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 2009/01/18 06:50:59 UTC

[jira] Commented: (HBASE-1058) Prevent runaway compactions

    [ https://issues.apache.org/jira/browse/HBASE-1058?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12664927#action_12664927 ] 

stack commented on HBASE-1058:
------------------------------

You can overrun a cluster by running a PE upload into a small cluster.  Do './bin/hadoop ......PerformanceEvaluation sequentialWrite 256' ... i.e. try loading 1/4 billion items.  Trying this I produced a region with 1k flush files to compact.  It won't recover, not inside 1G of heap.   Just OOMEs.

> Prevent runaway compactions
> ---------------------------
>
>                 Key: HBASE-1058
>                 URL: https://issues.apache.org/jira/browse/HBASE-1058
>             Project: Hadoop HBase
>          Issue Type: Bug
>            Reporter: stack
>
> A rabid upload will easily outrun our compaction ability dropping flushes faster than we can compact them up.  Fix.

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