[
https://issues.apache.org/jira/browse/HDFS-2085?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13052755#comment-13052755
]
Todd Lipcon commented on HDFS-2085:
-----------------------------------
That's an interesting point... there's not currently any test that checks for
that, and in fact END_LOG_SEGMENT is basically a placeholder at this point. The
idea is that in the future it might be used to contain some statistics about
what's in the log, etc.
On the pro side, it would be nice to do this to make it an invariant that
finalized log files include END_LOG_SEGMENT. On the con side, it involves
reopening a log for append which isn't an operation we currently do anywhere
else.
Mind if we open this as another JIRA since it doesn't currently cause any
issues that I can think of?
> 1073: finalize inprogress edit logs at startup
> ----------------------------------------------
>
> Key: HDFS-2085
> URL: https://issues.apache.org/jira/browse/HDFS-2085
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Components: name-node
> Affects Versions: Edit log branch (HDFS-1073)
> Reporter: Todd Lipcon
> Assignee: Todd Lipcon
> Fix For: Edit log branch (HDFS-1073)
>
> Attachments: hdfs-2085.txt
>
>
> With HDFS-2074, the NameNode can read through any "in-progress" logs it finds
> during startup to determine how many transactions they have. It can then
> re-name the file from its inprogress name to its finalized name. For example,
> if it finds a file edits_10_inprogress with 3 transactions, it can rename it
> to edits_10-12 at startup. This means that other parts of the system like
> edits-log-transfer don't need to worry about in-progress logs.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
