[jira] [Updated] (HBASE-14388) Compaction improvements: Avoid flush storms by jittering flush interval and max log files

["Vladimir Rodionov (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/HBASE-14388?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Vladimir Rodionov updated HBASE-14388:
--------------------------------------
    Attachment: HBASE-14388-v1.patch

* Periodic memstore flush changes
New configuration option: *hbase.regionserver.flush.jitter*. Default value: 0.5. This affects memstore flush interval (*hbase.regionserver.optionalcacheflushinterval*) and maximum changes per flush (*hbase.regionserver.flush.per.changes*) by randomizing them. After every flush operation, the new random value is assigned to memstore flush interval and to maximum changes per flush.

* Maximum WAL files changes
The configuration option *hbase.regionserver.maxlogs* is obsolete. The maximum number of WAL files is defined now dynamically:
{code}
 maxLogs = HEAP_SIZE * memstoreRatio * 2/ LogRollSize
{code} 


> Compaction improvements: Avoid flush storms by jittering flush interval and max log files 
> ------------------------------------------------------------------------------------------
>
>                 Key: HBASE-14388
>                 URL: https://issues.apache.org/jira/browse/HBASE-14388
>             Project: HBase
>          Issue Type: Improvement
>            Reporter: Vladimir Rodionov
>            Assignee: Vladimir Rodionov
>             Fix For: 2.0.0
>
>         Attachments: HBASE-14388-v1.patch
>
>




--
This message was sent by Atlassian JIRA
(v6.3.4#6332)