[
https://issues.apache.org/jira/browse/HBASE-16698?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15600905#comment-15600905
]
stack commented on HBASE-16698:
-------------------------------
On HBASE-3899, I remembered it as an experiment that was never finished (maybe
someone else has a better memory of what happened then). I went back and looked
at code and seems like it was not hooked up out in Apache HBase. Only
TestDelayedRpc used it ( It may have been working internally at FB). It was
reverted because it was not being used and much of the machinery it had wired
up had been moved or removed (HBaseServer).
Above I say owning the dfsclient is the way ouf our WAL perf trough but the
SEDA core, as you remind us [~carp84], would be another where threads can be
let go after dumping payload on the WAL to go bring on more data while waiting
on the syncs to come home.
> 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)
