You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@activemq.apache.org by "Philippe Cuvecle (JIRA)" <ji...@apache.org> on 2014/06/04 15:10:02 UTC
[jira] [Updated] (AMQ-5208) ActiveMQ restart when Locker looses
lock weakness
[ https://issues.apache.org/jira/browse/AMQ-5208?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Philippe Cuvecle updated AMQ-5208:
----------------------------------
Component/s: Broker
> ActiveMQ restart when Locker looses lock weakness
> -------------------------------------------------
>
> Key: AMQ-5208
> URL: https://issues.apache.org/jira/browse/AMQ-5208
> Project: ActiveMQ
> Issue Type: Improvement
> Components: Broker
> Affects Versions: 5.9.1
> Reporter: Philippe Cuvecle
>
> The feature implemented in AMQ-4526 does not help if the restart does not succeed
> Imagine the following scenario : 2 AMQ configured with jdbcPersistenceAdapter
> - At the beginning AMQ1 is master, AMQ2 is slave
> - database targeted by jdbcPersistenceAdapter shutsdown or connectivity is lost
> - AMQ1 restarts but because DB is not back during restart it dies
> - AMQ2 tries to get the lock every 10s and stay alive
> - DB is back, AMQ2 becomes master
> - DB shutdown again, AMQ2 restart and dies for the same reason AMQ1 died
> -> We don't have any AMQ up
> When a lock is down why don't you just deny AMQ connections and retry to become master on a regular basis ?
> More generally how can we have a master/slave high availability configuration that can tolerate long storage unavailability ?
--
This message was sent by Atlassian JIRA
(v6.2#6252)