[asterisk-bugs] [JIRA] (ASTERISK-26009) ARI Bug Crashes Asterisk ($1000 Bounty)

[Sidney VanNess (JIRA)]

    [ https://issues.asterisk.org/jira/browse/ASTERISK-26009?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=230577#comment-230577 ] 

Sidney VanNess commented on ASTERISK-26009:
-------------------------------------------

libjansson.so.4.5.0. Also fails with  libjansson.so.4.7.0
We're not using PJSIP

Also, we have a workaround for this in place, it appears. Looks like we do not see this crash if we do not subscribe all. However, the procedure outlined above definitely crashes Asterisk consistently, so it should be fixed. 

Taking down bounty given functional work-around.
sv

> ARI Bug Crashes Asterisk ($1000 Bounty)
> ---------------------------------------
>
>                 Key: ASTERISK-26009
>                 URL: https://issues.asterisk.org/jira/browse/ASTERISK-26009
>             Project: Asterisk
>          Issue Type: Bug
>      Security Level: None
>          Components: Core/Stasis, Resources/res_ari, Resources/res_ari_events
>    Affects Versions: 13.7.2
>         Environment: Ubuntu 14.04.4 LTS
>            Reporter: Sidney VanNess
>            Assignee: Sidney VanNess
>            Severity: Critical
>              Labels: crash
>         Attachments: backtrace.txt
>
>
> I've posted this on the Asterisk dev mailing list, and we are offering a $1000 Bounty for a fix to this issue. See: http://lists.digium.com/pipermail/asterisk-dev/2016-May/075517.html for details.
> ============
> *Description/How To Reproduce*
> $ wscat -c
> 'ws://localhost:8088/ari/events?api_key=<key>:<password>&app=<appname>&subscribeAll=1
> #
> Run 5 or 10 of these at once with appname different for each one with all
> of them subscribing to all events.
> Place a call which will emit stasisStart for one of the apps (maybe a few
> times).
> Asterisk should terminate at this point.
> *This was done using local channels
> json_mem_free in json.c seems to have the issue.  If you don't free the
> memory, then it does not terminate.
> Of note, the issue seems somewhat similar in nature to the following bug
> report:
> https://issues.asterisk.org/jira/browse/ASTERISK-25775
> Therefore, we expect the fix to our issue also would result in the
> resolution of this seemingly related bug report.



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