[
https://issues.apache.org/jira/browse/HBASE-4363?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13104104#comment-13104104
]
Jean-Daniel Cryans edited comment on HBASE-4363 at 9/14/11 12:34 AM:
---------------------------------------------------------------------
The stopper is in fact the region server itself, this.running is if that
particular source is actively replicating or not. One case where the RS is
running but not a source is when you close a replication stream (which happened
for us can caused the problem described in this jira).
Not a fan of isReplicating, sort of overloaded. Thought of "isSourceActive",
but it doesn't convey the fact that the region server could be the cause for
shutting down.
was (Author: jdcryans):
The stopper is in fact the region server itself, this.running is if that
particular source is actively replicating or not. One case where the RS is
running but not a source is when you close a replication stream (which happened
for us can caused the problem described in this jira).
Not a fin of isReplicating, sort of overloaded. Thought of "isSourceActive",
but it doesn't convey the fact that the region server could be the cause for
shutting down.
> [replication] ReplicationSource won't close if failing to contact the sink
> --------------------------------------------------------------------------
>
> Key: HBASE-4363
> URL: https://issues.apache.org/jira/browse/HBASE-4363
> Project: HBase
> Issue Type: Bug
> Affects Versions: 0.90.4
> Reporter: Jean-Daniel Cryans
> Fix For: 0.90.5
>
> Attachments: HBASE-4363-0.90.patch
>
>
> When trying to close a source, it will hang if it's already in shipEdits()
> and has issues reaching the sink. The reason is that in that method the while
> loop only checks if the RS is going down but not if the source was asked to
> shutdown.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
