[jira] Commented: (HADOOP-3856) Asynchronous IO Handling in Hadoop and HDFS

["Jonathan Gray (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/HADOOP-3856?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12663548#action_12663548 ] 

Jonathan Gray commented on HADOOP-3856:
---------------------------------------

Over in HBase land, we're hitting issues with this.  We've been recommending users to up their max xceivers, but now we're starting to see OOMEs of varying sorts (like unable to create new native thread) as each xceiver is taking up at least one thread.

To be efficient with random access, we must keep connections/handles open to our files.  However this does not scale with the current DataNode (and DFSClient) implementations.

It seems Netty would be preferable, LGPL license should be okay.

There are some HBase guys who are willing to spend time working on this, but it seems there are some splits about how to proceed.  The idea of larger changes seems to have silenced the issue, so maybe we can try an iterative approach and solve the primary issue first of not having a thread per connection in the DataNode.  And then look at optimizing DFSClient.

Looking for some feedback so we can get started.  Thanks.


> Asynchronous IO Handling in Hadoop and HDFS
> -------------------------------------------
>
>                 Key: HADOOP-3856
>                 URL: https://issues.apache.org/jira/browse/HADOOP-3856
>             Project: Hadoop Core
>          Issue Type: New Feature
>          Components: dfs, io
>            Reporter: Raghu Angadi
>         Attachments: GrizzlyEchoServer.patch, MinaEchoServer.patch
>
>
> I think Hadoop needs utilities or framework to make it simpler to deal with generic asynchronous IO in  Hadoop.
> Example use case :
> Its been a long standing problem that DataNode takes too many threads for data transfers. Each write operation takes up 2 threads at each of the datanodes and each read operation takes one irrespective of how much activity is on the sockets. The kinds of load that HDFS serves has been expanding quite fast and HDFS should handle these varied loads better. If there is a framework for non-blocking IO, read and write pipeline state machines could be implemented with async events on a fixed number of threads. 
> A generic utility is better since it could be used in other places like DFSClient. DFSClient currently creates 2 extra threads for each file it has open for writing.
> Initially I started writing a primitive "selector", then tried to see if such facility already exists. [Apache MINA|http://mina.apache.org] seemed to do exactly this. My impression after looking the the interface and examples is that it does not give kind control we might prefer or need.  First use case I was thinking of implementing using MINA was to replace "response handlers" in DataNode. The response handlers are simpler since they don't involve disk I/O. I [asked on MINA user list|http://www.nabble.com/Async-events-with-existing-NIO-sockets.-td18640767.html], but looks like it can not be done, I think mainly because the sockets are already created.
> Essentially what I have in mind is similar to MINA, except that read and write of the sockets is done by the event handlers. The lowest layer essentially invokes selectors, invokes event handlers on single or on multiple threads. Each event handler is is expected to do some non-blocking work. We would of course have utility handler implementations that do  read, write, accept etc, that are useful for simple processing.
> Sam Pullara mentioned that [xSockets|http://xsocket.sourceforge.net/] is more flexible. It is under GPL.
> Are there other such implementations we should look at?

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.