Mailing-List: contact issues-help@hbase.apache.org; run by ezmlm
Precedence: bulk
Date: Wed, 30 Mar 2016 15:26:25 +0000 (UTC)
From: "stack (JIRA)" <jira@apache.org>
To: issues@hbase.apache.org
Message-ID: <JIRA.12953093.1458793945000.89140.1459351585590@Atlassian.JIRA>
In-Reply-To: <JIRA.12953093.1458793945000@Atlassian.JIRA>
References: <JIRA.12953093.1458793945000@Atlassian.JIRA>
 <JIRA.12953093.1458793945361@arcas>
Subject: [jira] [Commented] (HBASE-15525) OutOfMemory could occur when using
 BoundedByteBufferPool during RPC bursts
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit


    [ https://issues.apache.org/jira/browse/HBASE-15525?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15218139#comment-15218139 ] 

stack commented on HBASE-15525:
-------------------------------

bq. This is some thing I discussed with you offline some time back 

Sorry (smile).

bq. The max #BBs in the pool can be defined by customer. (Default to 2 * handlers)..

You know what I am going to say here (smile). So, overflow will be on to heap. Thats good. If we give good feedback on effectiveness of offheap buffers and how much we are overflowing to onheap, that'll help.

Going by diagram over in HBASE-15555, we'll have more than one use of this pool? Or if we more than one use, we should allocate a new pool? Or if fixed size, then maybe should be one pool only?

This is great. How will we test it? (What would you suggest [~dvdreddy])?

> 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: Sub-task
>          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)