[
https://issues.apache.org/jira/browse/HADOOP-2038?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
stack updated HADOOP-2038:
--------------------------
Attachment: delay.patch
Here's a patch. Below is commit message.
{code}
D
src/contrib/hbase/src/test/org/apache/hadoop/hbase/TestCleanRegionServerExit.java
D src/contrib/hbase/src/test/org/apache/hadoop/hbase/TestRegionServerAbort.java
Removed.
A src/contrib/hbase/src/test/org/apache/hadoop/hbase/TestRegionServerExit.java
The two above tests have been amalgamated here.
M src/contrib/hbase/src/java/org/apache/hadoop/hbase/HMaster.java
(PendingRegionShutdown.getDelay): Return difference between creation time
and current time.
{code}
> [hbase] TestCleanRegionServerExit failed in patch build #927
> ------------------------------------------------------------
>
> Key: HADOOP-2038
> URL: https://issues.apache.org/jira/browse/HADOOP-2038
> Project: Hadoop
> Issue Type: Bug
> Components: contrib/hbase
> Reporter: stack
> Assignee: stack
> Priority: Minor
> Fix For: 0.15.0
>
> Attachments: delay.patch
>
>
> Looking at the hudson console logs, the region server with -ROOT- region went
> away. Usually this is not the region server that goes down. Its usually the
> one hosting other regions. I noticed a while ago that we don't recover from
> this -ROOT- gone situation (where once we used to). Chatting with Jim, a
> shutdown event gets queued and is supposed to be run after a 1/2 lease has
> expired. I noticed that shutdown's getDelay always returned 1/2 a lease.
> Reading DelayQueue, the queue will not return elements on poll unless the
> queue'd items getDelay returns a value of <= 0.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
