[ https://forge.continuent.org/jira/browse/HEDERA-47?page=all ]
Linas Virbalas resolved HEDERA-47:
----------------------------------
Resolution: Fixed
Fixed in CVS.
> JGroupsReliableGroupChannel should process SuspectEvent asynchronously
> ----------------------------------------------------------------------
>
> Key: HEDERA-47
> URL: https://forge.continuent.org/jira/browse/HEDERA-47
> Project: Hedera
> Type: Sub-task
> Versions: Hedera 2.0
> Reporter: Linas Virbalas
> Assignee: Linas Virbalas
> Fix For: Hedera 2.0
>
> Original Estimate: 3 days
> Remaining: 3 days
>
> Observations of Sequoia cluster which is ran on top of Hedera JGroups
> revealed that it might hang when one of the members fails. This is because
> the alive member receives a SuspectEvent about a failed member and it tries
> to cleanup all not completed transactions of it. It sends ResumeActivity
> message to a group (himself), waits for a reply, but never receives it,
> because it is processing SuspectEvent synchronously in the same thread. Thus
> this processing of SuspectEvent just hangs the GC channel.
> We should make SuspectEvent be processed asynchronously so new messages would
> still be received and potentially unblock SuspectEvent processing.
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
https://forge.continuent.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see:
http://www.atlassian.com/software/jira
_______________________________________________
Hedera mailing list
[email protected]
https://forge.continuent.org/mailman/listinfo/hedera
