[
https://issues.apache.org/jira/browse/ARTEMIS-4668?focusedWorklogId=908359&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-908359
]
ASF GitHub Bot logged work on ARTEMIS-4668:
-------------------------------------------
Author: ASF GitHub Bot
Created on: 05/Mar/24 15:50
Start Date: 05/Mar/24 15:50
Worklog Time Spent: 10m
Work Description: clebertsuconic commented on code in PR #4840:
URL: https://github.com/apache/activemq-artemis/pull/4840#discussion_r1513065063
##########
artemis-protocols/artemis-amqp-protocol/src/main/java/org/apache/activemq/artemis/protocol/amqp/proton/AMQPLargeMessageReader.java:
##########
@@ -77,39 +73,73 @@ public AMQPLargeMessageReader open() {
}
@Override
- public Message readBytes(Delivery delivery) throws Exception {
+ public boolean readBytes(Delivery delivery) throws Exception {
if (closed) {
throw new IllegalStateException("AMQP Large Message Reader is closed
and read cannot proceed");
}
+ serverReceiver.connection.requireInHandler();
+
final Receiver receiver = ((Receiver) delivery.getLink());
final ReadableBuffer dataBuffer = receiver.recv();
+ final AMQPSessionCallback sessionSPI =
serverReceiver.getSessionContext().getSessionSPI();
+
if (currentMessage == null) {
- final AMQPSessionCallback sessionSPI =
serverReceiver.getSessionContext().getSessionSPI();
final long id = sessionSPI.getStorageManager().generateID();
currentMessage = new AMQPLargeMessage(id,
delivery.getMessageFormat(), null,
sessionSPI.getCoreMessageObjectPools(),
sessionSPI.getStorageManager());
currentMessage.parseHeader(dataBuffer);
-
+ logger.trace("Initializing current message {} on {}", currentMessage,
this);
sessionSPI.getStorageManager().largeMessageCreated(id,
currentMessage);
+ sessionSPI.execute(() -> validateFile(currentMessage));
Review Comment:
It still needs to create the file. which should happen away from the Netty
Thread.
in case of an exception it will initialize the exception on the session...
The current writes are still going through though but the exception would be
initialized. I could interrupt the addBytes if there was an exception, but the
possibility of having an exception there is actually so low that if any issues
are happening with the storage, other things will happen to put the server down
anyways through IOExceptions.
Issue Time Tracking
-------------------
Worklog Id: (was: 908359)
Time Spent: 3h (was: 2h 50m)
> Move AMQP Large Message File Handling away from Netty thread
> ------------------------------------------------------------
>
> Key: ARTEMIS-4668
> URL: https://issues.apache.org/jira/browse/ARTEMIS-4668
> Project: ActiveMQ Artemis
> Issue Type: Bug
> Affects Versions: 2.32.0
> Reporter: Clebert Suconic
> Assignee: Clebert Suconic
> Priority: Major
> Fix For: 2.33.0
>
> Time Spent: 3h
> Remaining Estimate: 0h
>
> Operations like file.open, file.close, and file.sync should be moved away
> from the Netty Thread for AMQP Large Messages
> This task now is about moving the processing for AMQP Messages. we may in a
> near future also improve tunneled large messages. For now we will do for AMQP
> messages only.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
