[
https://issues.apache.org/jira/browse/HBASE-13819?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15209371#comment-15209371
]
stack commented on HBASE-13819:
-------------------------------
bq. We recently pulled this patch internally and are seeing some significant
side effects of BoundedByteBufferPool where the system runs out of direct
memory whenever there is some network congestion or some client side issues.
Because we are not returning BB to the pool? The pool is growing w/o bound?
bq. Buffer size settles to 512 kb over time(from debug statements we put in)
We should add these if not present at TRACE level.
bq. number of rpcs that were queued in the responder was around ~4000 and this
leads to exhaustion of the direct buffer space,
So 2G instead of 1G? But the pool is bounded?
The responder is not keeping up? It is not moving stuff out of the server fast
enough?
Where is pendingCallsQueue?
Did you observe the offheap size used growing? There s a metric IIRC.
I am interested in leaks; how they are happening.
bq. ...and having a fixed size buffers instead
Where would the fixed size be? In BBBP they eventually reach fixed size?
Thanks.
> Make RPC layer CellBlock buffer a DirectByteBuffer
> --------------------------------------------------
>
> Key: HBASE-13819
> URL: https://issues.apache.org/jira/browse/HBASE-13819
> Project: HBase
> Issue Type: Sub-task
> Components: Scanners
> Reporter: Anoop Sam John
> Assignee: Anoop Sam John
> Fix For: 2.0.0, 1.3.0
>
> Attachments: HBASE-13819.patch, HBASE-13819_branch-1.patch,
> HBASE-13819_branch-1.patch, HBASE-13819_branch-1.patch
>
>
> In RPC layer, when we make a cellBlock to put as RPC payload, we will make an
> on heap byte buffer (via BoundedByteBufferPool). The pool will keep upto
> certain number of buffers. This jira aims at testing possibility for making
> this buffers off heap ones. (DBB) The advantages
> 1. Unsafe based writes to off heap is faster than that to on heap. Now we are
> not using unsafe based writes at all. Even if we add, DBB will be better
> 2. When Cells are backed by off heap (HBASE-11425) off heap to off heap
> writes will be better
> 3. When checked the code in SocketChannel impl, if we pass a HeapByteBuffer
> to the socket channel, it will create a temp DBB and copy data to there and
> only DBBs will be moved to Sockets. If we make DBB 1st hand itself, we can
> avoid this one more level of copying.
> Will do different perf testing with changed and report back.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
