LISTSERV mailing list manager LISTSERV 16.5

Help for XROOTD-L Archives


XROOTD-L Archives

XROOTD-L Archives


XROOTD-L@LISTSERV.SLAC.STANFORD.EDU


View:

Message:

[

First

|

Previous

|

Next

|

Last

]

By Topic:

[

First

|

Previous

|

Next

|

Last

]

By Author:

[

First

|

Previous

|

Next

|

Last

]

Font:

Proportional Font

LISTSERV Archives

LISTSERV Archives

XROOTD-L Home

XROOTD-L Home

XROOTD-L  May 2011

XROOTD-L May 2011

Subject:

Re: Resending: debugging poor performance of xroot across WAN

From:

Lukasz Janyst <[log in to unmask]>

Date:

17 May 2011 15:16:40 +0200Tue, 17 May 2011 15:16:40 +0200

Content-Type:

text/plain

Parts/Attachments:

Parts/Attachments

text/plain (138 lines)

2011/5/16 Charles G Waldman <[log in to unmask]>:
>  >  When the traffic optimization is supposed to happen in XRootD, you
>  > would see things like a XRootD::Read of len=489, because any external
>  > application (ROOT in this case) just needs to state what data it needs
>  > and the task of figuring out how to fetch it in the most optimal way
>  > is entrusted to XRootD which will optimize the actual network reads.
>  > The tweaking can then be done by the parameters you mention. Since
>  > XRootD is a generic data access software it has no knowledge of the
>  > underlying data file format so all it can do is some statistical guess
>  > work that may be more or less optimal.
>
> Yes, but for sites which are running a particular type of job - for
> example, a USATLAS site - the jobs have similar data-access patterns
> and it makes sense to have one set of read-ahead params for all the
> jobs, controlled by an environment variable or setup file.  For
> example at the MWT2 site, we were using dcap access, and were having
> good results with WAN reads once we set a few environment variables:
>
>  export DC_LOCAL_CACHE_MEMORY_PER_FILE=10000000
>  export DC_LOCAL_CACHE_BLOCK_SIZE=32768
>  export DCACHE_RA_BUFFER=16000
>
> which were applied to all ATLAS jobs.  We have not been able to get
> the same level of performance with xrootd across the WAN as we had
> gotten with dcache with the above env. vars.  There should be controls
> which can be adjusted by (knowledgeable) site admins, who can tune
> this to get optimal performance for their given operating parameters
> (network latency, disk I/O performance, memory limitation, etc).

   Yes, this is quite correct. You can do similar adjustment with
XRootD. Unfortunately in the older versions (coming with ROOT<5.28)
you can do that only by putting a .rootrc file in the CWD of the job
that you're running or in $ROOTSYS/etc, in versions coming with ROOT
>5.28 you can set it via envvars.

   You can have three values for ReadAheadStrategy:

0) no read ahead
1) sequential - read some data ahead of currently requested buffer
(how much should be read is specified by ReadAheadSize)
2) sliding window - window centered on the recent average slides
through the file following the stream of the requests

   ReadCacheSize denotes how much data should be cached per file.

>  > 2) When the traffic optimization is supposed to happen in ROOT, you
>  > would most probably see xrootd being asked to perform vector reads
>  > (XrdClient::ReadV) of around 30 megs. This is far more optimal because
>  > ROOT knows its file format and can easily predict which parts of the
>  > file it will need in the nearest future so it is able to prefetch data
>  > before it is needed by the application. The parameters here are the
>  > ones that you mention below: tree->SetCacheSize and friends.
>
> OK, but as I understand it, there is no way to turn this on at the
> site level - this can't be enabled in system.rootrc or via an env.
> var, can it?  It seems this requires users to rewrite their code.

   Correct.

>  > > Question 1)
>  > > Why is ReadAheadSize set to 0 here?  And what's the best way to override this?
>  >
>  >    Because most probably it is assumed that the optimization will be
>  > done in ROOT and XRootD should not bother to do anything.
>  >
>  > >     We'd love to turn it on by default. But the read-ahead needs to put the data somewhere: we need some extra memory; 30 to 100MB are enough. But with the experiments' never ending quest for memory resources turning this on by default is not an option: a 4-core machine with 4 jobs would eat an additional 400MB. Too much.
>  >
>  >    Well, if you turn on the read-ahead or any prefetching you need to
>  > store the additional data somewhere and RAM is the easiest target. We
>  > work on some code that will hopefully be committed to ROOT soon which
>  > will enable it to prefetch the data blocks and store them on disk.
>
> This could affect performance negatively - for a multi-core host, you
> are now creating more I/O to the local disk, which can become a
> bottleneck.  This is exactly the reason we prefer remote-access to
> stage-in for job inputs - having 24 or more jobs all access the same
> local disk can create I/O contention.  For worker nodes which are not
> short on RAM, the additional memory usage of a (modest) cache is not
> a problem.  (Using 10MB/file for dCache gave a huge performance boost
> and did not cause undue memory pressure.  We did not experiment with
> lower values but I suspect even 1MB/file would be enough to help
> significantly).

   Depends on the usecase, it also introduces parallel prefetching of
TreeCache buffers so the first results are quite promising. In any
case it will be possible to disable/tweak it depending on the
particular needs.

>
>  > >     Instead you need to turn it on by yourself by calling
>  > >
>  > >       tree->SetCacheSize(10000000);
>  > >       tree->AddBranchToCache("*");
>  > >
>  > >
>  > > I don't think we can force users to do this, is there somewhere else in the stack
>  > > that this code could be inserted?
>  >
>  >    Not really, since it's dependent on the user data that is being
>  > read. You could argue that it should use some caching by default but
>  > it's debatable.
>
> Well, we are comparing performance of xrootd and dcache.  For jobs
> accessing data across LAN, the performance is comparable.  But once
> the WAN enters the picture, the xrootd performace is poor compared to
> what we get with dCache.  So, I'm trying to at least reproduce the
> dCache performance results ... (using ATLAS "Hammercloud" job
> performance as my metric).

   Do similar tweaking that you have done for dCache :)

>
>  > > Question 2)
>  > > Is this "session not found" the cause of the failures?
>  >
>  >    This is a request to end a session that apparently does not
>  > exist... I will have a closer look. What is your access pattern? Do
>  > you have long standing jobs that keep the connections open for a long
>  > time, or is it more like the jobs fetching the data they need to
>  > process and quiting after the processing is over?
>
> This is a "canned" ATLAS job from the Hammercloud testing system - using
>
>  Test template: 73 (stress) - Muon 16.0.3.3 PANDA default data-access
>
> I'll do some research to find out more about what this job is actually
> doing.
>

   As already commented by Andy: this is a bug that will be fixed.
Thanks for reporting!

Cheers,
   Lukasz



Top of Message | Previous Page | Permalink

Advanced Options


Options

Log In

Log In

Get Password

Get Password


Search Archives

Search Archives


Subscribe or Unsubscribe

Subscribe or Unsubscribe


Archives

March 2024
February 2024
January 2024
December 2023
November 2023
October 2023
September 2023
August 2023
July 2023
June 2023
May 2023
April 2023
March 2023
February 2023
January 2023
December 2022
October 2022
September 2022
August 2022
July 2022
June 2022
May 2022
April 2022
March 2022
February 2022
January 2022
December 2021
November 2021
October 2021
September 2021
August 2021
July 2021
June 2021
May 2021
April 2021
March 2021
February 2021
January 2021
December 2020
November 2020
October 2020
September 2020
August 2020
July 2020
June 2020
May 2020
April 2020
March 2020
February 2020
January 2020
December 2019
November 2019
October 2019
September 2019
August 2019
July 2019
June 2019
May 2019
April 2019
March 2019
February 2019
January 2019
December 2018
November 2018
October 2018
September 2018
August 2018
July 2018
June 2018
May 2018
April 2018
March 2018
February 2018
January 2018
December 2017
November 2017
October 2017
September 2017
August 2017
July 2017
June 2017
May 2017
April 2017
March 2017
February 2017
January 2017
December 2016
November 2016
October 2016
September 2016
August 2016
July 2016
May 2016
April 2016
March 2016
February 2016
January 2016
December 2015
November 2015
October 2015
September 2015
August 2015
July 2015
June 2015
May 2015
April 2015
March 2015
February 2015
January 2015
December 2014
November 2014
October 2014
September 2014
August 2014
July 2014
June 2014
May 2014
April 2014
March 2014
February 2014
January 2014
December 2013
November 2013
October 2013
September 2013
August 2013
July 2013
June 2013
May 2013
April 2013
March 2013
February 2013
January 2013
December 2012
November 2012
October 2012
September 2012
August 2012
July 2012
June 2012
May 2012
April 2012
March 2012
February 2012
January 2012
December 2011
November 2011
October 2011
September 2011
August 2011
July 2011
June 2011
May 2011
April 2011
March 2011
February 2011
January 2011
December 2010
October 2010
September 2010
August 2010
July 2010
June 2010
May 2010
April 2010
March 2010
February 2010
January 2010
December 2009
November 2009
October 2009
September 2009
July 2009
June 2009
May 2009
April 2009
March 2009
January 2009
December 2008
September 2008
August 2008
July 2008
June 2008
May 2008
April 2008
March 2008
February 2008
January 2008
December 2007
November 2007
October 2007
September 2007
August 2007
July 2007
June 2007
May 2007
April 2007
March 2007
February 2007
January 2007
December 2006
November 2006
October 2006
September 2006
August 2006
July 2006
June 2006
May 2006
April 2006
March 2006
February 2006
January 2006
December 2005
November 2005
October 2005
September 2005
August 2005
July 2005
June 2005
May 2005
April 2005
March 2005
February 2005
January 2005
December 2004
November 2004
October 2004
September 2004
August 2004

ATOM RSS1 RSS2



LISTSERV.SLAC.STANFORD.EDU

Secured by F-Secure Anti-Virus CataList Email List Search Powered by the LISTSERV Email List Manager

Privacy Notice, Security Notice and Terms of Use