[
https://issues.apache.org/jira/browse/HBASE-5074?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13210717#comment-13210717
]
Phabricator commented on HBASE-5074:
------------------------------------
stack has commented on the revision "[jira] [HBASE-5074] Support checksums in
HBase block cache".
I got about 15% through. Will do rest later. This stuff is great.
INLINE COMMENTS
src/main/java/org/apache/hadoop/hbase/HConstants.java:605 Nice doc. Lets
hoist up into the reference manual on commit.
src/main/java/org/apache/hadoop/hbase/fs/HFileSystem.java:1 Good. I think
its better having it in here.
src/main/java/org/apache/hadoop/hbase/fs/HFileSystem.java:115 I see we use
this writing the WAL. Reading we'll use whatever the readfs? Do we need to
expose this? Or the getReadRS even?
Or is it that you want different fs's for read and write? If so, should this
method be called getWriteFS?
src/main/java/org/apache/hadoop/hbase/fs/HFileSystem.java:129 Post creation,
invoking this method would have no effect? If so, remove, and make this data
member final?
src/main/java/org/apache/hadoop/hbase/io/hfile/AbstractHFileReader.java:44
Why change this comment? Do we care how it does checksumming?
src/main/java/org/apache/hadoop/hbase/io/hfile/AbstractHFileReader.java:49
Yeah, I wonder if upper tiers need worry about this stuff? Whether its
checksummed or not? Should they just be talking about readfs vs writefs? And
then its up to the configuration as to what the underlying fs does (in this
case its just turning off hdfs checksumming). Looks like actual checksumming
is over in HFileBlock... maybe HFile itself doesn't need to be concerned w/
checksumming?
No biggie. Just a comment.
REVISION DETAIL
https://reviews.facebook.net/D1521
> support checksums in HBase block cache
> --------------------------------------
>
> Key: HBASE-5074
> URL: https://issues.apache.org/jira/browse/HBASE-5074
> Project: HBase
> Issue Type: Improvement
> Components: regionserver
> Reporter: dhruba borthakur
> Assignee: dhruba borthakur
> Attachments: D1521.1.patch, D1521.1.patch, D1521.2.patch,
> D1521.2.patch, D1521.3.patch, D1521.3.patch, D1521.4.patch, D1521.4.patch,
> D1521.5.patch, D1521.5.patch, D1521.6.patch, D1521.6.patch
>
>
> The current implementation of HDFS stores the data in one block file and the
> metadata(checksum) in another block file. This means that every read into the
> HBase block cache actually consumes two disk iops, one to the datafile and
> one to the checksum file. This is a major problem for scaling HBase, because
> HBase is usually bottlenecked on the number of random disk iops that the
> storage-hardware offers.
--
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
