hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Todd Lipcon (Commented) (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HDFS-2882) DN continues to start up, even if block pool fails to initialize
Date Wed, 04 Apr 2012 22:50:22 GMT

    [ https://issues.apache.org/jira/browse/HDFS-2882?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13246810#comment-13246810

Todd Lipcon commented on HDFS-2882:

Trying to think through what the correct/expected behavior is here:

Scenario 1: user configures DN to point to a single cluster which doesn't match its storage
Possible results:
1) DN keeps running, with no block pools
2) DN shuts down once its one-and-only service fails.

I think #2 is fairly clearly correct here.


Scenario 2: user configures DN to point to one NN. The user adds an additional nameservice
to the config and issues a -refreshNamenodes call. The newly added nameservice is from the
wrong cluster.
Possible results:
1) DN tries to connect and fails. It logs a message indicating this, but keeps running with
its existing service.
2) DN tries to connect and fails. It aborts the whole datanode.

I think #1 is correct here.

Scenario 3: user configures DN to point to two different NNs which are on different clusters,
and starts up.
Possible results:
1) DN connects to cluster 1, and joins the cluster. Service to Cluster 2 fails, but the DN
stays running. The admin may issue refreshNodes to try to connect again.
2) DN connects to cluster 1 and joins the cluster. When service to cluster 2 is rejected,
the DN shuts down.

It's unclear what the correct results are here. My leaning is towards #2, but not certain.
> DN continues to start up, even if block pool fails to initialize
> ----------------------------------------------------------------
>                 Key: HDFS-2882
>                 URL: https://issues.apache.org/jira/browse/HDFS-2882
>             Project: Hadoop HDFS
>          Issue Type: Bug
>          Components: data-node
>    Affects Versions: 0.24.0
>            Reporter: Todd Lipcon
>            Assignee: Todd Lipcon
> I started a DN on a machine that was completely out of space on one of its drives. I
saw the following:
> 2012-02-02 09:56:50,499 FATAL org.apache.hadoop.hdfs.server.datanode.DataNode: Initialization
failed for block pool Block pool BP-448349972- (storage id DS-507718931-
> 42002148) service to styx01.sf.cloudera.com/
> java.io.IOException: Mkdirs failed to create /data/1/scratch/todd/styx-datadir/current/BP-448349972-
>         at org.apache.hadoop.hdfs.server.datanode.FSDataset$BlockPoolSlice.<init>(FSDataset.java:335)
> but the DN continued to run, spewing NPEs when it tried to do block reports, etc. This
was on the HDFS-1623 branch but may affect trunk as well.

This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira


View raw message