[
https://issues.apache.org/jira/browse/HBASE-2315?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12844634#action_12844634
]
stack commented on HBASE-2315:
------------------------------
Thanks for the interesting comments.
WAL is per regionserver, not per region.
On WAL throughput, since the 'append' to the WAL is what takes the bulk of the
time making an hbase update, depending on hardware and how often we flush,
we'll see variously 1-10k writes a second. We wil see hdfs holding up writes
on occasion when struggling.. with latency going way up.
I'm up for more exploration if you fellas are. Let me take a look at this
patch of yours Flavio.
Thanks for happening by lads.
> BookKeeper for write-ahead logging
> ----------------------------------
>
> Key: HBASE-2315
> URL: https://issues.apache.org/jira/browse/HBASE-2315
> Project: Hadoop HBase
> Issue Type: New Feature
> Components: regionserver
> Reporter: Flavio Paiva Junqueira
> Attachments: HBASE-2315.patch, 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.
-
You can reply to this email to add a comment to the issue online.
