[asterisk-bugs] [JIRA] (ASTERISK-28956) res_odbc: ODBC connection does not always reconnect

[Luke Escude (JIRA)]

    [ https://issues.asterisk.org/jira/browse/ASTERISK-28956?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=251290#comment-251290 ] 

Luke Escude commented on ASTERISK-28956:
----------------------------------------

Joshua, is there a way to disable the lockfile the voicemail spool folders? Since it's using ODBC, I don't think it's really necessary to have that.

Apparently if the SQL server goes down for a bit during a voicemail transaction, the lockfile will remain, preventing further access until I manually remove it.

I'm thinking about creating a cronjob every morning at 3am to wipe out all lockfiles in voicemail but this might do more harm than good.

> res_odbc: ODBC connection does not always reconnect
> ---------------------------------------------------
>
>                 Key: ASTERISK-28956
>                 URL: https://issues.asterisk.org/jira/browse/ASTERISK-28956
>             Project: Asterisk
>          Issue Type: Bug
>      Security Level: None
>          Components: CDR/cdr_adaptive_odbc
>    Affects Versions: 16.8.0
>         Environment: CentOS 7 x64
>            Reporter: Luke Escude
>            Assignee: Unassigned
>
> So I am not too sure how to diagnose this one, but I've seen it happen too many times to not report it.
> Environment: Our Asterisk instances use an ODBC connection for CDR (adaptive_odbc) as well as an ODBC connection for Voicemail. The DSNs for each of these is separate, since they point to separate MySQL servers.
> Failure Scenario: A SQL server goes down for some X minutes. During those X minutes, all Asterisk instances that are supposed to connect to that particular one are aware of the issue, and odbc show will indicate as such. This is expected behavior, and happens for both the Voicemail ODBC and the CDR ODBC connection.
> Recovery Scenario: When a Voicemail SQL server comes back up, ALL Asterisk instances see it and reconnect. This is expected. However, when the CDR SQL server comes back up, only a majority of Asterisk instances reconnect. The remaining ones will say 1 out of 1 (meaning ODBC is connected) but will not log CDR at all until performing a module reload res_adaptive_odbc command. Some of them will still say 0 out of 1, indicating the connection is indeed dead, but they never attempt to reconnect.
> While the issue is happening, are there any other useful commands I can run to assist it diagnosing this? I don't want to have to restart asterisk since obviously that will fix it.



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