hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jonathan Gray (JIRA)" <j...@apache.org>
Subject [jira] Commented: (HBASE-2468) Improvements to prewarm META cache on clients
Date Sat, 22 May 2010 22:41:17 GMT

    [ https://issues.apache.org/jira/browse/HBASE-2468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12870330#action_12870330

Jonathan Gray commented on HBASE-2468:

A single table could be made up of many thousands even hundreds of thousands of regions? 
It could be significant.  Even if 1000 regions it could add a huge overhead to instantiating
an HTable a user may only use for a single Get.

As for the look-ahead, I think it makes sense if there is a cached value that gets evicted
to do a look-ahead of an extra row so you have a good chance of picking up the other side
of a split.  Beyond that, I don't think it's a good idea to add a default overhead because
this fetch is likely blocking a user data request waiting for a region to open and refresh
the location from meta.

I love these features and have wanted them for a while.  Just don't think we should change
default HTable behavior significantly.

> Improvements to prewarm META cache on clients
> ---------------------------------------------
>                 Key: HBASE-2468
>                 URL: https://issues.apache.org/jira/browse/HBASE-2468
>             Project: Hadoop HBase
>          Issue Type: Improvement
>          Components: client
>            Reporter: Todd Lipcon
>            Assignee: Mingjie Lai
>             Fix For: 0.21.0
>         Attachments: HBASE-2468-trunk.patch
> A couple different use cases cause storms of reads to META during startup. For example,
a large MR job will cause each map task to hit meta since it starts with an empty cache.
> A couple possible improvements have been proposed:
>  - MR jobs could ship a copy of META for the table in the DistributedCache
>  - Clients could prewarm cache by doing a large scan of all the meta for the table instead
of random reads for each miss
>  - Each miss could fetch ahead some number of rows in META

This message is automatically generated by JIRA.
You can reply to this email to add a comment to the issue online.

View raw message