[
https://issues.apache.org/jira/browse/AMQ-9547?focusedWorklogId=929432&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-929432
]
ASF GitHub Bot logged work on AMQ-9547:
---------------------------------------
Author: ASF GitHub Bot
Created on: 08/Aug/24 16:16
Start Date: 08/Aug/24 16:16
Worklog Time Spent: 10m
Work Description: cshannon opened a new pull request, #1278:
URL: https://github.com/apache/activemq/pull/1278
This method always throws an exception so it should be removed and no longer
used. Places in PageFile that called the method have been updated to either
remove the usage or throw an error.
Trying to add a test would be very difficult because it's not easy to stop
KahaDB with an unclean shutdown and restart and also because KahaDB will just
detect and recovery a corrupt index which makes asserting an issue difficult.
Issue Time Tracking
-------------------
Worklog Id: (was: 929432)
Remaining Estimate: 0h
Time Spent: 10m
> PageFile can call setLength() on the recovery file which always throws an
> exception
> -----------------------------------------------------------------------------------
>
> Key: AMQ-9547
> URL: https://issues.apache.org/jira/browse/AMQ-9547
> Project: ActiveMQ Classic
> Issue Type: Bug
> Components: KahaDB
> Affects Versions: 5.18.5, 6.1.3
> Reporter: Christopher L. Shannon
> Assignee: Christopher L. Shannon
> Priority: Major
> Fix For: 6.2.0, 5.18.6, 6.1.4
>
> Time Spent: 10m
> Remaining Estimate: 0h
>
> There are 3 places in PageFile that call setLength() in
> RecoverableRandomAccessFile. This method was changed in AMQ-5578 so that it
> always throws an exception when called. This method should just be removed
> entirely since it always throws an error and in the places where it is used
> we need to either stop calling it if we can or throw an exception if the
> length is unexpected.
> Currently when KahaDB tries to redo recovery updates, if the file length is
> encountered as 0 it will call this method which leads to throwing an
> exception. This causes the store to detect the error and rebuild the index.
> We should be able to just continue without calling this method and recover.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
For further information, visit: https://activemq.apache.org/contact
