hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Todd Lipcon (JIRA)" <j...@apache.org>
Subject [jira] Commented: (HDFS-955) FSImage.saveFSImage can lose edits
Date Wed, 10 Feb 2010 21:48:27 GMT

    [ https://issues.apache.org/jira/browse/HDFS-955?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12832220#action_12832220
] 

Todd Lipcon commented on HDFS-955:
----------------------------------

Unfortunately HDFS-957 I think provides a fix for both of the examples above where the correct
recovery in state 5 is IMAGE_NEW. However, in the case where we've just received an actual
checkpoint but haven't rolled yet, the correct recovery is IMAGE_NEW + EDITS_NEW. I don't
have a good way of distinguishing that situation from the manual saveNamespace where we don't
want to recover EDITS_NEW.

I don't know that we can actually fix this with the current set of files and determinable
state. I'll continue to think.

> FSImage.saveFSImage can lose edits
> ----------------------------------
>
>                 Key: HDFS-955
>                 URL: https://issues.apache.org/jira/browse/HDFS-955
>             Project: Hadoop HDFS
>          Issue Type: Bug
>    Affects Versions: 0.20.1, 0.21.0, 0.22.0
>            Reporter: Todd Lipcon
>            Assignee: Todd Lipcon
>            Priority: Blocker
>         Attachments: hdfs-955-unittest.txt, PurgeEditsBeforeImageSave.patch
>
>
> This is a continuation of a discussion from HDFS-909. The FSImage.saveFSImage function
(implementing dfsadmin -saveNamespace) can corrupt the NN storage such that all current edits
are lost.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message