[jira] [Comment Edited] (AMQ-6277) KahaDB does journal recovery for last append in error in normal restart case

["Gary Tully (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/AMQ-6277?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15270452#comment-15270452 ] 

Gary Tully edited comment on AMQ-6277 at 5/4/16 9:13 PM:
---------------------------------------------------------

think this fix can break the forceRecover index scenario if the firstLocation is returned twice.

AMQ4212Test proved this out!


was (Author: gtully):
think this fix can break the forceRecover index scenario if the firstLocation is returned twice.

Both the ackMessagefileMap and producerAudit may need recovery in an abnormal shutdown case because they will miss updates since their last checkpoint.

> KahaDB does journal recovery for last append in error in normal restart case
> ----------------------------------------------------------------------------
>
>                 Key: AMQ-6277
>                 URL: https://issues.apache.org/jira/browse/AMQ-6277
>             Project: ActiveMQ
>          Issue Type: Improvement
>          Components: KahaDB, Message Store
>    Affects Versions: 5.13.0
>            Reporter: Gary Tully
>            Assignee: Gary Tully
>            Priority: Minor
>             Fix For: 5.14.0
>
>
> On a normal restart - the journal is replayed from the last append location in error. Reporting some unnecessary info logging of the form{code}INFO | Recovering from the journal @1:503
>  INFO | Recovery replayed 1 operations from the journal in 0.0 seconds.{code}
> Recovery is only required when the last append location is different from the recovery location.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)