[
https://issues.apache.org/jira/browse/HBASE-16698?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15574288#comment-15574288
]
Allan Yang commented on HBASE-16698:
------------------------------------
Why handlers stuck at CountDownLatch? In my opinion, if the write operation in
{{doMiniBatchMutation}} need to sync, after waiting for syncing, the seqid is
assigned in RingBufferEventHandler as you pointed out. So when
{{mvcc.completeMemstoreInsertWithSeqNum}} is called, the latch is released
already.
Unless you set the batch's durability to ASYNC or NOSYNC, then there will be a
problem. Though you don't want to wait for the sync, in order to get the right
mvcc readpoint, you have to wait for the previous transaction to finish the
sync. So that will cause the handlers stuck at CountDownLatch.
I think people set the transaction's durability to ASYNC or NOSYNC is to gain
speed. But in branch 1 after merging the mvcc id with the seqid, sync no wal
still have to wait other transaction to sync, that's a problem need to be fixed.
> 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: 1.1.6, 1.2.3
> Reporter: Yu Li
> Assignee: Yu Li
> Fix For: 2.0.0
>
> Attachments: HBASE-16698.branch-1.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)
