[jira] [Commented] (ARTEMIS-2918) Support Server restart on I/O errors

["Francesco Nigro (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/ARTEMIS-2918?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17211562#comment-17211562 ] 

Francesco Nigro commented on ARTEMIS-2918:
------------------------------------------

After spent some time working on porting this feature from AMQ6 I've realized how much it's prone to have memory leaks and many corner cases that can make it unstable and dangerous: I'm creating a new task, after the comment of [~gtully], to a new task, that I'll link right after.

> Support Server restart on I/O errors
> ------------------------------------
>
>                 Key: ARTEMIS-2918
>                 URL: https://issues.apache.org/jira/browse/ARTEMIS-2918
>             Project: ActiveMQ Artemis
>          Issue Type: New Feature
>          Components: Broker
>    Affects Versions: 2.15.0
>            Reporter: Francesco Nigro
>            Assignee: Francesco Nigro
>            Priority: Major
>          Time Spent: 20m
>  Remaining Estimate: 0h
>
> Currently the JDBC journal already provide a periodic lock evaluation both for live and backup using HA shared store, but the default strategy on failures is to suicide the broker: a different strategy could be to unify the behavior seen for the file-based version (https://issues.apache.org/jira/browse/ARTEMIS-2421) and making use of the recent changes on https://issues.apache.org/jira/browse/ARTEMIS-2823 using connection pooling to be more resilient to network glitchs too.
>  
> This translate into supporting a feature similar to classic ActiveMQ's restartAllowed one, although it could be applied only on I/O failures because Artemis doesn't use a container to manage its instance.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)