hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "nkeywal (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HDFS-3703) Decrease the datanode failure detection time
Date Wed, 05 Sep 2012 18:53:08 GMT

    [ https://issues.apache.org/jira/browse/HDFS-3703?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13448997#comment-13448997
] 

nkeywal commented on HDFS-3703:
-------------------------------

Hi,

I've done a test on HBase, with the minicluster. Basically it works :-). I've added a check
on the write path as well.
The test consists of stopping a datanode and a regionserver while keeping the socket opened
to get timeouts on a minimal cluster (3 nodes, replication level 2), after having written
10 millions lines.  Without the fix, HBase needs 15 minutes to recover: it gets timeouts all
other the place because it goes to the dead datanodes. It starts to works only after the node
is marked as dead in the NN. With the fix, in 1 minute it's done.

I had to add the write path as well, because during the recovery HBase writes many small files
(hundreds actually: that's splitting the logs). Without the check on the write path we get
directed to the dead DN for all the files.

Attached: the original patch rebased on branch-2 + the write path.










                
> Decrease the datanode failure detection time
> --------------------------------------------
>
>                 Key: HDFS-3703
>                 URL: https://issues.apache.org/jira/browse/HDFS-3703
>             Project: Hadoop HDFS
>          Issue Type: Improvement
>          Components: data-node, name-node
>    Affects Versions: 1.0.3, 2.0.0-alpha
>            Reporter: nkeywal
>            Assignee: Suresh Srinivas
>         Attachments: HDFS-3703-branch2.patch, HDFS-3703.patch
>
>
> By default, if a box dies, the datanode will be marked as dead by the namenode after
10:30 minutes. In the meantime, this datanode will still be proposed  by the nanenode to write
blocks or to read replicas. It happens as well if the datanode crashes: there is no shutdown
hooks to tell the nanemode we're not there anymore.
> It especially an issue with HBase. HBase regionserver timeout for production is often
30s. So with these configs, when a box dies HBase starts to recover after 30s and, while 10
minutes, the namenode will consider the blocks on the same box as available. Beyond the write
errors, this will trigger a lot of missed reads:
> - during the recovery, HBase needs to read the blocks used on the dead box (the ones
in the 'HBase Write-Ahead-Log')
> - after the recovery, reading these data blocks (the 'HBase region') will fail 33% of
the time with the default number of replica, slowering the data access, especially when the
errors are socket timeout (i.e. around 60s most of the time). 
> Globally, it would be ideal if HDFS settings could be under HBase settings. 
> As a side note, HBase relies on ZooKeeper to detect regionservers issues.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

Mime
View raw message