[
https://issues.apache.org/jira/browse/HBASE-20734?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16606132#comment-16606132
]
Reid Chan commented on HBASE-20734:
-----------------------------------
* {CommonFSUtils}#constructWALRegionDirFromRegionInfo
This name should get aligned with others, like getWALRegionDir for your
reference.
Please add java doc on new methods.
* {AbstractStateMachineTableProcedure.java}#getRegionDir is no longer used.
I think it is ok to remove it, since AbstractStateMachineTableProcedure is
IA.Private.
* {HRegion}
Please add java doc on new methods, and avoid unnecessary style changes like
below:
{code}
- + " recovered edits file(s) under " + regiondir);
+ + " recovered edits file(s) under " + regionDir);
{code}
Please update the description, since may not be HDFS.
{code}
// Store SeqId in HDFS when a region closes
// checking region folder exists is due to many tests which delete the
table folder while a
// table is still online
- if (this.fs.getFileSystem().exists(this.fs.getRegionDir())) {
- WALSplitter.writeRegionSequenceIdFile(this.fs.getFileSystem(),
this.fs.getRegionDir(),
+ if (getWalFileSystem().exists(getWALRegionDir())) {
{code}
Don't we add two new reference? i'm not sure about the ref count, what if
{regionDir} and {walFS} never get initialized which may not happen in real
world, but it is a logical problem.
{code}
+ 52 * ClassSize.REFERENCE
{code}
BTW, due to HBASE-21138, may need some efforts to fix conflicts.
> Colocate recovered edits directory with hbase.wal.dir
> -----------------------------------------------------
>
> Key: HBASE-20734
> URL: https://issues.apache.org/jira/browse/HBASE-20734
> Project: HBase
> Issue Type: Improvement
> Components: MTTR, Recovery, wal
> Reporter: Ted Yu
> Assignee: Zach York
> Priority: Major
> Fix For: 3.0.0
>
> Attachments: HBASE-20734.branch-1.001.patch,
> HBASE-20734.branch-1.002.patch, HBASE-20734.branch-1.003.patch,
> HBASE-20734.master.001.patch, HBASE-20734.master.002.patch,
> HBASE-20734.master.003.patch, HBASE-20734.master.004.patch,
> HBASE-20734.master.005.patch, HBASE-20734.master.006.patch,
> HBASE-20734.master.007.patch, HBASE-20734.master.008.patch,
> HBASE-20734.master.009.patch, HBASE-20734.master.010.patch
>
>
> During investigation of HBASE-20723, I realized that we wouldn't get the best
> performance when hbase.wal.dir is configured to be on different (fast) media
> than hbase rootdir w.r.t. recovered edits since recovered edits directory is
> currently under rootdir.
> Such setup may not result in fast recovery when there is region server
> failover.
> This issue is to find proper (hopefully backward compatible) way in
> colocating recovered edits directory with hbase.wal.dir .
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
