perhaps this is helpful or jogs the memory of someone...

> res_pjsip: Crash when locking group lock when sending stateful response
> -----------------------------------------------------------------------
>
>                 Key: ASTERISK-28689
>                 URL: https://issues.asterisk.org/jira/browse/ASTERISK-28689
>             Project: Asterisk
>          Issue Type: Bug
>      Security Level: None
>          Components: pjproject/pjsip
>    Affects Versions: 13.20.0, 13.30.0
>         Environment: Ubuntu 16.04
> 4.4.0-119-generic
>            Reporter: Jesse Ross
>            Assignee: Unassigned
>            Severity: Major
>         Attachments: console, console, core2.tar.gz, core-brief.txt, core-brief.txt, core-full.txt, core-full.txt, core-locks.txt, core-locks.txt, core.tar.gz, core-thread1.txt, core-thread1.txt
>
>
> Hi,
> We are getting crashes like this randomly:
> Jan 13 21:33:21 voip kernel: [46434830.037815] traps: asterisk[24067] bq. general protection ip:7fc0ecf762b5 sp:7fc02d86ebc8 error:0 in libasteriskpj.so.2[7fc0ece5f000+16d000]
> Jan 13 21:44:48 voip kernel: [46435517.854842] asterisk[26382]: segfault at 18 ip 00007f3a5467ba28 sp 00007f39990b4bd0 error 4 in libasteriskpj.so.2[7f3a54564000+16d000]
> It didn't happen at all before but started to do it a couple times a day. Now it is doing it around 20 times per day. But only in the evening for some reason... 
> I've attached the core dump info.
> I upgraded to 13.30 from 13.20 and installed the latest libpri and dahdi but it didn't help.
> Thanks



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