[
https://issues.apache.org/jira/browse/HADOOP-5193?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12671716#action_12671716
]
Nigel Daley commented on HADOOP-5193:
-------------------------------------
{quote}
-1 tests included. The patch doesn't appear to include any new or modified
tests.
Please justify why no tests are needed for this patch.
{quote}
Konstantin, why no regression test?
Boris, why didn't you ask about this in the code review? Our checklist is
here: http://wiki.apache.org/hadoop/CodeReviewChecklist
> SecondaryNameNode does not rollImage because of incorrect calculation of
> edits modification time.
> -------------------------------------------------------------------------------------------------
>
> Key: HADOOP-5193
> URL: https://issues.apache.org/jira/browse/HADOOP-5193
> Project: Hadoop Core
> Issue Type: Bug
> Components: dfs
> Affects Versions: 0.19.0
> Reporter: Konstantin Shvachko
> Assignee: Konstantin Shvachko
> Fix For: 0.19.1
>
> Attachments: EditsLength.patch
>
>
> Secondary name-node cannot complete the second phase of the checkpoint
> because getFsEditsTime() returns the mod time of {{edits.new}} rather than
> {{edits}} file.
> The difference is that {{edits}} remains unchanged during the whole
> checkpoint process an therefore can serve as an invariant. On the contrary
> {{edits.new}} is changing all the time since it is the target of the edits
> log during checkpoint. So comparison of the mod time of {{edits.new}} before
> and after checkpoint fail and name-node does not upload new image file from
> the secondary node and does not truncate edits files.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
