[ 
https://issues.apache.org/jira/browse/HBASE-23326?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17889480#comment-17889480
 ] 

Vineet Kumar Maheshwari commented on HBASE-23326:
-------------------------------------------------

Please find configuration details and observations:

hbase. master. store. region. compact. min = 4 (default value)

Number of store files are higher than this value, triggering the compaction.

But in 
[https://github.com/apache/hbase/blob/master/hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/compactions/SortedCompactionPolicy.java#L228]
 minFiles criteria does not match. So compaction does not happen. Here minFiles 
is calculated based on configurations: hbase. hstore. compaction. min or hbase. 
hstore. compactionThreshold

Value for hbase. hstore. compactionThreshold = 10.

This leads to compaction getting triggered continuously.

 

Observations:

1. Debug log is written continously

2. Config mismatch for compaction, hbase. master. store. region. compact. min 
is not used by below layer code leading to no compaction

3. IMO we can change compact function in MasterRegionFlusherAndCompactor, 
compact function no need to schedule next compaction. Let it get schedule based 
on flusher flow.

> Implement a ProcedureStore which stores procedures in a HRegion
> ---------------------------------------------------------------
>
>                 Key: HBASE-23326
>                 URL: https://issues.apache.org/jira/browse/HBASE-23326
>             Project: HBase
>          Issue Type: Improvement
>          Components: proc-v2
>            Reporter: Duo Zhang
>            Assignee: Duo Zhang
>            Priority: Critical
>             Fix For: 3.0.0-alpha-1, 2.3.0
>
>
> So we can resue the code in HRegion for persisting the procedures, and also 
> the optimized WAL implementation for better performance.
> This requires we merge the hbase-procedure module to hbase-server, which is 
> an anti-pattern as we make the hbase-server module more overloaded. But I 
> think later we can first try to move the WAL stuff out.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

Reply via email to