[
https://issues.apache.org/jira/browse/HBASE-15525?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15214538#comment-15214538
]
stack commented on HBASE-15525:
-------------------------------
Oh, so, it addresses this issue because sizing of the offheap pool is fixed and
if a write burst, new allocations will be from heap, not from offheap?
How we come up what the fixedsize of a ByteBuffer will be? In this case, the
write s usually 12k in size but there were some sprinkling of 512k writes so
the BBs took the 512k form. Would the FixedSize buffers be 512k or 12k? Or
something else? And how many of these BBs will we keep around?
> OutOfMemory could occur when using BoundedByteBufferPool during RPC bursts
> --------------------------------------------------------------------------
>
> Key: HBASE-15525
> URL: https://issues.apache.org/jira/browse/HBASE-15525
> Project: HBase
> Issue Type: Bug
> Components: IPC/RPC
> Reporter: deepankar
> Assignee: Anoop Sam John
> Priority: Critical
> Attachments: WIP.patch
>
>
> After HBASE-13819 the system some times run out of direct memory whenever
> there is some network congestion or some client side issues.
> This was because of pending RPCs in the RPCServer$Connection.responseQueue
> and since all the responses in this queue hold a buffer for cellblock from
> BoundedByteBufferPool this could takeup a lot of memory if the
> BoundedByteBufferPool's moving average settles down towards a higher value
> See the discussion here
> [HBASE-13819-comment|https://issues.apache.org/jira/browse/HBASE-13819?focusedCommentId=15207822&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-15207822]
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
