You are viewing a plain text version of this content. The canonical link for it is here.
Posted to hdfs-issues@hadoop.apache.org by "Stephen O'Donnell (Jira)" <ji...@apache.org> on 2021/02/19 16:48:00 UTC

[jira] [Commented] (HDFS-15422) Reported IBR is partially replaced with stored info when queuing.

    [ https://issues.apache.org/jira/browse/HDFS-15422?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17287194#comment-17287194 ] 

Stephen O'Donnell commented on HDFS-15422:
------------------------------------------

We have had a report of something that sounds very like this problem. Appended blocks, failover and the blocks are marked corrupt. They are still readable etc. I suspect if we fail back over and restart the new SBNN it will clear it, but I am waiting to confirm that.

This is on Cloudera CDP 7.x, which is a heavily patched 3.1 build.

It looks like this problem is still there on trunk. From earlier comments, it sounds like a unit test for this is very difficult, so I will post a trunk patch with the small change [~kihwal] suggested and see what Yetus says.

> Reported IBR is partially replaced with stored info when queuing.
> -----------------------------------------------------------------
>
>                 Key: HDFS-15422
>                 URL: https://issues.apache.org/jira/browse/HDFS-15422
>             Project: Hadoop HDFS
>          Issue Type: Bug
>          Components: namenode
>            Reporter: Kihwal Lee
>            Priority: Critical
>         Attachments: HDFS-15422-branch-2.10.001.patch
>
>
> When queueing an IBR (incremental block report) on a standby namenode, some of the reported information is being replaced with the existing stored information.  This can lead to false block corruption.
> We had a namenode, after transitioning to active, started reporting missing blocks with "SIZE_MISMATCH" as corrupt reason. These were blocks that were appended and the sizes were actually correct on the datanodes. Upon further investigation, it was determined that the namenode was queueing IBRs with altered information.
> Although it sounds bad, I am not making it blocker 



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

---------------------------------------------------------------------
To unsubscribe, e-mail: hdfs-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-help@hadoop.apache.org