[
https://issues.apache.org/jira/browse/HBASE-15264?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15149994#comment-15149994
]
stack commented on HBASE-15264:
-------------------------------
bq. I will stop renew lease for this file first, and then call
recoverFileLease.
There is that messyness where if you ask to recover the lease while the NN is
recovering the lease, it will start the recovery over from scratch. Do you
think that with the stopping of the lease renewal, maybe the recovery will run
faster. Four seconds seemed safe time to wait on lease recovery IIRC.
bq. ...I think finally namenode will close the file because of lease expired
even if the recoverFileLease failed?
Seems reasonable [~Apache9] We (or I) can test it. If all unacked edits are
into the new file, that should work. I think WAL replay will deal properly
where a WAL in the middle of a set of WALs needs its lease recovered.
That Payload looks like RingBufferTruck
Is the RingBuffer in your way? Let all the handlers come together on your queue?
Just skimmed AsyncFSWAL. It looks great. I love how it is so little code.
Looking forward to trying it out.
> Implement a fan out HDFS OutputStream
> -------------------------------------
>
> Key: HBASE-15264
> URL: https://issues.apache.org/jira/browse/HBASE-15264
> Project: HBase
> Issue Type: Sub-task
> Reporter: Duo Zhang
> Assignee: Duo Zhang
> Attachments: HBASE-15264-v1.patch, HBASE-15264-v2.patch,
> HBASE-15264.patch
>
>
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
