Re: Proof Asynchronous Mode

From: Sinan Kuday <kuday_at_cern.ch>
Date: Tue, 23 Aug 2011 17:52:23 +0300


Thanks a lot.

Creating chains with "http://" solved my issue but naturally slowed the analysis down.
I have installed a small Proof cluster getting together several computers using PoD and i was planning to keep datasets in the master node.
Now i also installed afdsmgrd for dataset management but can't find enough documentation for stand-alone running. On the other i really don't understand why workers don't accept usual paths.
Any suggestions?

Best,
Sinan Kuday

On 23.Ašu.2011, at 11:54, Gerardo Ganis wrote:

>
> Hello,
>
> Yes, the workers must be able to open the files, either from a
> shared file system, or from a server
> (http://, root://, ...) or from a local file system.
>
> in the specific case the files are available at
>
> http://root.cern.ch/files/h1
>
> so you can create the chain with the macro under tutorials/tree
>
> root [] .x $ROOTSYS/tutorials/tree/h1chain.C("http://root.cern.ch/files/h1
> ")
>
> and your problem should go away.
>
> In the general case, files are read out from ad hoc servers,
> typically based on xrootd, some tool is used
> to populate the local disks on the machines. For example, the
> dataset manager daemon under
> $ROOTSYS/proof/afdsmgrd can be used in conjunction with xrootd to
> do that (see also
> http://code.google.com/p/afdsmgrd/).
>
> G. Ganis
>
> On 8/22/11 8:20 PM, Sinan Kuday wrote:
>> Hello,
>>
>> I am using Root 5.30.
>> By the way when i looked at log files, i saw the following errors:
>>
>>
>> So i guess workers can not find the root files dstarmb.root,
>> dstarp1a.root, dstarp1b.root ...etc. and that must be the problem.
>> But where exactly should i put these files? Should i upload them in
>> a par file?
>>
>> Thanks,
>> Sinan Kuday
>>
>>
>> On 22.Ašu.2011, at 20:53, Gerardo Ganis wrote:
>>
>>>
>>> Hello,
>>>
>>> I am not able to reproduce your problem. For me the example works
>>> fine, also with your
>>> configuration settings.
>>> Which version are you using?
>>>
>>> G. Ganis
>>>
>>> Ps: for what relates the dynamic startup, there was a crucial bug
>>> affecting it which is now fixed
>>> i the trunk, 5-30-patches and 5-28-00-patches. Thanks for
>>> reporting.
>>>
>>>
>>>
>>> On 8/18/11 9:07 PM, Sinan Kuday wrote:
>>>> Hello,
>>>>
>>>> I'd like to run following example on a local cluster or laptop:
>>>>
>>>> http://root.cern.ch/drupal/content/running-queries-asynchronous-batch-mode
>>>>
>>>> My configuration file:
>>>>
>>>> xrd.protocol xproofd libXrdProofd.so
>>>> xpd.schedparam selopt:roundrobin fraction:0.7
>>>> xpd.schedparam queue:fifo mxrun:5
>>>> #xpd.putrc Proof.DynamicStartup 1
>>>>
>>>> I have commented out "putrc" part since it doesn't work for me.
>>>> Here is what i am doing after running xrootd.
>>>>
>>>>
>>>> root [0] TProof *p = TProof::Open("localhost");
>>>> Starting master: opening connection ...
>>>> Starting master: OK
>>>> Opening connections to workers: OK (2 workers)
>>>> Setting up worker servers: OK (2 workers)
>>>> PROOF set to parallel mode (2 workers)
>>>> root [1] .x h1chain.C
>>>> root [2] chain.SetProof();
>>>> root [3] chain.Process("h1analysis.C","ASYN");
>>>> root [4] chain.Process("h1analysis.C","ASYN");
>>>> root [5]
>>>> | session: kuday.default.17720.status terminated by peer
>>>> Info in <TXSlave::HandleError>: 0x22d9970:localhost:0 got
>>>> called ... fProof: 0x1841a00, fSocket: 0x1860e00 (valid: 1)
>>>> Info in <TXSlave::HandleError>: 0x22d9970: proof: 0x1841a00
>>>> TXSlave::HandleError: 0x22d9970: DONE ...
>>>> Info in <TProof::MarkBad>:
>>>> +++ Message from local session : marking localhost:1093 (0) as bad
>>>> +++ Reason: received kPROOF_FATAL
>>>>
>>>> +++ Message from local session : marking localhost:1093 (0) as bad
>>>> +++ Reason: received kPROOF_FATAL
>>>>
>>>> +++ Most likely your code crashed
>>>> +++ Please check the session logs for error messages either using
>>>> +++ the 'Show logs' button or executing
>>>> +++
>>>> +++ root [] TProof::Mgr("kuday_at_localhost:1093")->GetSessionLogs()-
>>>> >Display("*")
>>>>
>>>>
>>>> where am i doing wrong??
>>>> Thanks in advance,
>>>> Sinan Kuday
>>>>
>>>
>>>
>>> --
>>> +
>>> --------------------------------------------------------------------------+
>>> Gerardo GANIS CERN, PH Dept, SFT group, CH 1211 Geneve 23
>>> room: 32-RC-017, tel: +41 22 7676439
>>> email: gerardo.ganis_at_cern.ch, fax: +41 22 7669133
>>> +
>>> --------------------------------------------------------------------------+
>>>
>>
>
>
> --
> +
> --------------------------------------------------------------------------+
> Gerardo GANIS CERN, PH Dept, SFT group, CH 1211 Geneve 23
> room: 32-RC-017, tel: +41 22 7676439
> email: gerardo.ganis_at_cern.ch, fax: +41 22 7669133
> +
> --------------------------------------------------------------------------+
>
Received on Tue Aug 23 2011 - 17:48:51 CEST

This archive was generated by hypermail 2.2.0 : Fri Aug 26 2011 - 23:50:02 CEST