[
https://issues.apache.org/jira/browse/ARTEMIS-3045?focusedWorklogId=551814&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-551814
]
ASF GitHub Bot logged work on ARTEMIS-3045:
-------------------------------------------
Author: ASF GitHub Bot
Created on: 12/Feb/21 13:47
Start Date: 12/Feb/21 13:47
Worklog Time Spent: 10m
Work Description: franz1981 edited a comment on pull request #3392:
URL: https://github.com/apache/activemq-artemis/pull/3392#issuecomment-778040106
The change seems ok CI-wise, but, given that's running its logic on the
Netty event loop we need a couple of questions to be answered:
- what happen to the Netty channel writability with racing sends on a
different artemis `Channel` (eg `PING`)?
The `REPLICATION` channel isn't the only one using the underlying `Netty`
connection and that means that being awaken that a Netty channel is writable
again can make it not writable while attempting to write the replication
packets, because a concurrent write has filled it again!
- `ChannelImpl::send` can block? If yes, it can cause some trouble because
the cluster connection isn't the only citizen of the Netty event loop thread
and this can cause other connections to starve (to not mention the same
connection responses to be read!)
@clebertsuconic @jbertram @gtully
I believe that answering to this questions is key to understand how to make
this change safely or to verify that's already ok as it is...
I see that `ChannelImpl.CHANNEL_ID.REPLICATION` Artemis `Channel` is always
created with a `confWindowSize == -1` so it seems it won't have any blocking
behaviour caused by the `resendCache` or `responseAsyncCache` and given that it
cannot failover nor is using `sendBlocking`, its `ChannelImpl::lock` isn't
used, but is still brittle and I would like to enforce it.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
Issue Time Tracking
-------------------
Worklog Id: (was: 551814)
Time Spent: 8h 20m (was: 8h 10m)
> ReplicationManager can batch sent replicated packets
> ----------------------------------------------------
>
> Key: ARTEMIS-3045
> URL: https://issues.apache.org/jira/browse/ARTEMIS-3045
> Project: ActiveMQ Artemis
> Issue Type: Improvement
> Reporter: Francesco Nigro
> Assignee: Francesco Nigro
> Priority: Major
> Time Spent: 8h 20m
> Remaining Estimate: 0h
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
