[
https://issues.apache.org/jira/browse/HBASE-16698?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15599149#comment-15599149
]
Yu Li commented on HBASE-16698:
-------------------------------
bq. Lets not backport to 1.2 until it in 1.3. Thats how we generally do it.
Else the discontinuity confuses.
Ok, got it, thanks for the confirmation [~stack]
bq. On master, when I do a jstack with some load, almost all the handlers are
waiting for sync()... For async, we still have to have the latch I think.
I see, and makes sense. Let me test to make sure 1) the patch here introduce no
perf regression for SYNC_WAL and 2) it benefits ASYNC_WAL, for master branch.
bq. The other reason I was asking about this is that I have a hacked up patch
which divides the batchMutate() into 3 phases... After sync some other handler
or thread will complete the work.
Thanks for bring this up and mentioning the paper [~enis], I think this
cooperates the idea of "SEDA" JIRA mentioned weeks ago, and we also have some
initial work in progress here in Alibaba-search. I believe this is something
able to increase our overall throughput and worth a standalone JIRA for further
discussion (smile).
Also glanced at HBASE-3899, seems like a similar idea but somehow commit
reverted... mind telling the whole story sir [~stack]?
> Performance issue: handlers stuck waiting for CountDownLatch inside
> WALKey#getWriteEntry under high writing workload
> --------------------------------------------------------------------------------------------------------------------
>
> Key: HBASE-16698
> URL: https://issues.apache.org/jira/browse/HBASE-16698
> Project: HBase
> Issue Type: Improvement
> Components: Performance
> Affects Versions: 1.2.3
> Reporter: Yu Li
> Assignee: Yu Li
> Fix For: 2.0.0
>
> Attachments: HBASE-16698.branch-1.patch,
> HBASE-16698.branch-1.v2.patch, HBASE-16698.branch-1.v2.patch,
> HBASE-16698.patch, HBASE-16698.v2.patch, hadoop0495.et2.jstack
>
>
> As titled, on our production environment we observed 98 out of 128 handlers
> get stuck waiting for the CountDownLatch {{seqNumAssignedLatch}} inside
> {{WALKey#getWriteEntry}} under a high writing workload.
> After digging into the problem, we found that the problem is mainly caused by
> advancing mvcc in the append logic. Below is some detailed analysis:
> Under current branch-1 code logic, all batch puts will call
> {{WALKey#getWriteEntry}} after appending edit to WAL, and
> {{seqNumAssignedLatch}} is only released when the relative append call is
> handled by RingBufferEventHandler (see {{FSWALEntry#stampRegionSequenceId}}).
> Because currently we're using a single event handler for the ringbuffer, the
> append calls are handled one by one (actually lot's of our current logic
> depending on this sequential dealing logic), and this becomes a bottleneck
> under high writing workload.
> The worst part is that by default we only use one WAL per RS, so appends on
> all regions are dealt with in sequential, which causes contention among
> different regions...
> To fix this, we could also take use of the "sequential appends" mechanism,
> that we could grab the WriteEntry before publishing append onto ringbuffer
> and use it as sequence id, only that we need to add a lock to make "grab
> WriteEntry" and "append edit" a transaction. This will still cause contention
> inside a region but could avoid contention between different regions. This
> solution is already verified in our online environment and proved to be
> effective.
> Notice that for master (2.0) branch since we already change the write
> pipeline to sync before writing memstore (HBASE-15158), this issue only
> exists for the ASYNC_WAL writes scenario.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
