[
https://issues.apache.org/jira/browse/ARTEMIS-4193?focusedWorklogId=850030&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-850030
]
ASF GitHub Bot logged work on ARTEMIS-4193:
-------------------------------------------
Author: ASF GitHub Bot
Created on: 09/Mar/23 10:18
Start Date: 09/Mar/23 10:18
Worklog Time Spent: 10m
Work Description: gemmellr commented on code in PR #4395:
URL: https://github.com/apache/activemq-artemis/pull/4395#discussion_r1130770649
##########
artemis-server/src/main/java/org/apache/activemq/artemis/core/server/impl/ActiveMQServerImpl.java:
##########
@@ -3381,10 +3382,24 @@ synchronized void initialisePart2(boolean scalingDown)
throws Exception {
pagingManager.reloadStores();
- JournalLoadInformation[] journalInfo = loadJournals();
+ Set<Long> storedLargeMessages = new HashSet<>();
+ JournalLoadInformation[] journalInfo = loadJournals(storedLargeMessages);
if (rebuildCounters) {
- pagingManager.rebuildCounters();
+ pagingManager.rebuildCounters(storedLargeMessages);
+
+ pagingManager.execute(() -> {
+ storedLargeMessages.forEach(id -> {
+ try {
+ SequentialFile file =
storageManager.createFileForLargeMessage(id, true);
+ logger.debug("Removing pending large message for file={}",
file);
+ file.delete();
+ } catch (Exception e) {
+ // this shouldn't really happen, unless something is off
with the storage
+ logger.warn("It was not possible to remove previously stored
large message on folder::It will be retried on next startup::" +
e.getMessage(), e);
Review Comment:
Dont think you need to add the exception message, when adding the exception
itself, just goign to end up with the message twice. (But if doing so, should
use a placeholder rather than concat).
It would however perhaps be good to add at least the message ID that the
failure related to though otherwise it may just end up printing a bunch of
similar or identical entries that dont make clear which message/files it
relates to.
Issue Time Tracking
-------------------
Worklog Id: (was: 850030)
Time Spent: 2h 20m (was: 2h 10m)
> Interrupting Large Message Streaming with a server kill may leave orphaned
> files
> --------------------------------------------------------------------------------
>
> Key: ARTEMIS-4193
> URL: https://issues.apache.org/jira/browse/ARTEMIS-4193
> Project: ActiveMQ Artemis
> Issue Type: Bug
> Affects Versions: 2.28.0
> Reporter: Clebert Suconic
> Priority: Major
> Fix For: 2.29.0
>
> Time Spent: 2h 20m
> Remaining Estimate: 0h
>
> There's a schema in the journal to store pending records before a file is
> created.
> However the sync is not properly applied and if the server is killed it could
> leave a few messages orphaned in the file system.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
