[
https://issues.apache.org/jira/browse/HDFS-4960?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13702832#comment-13702832
]
Varun Sharma commented on HDFS-4960:
------------------------------------
I don't know about the future plans of .meta - however, I think currently its
only being used for storing checksums and if checksums are not being asked for,
there is no need for seeking into the .meta file. I think the FB branch already
has this change. I personally think that inlining metadata in blocks is the way
to go, for the future, instead of storing separately in a .meta file - it
cripples hdfs for real time work loads.
I did not have a large enough data set so probably all the .meta files were in
fs cache.
lseek in .meta ~50 microseconds
lseek in block ~50 microseconds
read .meta ~50 microseconds
read block ~ 300 microseconds
So, just looking from system calls perspective, it is ~ 20 % however, it does
not show up in the end to end test, because in general, there are a bunch of
other contention issues inside HDFS + HBase which hog CPU resources.
> Unnecessary .meta seeks even when skip checksum is true
> -------------------------------------------------------
>
> Key: HDFS-4960
> URL: https://issues.apache.org/jira/browse/HDFS-4960
> Project: Hadoop HDFS
> Issue Type: Bug
> Affects Versions: 3.0.0, 2.1.0-beta
> Reporter: Varun Sharma
> Assignee: Varun Sharma
> Attachments: 4960-branch2.patch, 4960-trunk.patch
>
>
> While attempting to benchmark an HBase + Hadoop 2.0 setup on SSDs, we found
> unnecessary seeks into .meta files, each seek was a 7 byte read at the head
> of the file - this attempts to validate the version #. Since the client is
> requesting no-checksum, we should not be needing to touch the .meta file at
> all.
> Since the purpose of skip checksum is to also avoid the performance penalty
> of the extra seek, we should not be seeking into .meta if skip checksum is
> true
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
