[
https://issues.apache.org/jira/browse/HDFS-12641?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16383065#comment-16383065
]
Arpit Agarwal commented on HDFS-12641:
--------------------------------------
+1 on the change, it looks correct to me. I also see the new test passes even
after reverting the key part of the fix.
{code}
- if (!namesystem.isPopulatingReplQueues()) {
+ if (!namesystem.isPopulatingReplQueues() || !block.isComplete()) {
{code}
I haven't yet debugged what needs to be fixed in the test case.
> Backport HDFS-11755 into branch-2.7 to fix a regression in HDFS-11445
> ---------------------------------------------------------------------
>
> Key: HDFS-12641
> URL: https://issues.apache.org/jira/browse/HDFS-12641
> Project: Hadoop HDFS
> Issue Type: Task
> Affects Versions: 2.7.4
> Reporter: Wei-Chiu Chuang
> Assignee: Wei-Chiu Chuang
> Priority: Blocker
> Labels: release-blocker
> Attachments: HDFS-12641.branch-2.7.001.patch
>
>
> Our internal testing caught a regression in HDFS-11445 when we cherry picked
> the commit into CDH. Basically, it produces bogus missing file warnings.
> Further analysis revealed that the regression is actually fixed by HDFS-11755.
> Because of the order commits are merged in branch-2.8 ~ trunk (HDFS-11755 was
> committed before HDFS-11445), the regression was never actually surfaced for
> Hadoop 2.8/3.0.0-(alpha/beta) users. Since branch-2.7 has HDFS-11445 but no
> HDFS-11755, I suspect the regression is more visible for Hadoop 2.7.4.
> I am filing this jira to raise more awareness, than simply backporting
> HDFS-11755 into branch-2.7.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
