[
https://issues.apache.org/jira/browse/HBASE-10000?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13841843#comment-13841843
]
Ted Yu commented on HBASE-10000:
--------------------------------
Nicolas: thanks for the valuable comments.
Patch v4 should address all your comments. Let me answer the last several here
since they're important.
bq. may be it should be outside of the loop
Done.
bq. the previous versions was doing multiple calls.
I modified the condition so that calls are made when nbAttempt > 0.
bq. but ifFileClosed is not available, we will never succeed.
The behavior is not changed: we would rely on recoverLease() to be successful.
I modified the condition for the initial sleep according to your comments above.
bq. Could we say that 'isFileClosed' will be mandatory there?
There hasn't been consensus as to which release would have such pre-requisite.
So I want to keep the changes cover all supported hadoop releases.
bq. I looked especially at FSHDFSUtils
This is the part where more attention should be paid.
> Initiate lease recovery for outstanding WAL files at the very beginning of
> recovery
> -----------------------------------------------------------------------------------
>
> Key: HBASE-10000
> URL: https://issues.apache.org/jira/browse/HBASE-10000
> Project: HBase
> Issue Type: Improvement
> Reporter: Ted Yu
> Assignee: Ted Yu
> Fix For: 0.98.1
>
> Attachments: 10000-recover-ts-with-pb-2.txt,
> 10000-recover-ts-with-pb-3.txt, 10000-recover-ts-with-pb-4.txt, 10000-v1.txt,
> 10000-v4.txt, 10000-v5.txt, 10000-v6.txt
>
>
> At the beginning of recovery, master can send lease recovery requests
> concurrently for outstanding WAL files using a thread pool.
> Each split worker would first check whether the WAL file it processes is
> closed.
> Thanks to Nicolas Liochon and Jeffery discussion with whom gave rise to this
> idea.
--
This message was sent by Atlassian JIRA
(v6.1#6144)
