[
https://issues.apache.org/jira/browse/HADOOP-1710?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Jim Kellerman resolved HADOOP-1710.
-----------------------------------
Resolution: Fixed
Commited changes for this issue.
They run on Fedora Core 6 (dual CPU) and Windows XP Pro (single CPU)
Changes:
HTable
- deprecate all the batch methods: startBatchUpdate, abortBatch,
commitBatch. These methods just fall through to startUpdate, abort,
commit respectively as they implement batch update now.
- renewLease is deprecated and now a no-op
HRegionInterface
- deprecate all the single rpc update methods: startUpdate, put,
delete, abort, commit, renewLease. HRegionServer clients should use
the batchUpdate interface instead.
HMaster
- make all updates batch updates. Since HMaster uses HRegionInterface
directly instead of HTable, it must use HRegionInterface.batchUpdate
instead of the deprecated methods on HRegionInterface
HConnectionManager
- fix bug in which meta servers were being "re-found" but the server
cache was not being updated.
HClient, HMerge, HRegionServer, TestBatchUpdate, TestHTable,
TestScanner2, TestSplit
- update corresponding to changes in HTable
> [hbase] All updates should be batch updates
> -------------------------------------------
>
> Key: HADOOP-1710
> URL: https://issues.apache.org/jira/browse/HADOOP-1710
> Project: Hadoop
> Issue Type: Improvement
> Components: contrib/hbase
> Affects Versions: 0.15.0
> Reporter: Jim Kellerman
> Assignee: Jim Kellerman
> Fix For: 0.15.0
>
>
> All updates whether done by the master, region server or client should be
> batch updates
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
