[
https://issues.apache.org/jira/browse/AMQ-9484?focusedWorklogId=916084&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-916084
]
ASF GitHub Bot logged work on AMQ-9484:
---------------------------------------
Author: ASF GitHub Bot
Created on: 23/Apr/24 17:42
Start Date: 23/Apr/24 17:42
Worklog Time Spent: 10m
Work Description: cshannon commented on PR #1207:
URL: https://github.com/apache/activemq/pull/1207#issuecomment-2073017418
I only took a quick glance but 2 things stand out to me immediately that are
issues:
1. The current PR is using recoverNextMessages() which changes state of the
cursor, so this definitely won't work in an online broker without breaking the
queue if the intent is for a user to use this to export while a broker is
online.
2. This scans through the index and just skips when you could just use
MessageOrderIndex with the order iterator and skip ahead by setting the batch
location on the index
This needs a bit more context with use cases to see how this is intended to
be used to probably know the best way of doing some of this
Issue Time Tracking
-------------------
Worklog Id: (was: 916084)
Time Spent: 20m (was: 10m)
> Support exporting kahadb messages from a queue with an offset
> -------------------------------------------------------------
>
> Key: AMQ-9484
> URL: https://issues.apache.org/jira/browse/AMQ-9484
> Project: ActiveMQ Classic
> Issue Type: New Feature
> Reporter: Matt Pavlovich
> Assignee: Matt Pavlovich
> Priority: Minor
> Fix For: 6.2.0, 5.18.5
>
> Time Spent: 20m
> Remaining Estimate: 0h
>
> Support exporting and pulling messages from a MessageStore with an offset, to
> prevent loading _every_ message into memory (and subsequent required gc).
> For example, if a user wants to export the last message from a queue with
> 1,000 messages, they can scan over the index skipping most messages vs
> loading all messages into memory to get to the last message.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
