hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Lars Hofhansl (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HBASE-13082) Coarsen StoreScanner locks to RegionScanner
Date Mon, 02 Mar 2015 05:21:05 GMT

    [ https://issues.apache.org/jira/browse/HBASE-13082?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14342765#comment-14342765
] 

Lars Hofhansl commented on HBASE-13082:
---------------------------------------

Oh I get that :)
What I meant to say is that under no circumstances should fewer locks be slower than more
unless there is contention. Lemme lock at HBASE-13071 to see if something sticks out.

Also check out HBASE-13109. Not entirely related, but it finally solves the SKIP vs. SEEK
performance issues that have been plaguing us (SEEK is inefficient unless it can at least
seek into the next block).

With these three we should see some nice perf gains for scanning.


> Coarsen StoreScanner locks to RegionScanner
> -------------------------------------------
>
>                 Key: HBASE-13082
>                 URL: https://issues.apache.org/jira/browse/HBASE-13082
>             Project: HBase
>          Issue Type: Bug
>            Reporter: Lars Hofhansl
>         Attachments: 13082-test.txt, 13082.txt, 13082.txt, gc.png, gc.png, next.png,
next.png
>
>
> Continuing where HBASE-10015 left of.
> We can avoid locking (and memory fencing) inside StoreScanner by deferring to the lock
already held by the RegionScanner.
> In tests this shows quite a scan improvement and reduced CPU (the fences make the cores
wait for memory fetches).
> There are some drawbacks too:
> * All calls to RegionScanner need to be remain synchronized
> * Implementors of coprocessors need to be diligent in following the locking contract.
For example Phoenix does not lock RegionScanner.nextRaw() and required in the documentation
(not picking on Phoenix, this one is my fault as I told them it's OK)
> * possible starving of flushes and compaction with heavy read load. RegionScanner operations
would keep getting the locks and the flushes/compactions would not be able finalize the set
of files.
> I'll have a patch soon.



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

Mime
View raw message