[
https://issues.apache.org/jira/browse/HDFS-1073?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12928488#action_12928488
]
Konstantin Shvachko commented on HDFS-1073:
-------------------------------------------
- Sanjay, Todd means that when a checkpoint starts it triggers rollEdits(),
which is the cut off point for the new checkpoint. The checkpoint of course can
use the latest rolled edits instead, but then you get the problem of
synchronizing the start of a checkpoint and the edits roll event. Otherwise
checkpoints may become way behind the current namespace state.
- I agree with Rob that edits_100-100 should not be a special case to avoid. In
practice we will not see it, but if it happens the system should just absorb
it. Todd correctly points out that if the system is idle for a very long time
NN may try to create edits_100-100 the second time, but this could be just
avoided based on name collision.
- I do not see or did not understand the rational for "I'm quitting!" record.
Why should NN care whether last record was lost or not, just keep going with
what it has. Worked so far.
> Simpler model for Namenode's fs Image and edit Logs
> ----------------------------------------------------
>
> Key: HDFS-1073
> URL: https://issues.apache.org/jira/browse/HDFS-1073
> Project: Hadoop HDFS
> Issue Type: Improvement
> Reporter: Sanjay Radia
> Assignee: Todd Lipcon
> Attachments: hdfs-1073.txt, hdfs1073.pdf
>
>
> The naming and handling of NN's fsImage and edit logs can be significantly
> improved resulting simpler and more robust code.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
