You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@hbase.apache.org by "stack (JIRA)" <ji...@apache.org> on 2008/07/26 20:40:31 UTC
[jira] Created: (HBASE-779) Test changing
hbase.hregion.memcache.block.multiplier to 2
Test changing hbase.hregion.memcache.block.multiplier to 2
----------------------------------------------------------
Key: HBASE-779
URL: https://issues.apache.org/jira/browse/HBASE-779
Project: Hadoop HBase
Issue Type: Improvement
Reporter: stack
Priority: Blocker
Fix For: 0.2.0
Currently its set to 1. Under load, seeing fill-cache/pause-while-cache-is-flushed cycle. Changing multiplier to 2 could make hbase take on load faster (Was set to 1 because compactions used to overwhelm the system but now HBASE-745 added a not-so-dumb compaction algorithim, we might be able to run w/ a value of 2).
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Assigned: (HBASE-779) Test changing
hbase.hregion.memcache.block.multiplier to 2
Posted by "stack (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-779?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
stack reassigned HBASE-779:
---------------------------
Assignee: stack
> Test changing hbase.hregion.memcache.block.multiplier to 2
> ----------------------------------------------------------
>
> Key: HBASE-779
> URL: https://issues.apache.org/jira/browse/HBASE-779
> Project: Hadoop HBase
> Issue Type: Improvement
> Reporter: stack
> Assignee: stack
> Priority: Blocker
> Fix For: 0.2.0
>
>
> Currently its set to 1. Under load, seeing fill-cache/pause-while-cache-is-flushed cycle. Changing multiplier to 2 could make hbase take on load faster (Was set to 1 because compactions used to overwhelm the system but now HBASE-745 added a not-so-dumb compaction algorithim, we might be able to run w/ a value of 2).
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Resolved: (HBASE-779) Test changing
hbase.hregion.memcache.block.multiplier to 2
Posted by "stack (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-779?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
stack resolved HBASE-779.
-------------------------
Resolution: Fixed
Committed. Ran on cluster (with optional flushing disabled so didn't trigger HBASE-751). Ran with 8 clients. Ran without blocking.
> Test changing hbase.hregion.memcache.block.multiplier to 2
> ----------------------------------------------------------
>
> Key: HBASE-779
> URL: https://issues.apache.org/jira/browse/HBASE-779
> Project: Hadoop HBase
> Issue Type: Improvement
> Reporter: stack
> Priority: Blocker
> Fix For: 0.2.0
>
>
> Currently its set to 1. Under load, seeing fill-cache/pause-while-cache-is-flushed cycle. Changing multiplier to 2 could make hbase take on load faster (Was set to 1 because compactions used to overwhelm the system but now HBASE-745 added a not-so-dumb compaction algorithim, we might be able to run w/ a value of 2).
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HBASE-779) Test changing
hbase.hregion.memcache.block.multiplier to 2
Posted by "stack (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-779?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12617221#action_12617221 ]
stack commented on HBASE-779:
-----------------------------
In todd lipcon testing, this parameter set to 2 seems to increase rate at which we take on writes considerably (duh!) and until he hit HBASE-751, all seemed to be running smoothly.
> Test changing hbase.hregion.memcache.block.multiplier to 2
> ----------------------------------------------------------
>
> Key: HBASE-779
> URL: https://issues.apache.org/jira/browse/HBASE-779
> Project: Hadoop HBase
> Issue Type: Improvement
> Reporter: stack
> Priority: Blocker
> Fix For: 0.2.0
>
>
> Currently its set to 1. Under load, seeing fill-cache/pause-while-cache-is-flushed cycle. Changing multiplier to 2 could make hbase take on load faster (Was set to 1 because compactions used to overwhelm the system but now HBASE-745 added a not-so-dumb compaction algorithim, we might be able to run w/ a value of 2).
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.