accumulo-notifications mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Keith Turner (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (ACCUMULO-1172) Increase metadata table split threshold
Date Tue, 12 Mar 2013 19:27:12 GMT

    [ https://issues.apache.org/jira/browse/ACCUMULO-1172?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13600372#comment-13600372
] 

Keith Turner commented on ACCUMULO-1172:
----------------------------------------

bq. Make the datablock cache default the same size so it's in memory in the default case.

Good point.  Also the cache holds uncompressed blocks.  So maybe set splits size to 64M and
 cache size to 128M by default.
                
> Increase metadata table split threshold
> ---------------------------------------
>
>                 Key: ACCUMULO-1172
>                 URL: https://issues.apache.org/jira/browse/ACCUMULO-1172
>             Project: Accumulo
>          Issue Type: Bug
>            Reporter: Keith Turner
>            Assignee: Keith Turner
>             Fix For: 1.5.0
>
>
> The metadata table split threshold is initialized to 4M.  This is very small and can
cause a system to have alots of metadata table tablets.  This was set small a long time ago
when tablets did not have read parallelism.  That has not been the case for a long time, but
this threshold is still small.  I propose increasing it to 64M or 128M.

--
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

Mime
View raw message