You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@cassandra.apache.org by "Jonathan Ellis (JIRA)" <ji...@apache.org> on 2015/05/12 17:36:01 UTC
[jira] [Commented] (CASSANDRA-9145) First RepairJob unregisters
RepairSession from FailureDetector
[ https://issues.apache.org/jira/browse/CASSANDRA-9145?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14540050#comment-14540050 ]
Jonathan Ellis commented on CASSANDRA-9145:
-------------------------------------------
Fix LGTM. Not sure that writing a dtest is feasible, but have you tested manually?
Nits: Extra newline left at the beginning of the AsyncFunction where the old code was removed, and some bogus indentation changed in StreamPlan.
> First RepairJob unregisters RepairSession from FailureDetector
> --------------------------------------------------------------
>
> Key: CASSANDRA-9145
> URL: https://issues.apache.org/jira/browse/CASSANDRA-9145
> Project: Cassandra
> Issue Type: Bug
> Components: Core
> Reporter: sankalp kohli
> Assignee: Yuki Morishita
> Fix For: 2.2 beta 1
>
> Attachments: 0001-unregister-FD-when-all-validations-are-done.patch
>
>
> Due to CASSANDRA-3569, we unregister FD when all Merkle trees are received. Till 2.1, we do this on last repair job of a session.
> In trunk(3.0), we are doing it after any one repair job has got all Merkle trees. This looks like a regression.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)