[ 
https://issues.apache.org/jira/browse/HBASE-23349?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17001958#comment-17001958
 ] 

Lars Hofhansl commented on HBASE-23349:
---------------------------------------

I think we should step back and remember why we have the ref counting in the 
first place. This came from a discussion started in HBASE-13082 and 
HBASE-10060, namely too much synchronization.

If any changes we make now needs new synchronization in the scanner.next(...) 
path we're back to where we started and in that case we should remove the ref 
counting and bring back the old notification and scanner switching we had 
before.

My apologies that I had triggered the original discussion, and then completely 
dropped off (worked on other stuff) when we attempted to fix it. Reference 
counting is bad (I've never seen this successful implemented), if we can avoid 
it we should a bit of performance drop is acceptable.

Long story for: If we bring back scanner notification then let's get rid of ref 
counting completely.

 

> Low refCount preventing archival of compacted away files
> --------------------------------------------------------
>
>                 Key: HBASE-23349
>                 URL: https://issues.apache.org/jira/browse/HBASE-23349
>             Project: HBase
>          Issue Type: Improvement
>    Affects Versions: 3.0.0, 2.3.0, 1.6.0
>            Reporter: Viraj Jasani
>            Assignee: Viraj Jasani
>            Priority: Major
>             Fix For: 3.0.0, 2.3.0, 1.6.0
>
>
> We have observed that refCount on compacted away store files as low as 1 is 
> prevent archival.
> {code:java}
> regionserver.HStore - Can't archive compacted file 
> hdfs://{{root-dir}}/hbase/data/default/t1/12a9e1112e0371955b3db8d3ebb2d298/cf1/73b72f5ddfce4a34a9e01afe7b83c1f9
>  because of either isCompactedAway=true or file has reference, 
> isReferencedInReads=true, refCount=1, skipping for now.
> {code}
> We should come up with core code (run as part of discharger thread) 
> gracefully resolve reader lock issue by resetting ongoing scanners to start 
> pointing to new store files instead of compacted away store files.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

Reply via email to