[
https://issues.apache.org/jira/browse/HBASE-2315?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13412829#comment-13412829
]
stack commented on HBASE-2315:
------------------------------
Flavio:
You have what you need? Jesse and Jon I think did a good job above filling in
bits I waved hands over when we talked (I like Jesse's suggestion undoing even
our dependency of FS).
When we talked, we also figured that hbase would need access to bk or the bkfs
or the bk interface at replay of edits time, on region open. In particular,
we'd need to have a factory like the WAL log factory here abouts in the code:
http://hbase.apache.org/xref/org/apache/hadoop/hbase/regionserver/HRegion.html#2721
This is called when the region is opened by the regionserver. It does all
its init and as part of the init, it looks to see if there are edits to replay.
If edits-to-replay are out in bk, we need to be able to get at them at this
very point.
> BookKeeper for write-ahead logging
> ----------------------------------
>
> Key: HBASE-2315
> URL: https://issues.apache.org/jira/browse/HBASE-2315
> Project: HBase
> Issue Type: New Feature
> Components: regionserver
> Reporter: Flavio Junqueira
> Attachments: HBASE-2315.patch, bookkeeperOverview.pdf,
> zookeeper-dev-bookkeeper.jar
>
>
> BookKeeper, a contrib of the ZooKeeper project, is a fault tolerant and high
> throughput write-ahead logging service. This issue provides an implementation
> of write-ahead logging for hbase using BookKeeper. Apart from expected
> throughput improvements, BookKeeper also has stronger durability guarantees
> compared to the implementation currently used by hbase.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
