[asterisk-bugs] [JIRA] (ASTERISK-20347) Asterisk 1.8.16rc1 deadlock in cdr_mysql

[Alex Villacís Lasso (JIRA)]

     [ https://issues.asterisk.org/jira/browse/ASTERISK-20347?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Alex Villacís Lasso updated ASTERISK-20347:
-------------------------------------------

    Attachment: asterisk-11.17.0-cdr-mysql-reload-avoid-deadlock.patch

I can confirm (through the output of core show locks) that this issue remains in asterisk-11.17.0. This patch against 11.17.10 is an attempt to fix the root cause of the deadlock: cdr_mysql is calling ast_cdr_unregister() with its internal lock held, while a hangup requests the cdr lock and then the cdr_mysql internal lock.

> Asterisk 1.8.16rc1 deadlock in cdr_mysql
> ----------------------------------------
>
>                 Key: ASTERISK-20347
>                 URL: https://issues.asterisk.org/jira/browse/ASTERISK-20347
>             Project: Asterisk
>          Issue Type: Bug
>      Security Level: None
>          Components: Addons/cdr_mysql
>    Affects Versions: 1.8.16.0
>         Environment: Centos 5.8 x64, 8core, 8GB RAM
>            Reporter: David Brillert
>            Severity: Critical
>         Attachments: asterisk-11.17.0-cdr-mysql-reload-avoid-deadlock.patch, coreshowlocks.txt, thread-apply-all-bt.txt
>
>
> As per Matt Jordan I am opening a bug report for a deadlock found in Asterisk 1.8.16rc and cdr_mysql.
> All I have for now is a gdb backtrace unoptimized but should soon be able to provide a non-optimized bt and core show locks output.  Opening ticket without that for now...



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