hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Suresh Srinivas (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HDFS-4208) NameNode could be stuck in SafeMode due to never-created blocks in branch-1
Date Tue, 20 Nov 2012 22:24:58 GMT

    [ https://issues.apache.org/jira/browse/HDFS-4208?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13501548#comment-13501548
] 

Suresh Srinivas commented on HDFS-4208:
---------------------------------------

Comments:
# getCompleteBlocksTotal()
#* Should we name this as getSafeBlockCount()?
#* javadoc could be "Get the total number of COMPLETE blocks in the system. The last block
of files that are under construction that are zero length are not counted towards the COMPLETE
blocks. This avoids newly created blocks that might "
#* Instead of adding assert, we should count under construction blocks only if the inode is
not null and is underconstruction. The current code could have class cast exception or NPE
and might results in namenode not coming up, in case some weird errors.
#* We should also add to the LOG.info number of safeblocks and total blocks.
# To make the test case more clear, please add safeblock threshold to the configuration.

                
> NameNode could be stuck in SafeMode due to never-created blocks in branch-1
> ---------------------------------------------------------------------------
>
>                 Key: HDFS-4208
>                 URL: https://issues.apache.org/jira/browse/HDFS-4208
>             Project: Hadoop HDFS
>          Issue Type: Bug
>          Components: name-node
>    Affects Versions: 1.2.0
>            Reporter: Brandon Li
>            Assignee: Brandon Li
>         Attachments: HDFS-4208.branch-1.patch
>
>
> In one test case, NameNode allocated a block and then was killed before the client got
the addBlock response. After NameNode restarted, it couldn't get out of SafeMode waiting for
the block which was never created. In trunk, NameNode can get out of SafeMode since it only
counts complete blocks. However branch-1 doesn't have the clear notion of under-constructioned-block
in Namenode. 
> JIRA HDFS-4212 is to track the never-created-block issue and this JIRA is to fix NameNode
in branch-1 so it can get out of SafeMode when never-created-block exists.
> The proposed idea is for SafeMode not to count the zero-sized last block in an under-construction
file as part of total blcok count.

--
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