[
https://issues.apache.org/jira/browse/HBASE-5074?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13213884#comment-13213884
]
Phabricator commented on HBASE-5074:
------------------------------------
stack has commented on the revision "[jira] [HBASE-5074] Support checksums in
HBase block cache".
Answering Dhruba.
INLINE COMMENTS
src/main/java/org/apache/hadoop/hbase/fs/HFileSystem.java:115 Seems like we
could have better names for these methods, ones that give more of a clue as to
what they are about. getBackingFS, getNoChecksumFS?
Maybe you are keepign them generic like this because you will be back in this
area again soon doing another beautiful speedup on top of this checksumming fix
(When we going to do read-ahead? Would that speed scanning?)
src/main/java/org/apache/hadoop/hbase/io/hfile/AbstractHFileReader.java:44
ok. np.
src/main/java/org/apache/hadoop/hbase/io/hfile/AbstractHFileReader.java:49
Ok. So, two readers. Our file count is going to go up? We should release
note this as side effect of enabling this feature (previous you may have been
well below xceivers limit but now you could go over the top?) I didn't notice
this was going on. Need to foreground it I'd say.
src/main/java/org/apache/hadoop/hbase/io/hfile/ChecksumUtil.java:84 I
figured. Its fine as is.
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, D1521.7.patch,
> D1521.7.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
