[
https://issues.apache.org/jira/browse/HBASE-13082?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14964476#comment-14964476
]
Anoop Sam John commented on HBASE-13082:
----------------------------------------
Not getting cells from bulk loaded files which is loaded in btw a scan seems ok
and seems that is the correct way.
Regarding flush
bq.On flush, what if we let all Scans complete before letting go the snapshot?
More memory pressure. Simpler implementation
Yes simple but heap pressure will be much more. And when the scan will get over
is not really in our hands. I dont think we can think of doing this :-)
The other option you said seems fine. We need to see how this will look.. Will
be much more complex..
So how much is the perf gain we get by fully avoiding this lock? Worth of
doing all these complex stuff. I think it is still worth as the gain may be
high.
And this will allow to make full use of the cached blocks of compacted files is
another adv.
Again nice doc Ram.. Just in 2 mins we can understand the whole change and why
to do it.. Getting the impl details reading through the patch is much harder.
Thanks man.
> 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
> Assignee: ramkrishna.s.vasudevan
> Attachments: 13082-test.txt, 13082-v2.txt, 13082-v3.txt,
> 13082-v4.txt, 13082.txt, 13082.txt, HBASE-13082.pdf, HBASE-13082_1_WIP.patch,
> HBASE-13082_2_WIP.patch, HBASE-13082_3.patch, HBASE-13082_4.patch, gc.png,
> gc.png, gc.png, hits.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)
