[Clas_offline] user_ana

Silvia Niccolai silvia at jlab.org
Wed Mar 26 11:31:18 EDT 2014 

Hi Will,
Good to know that someone still has claslib permissions :-)
If I can give my opinion, I think that the specif fix under discussion (the one that allows user_ana to be run on the batch farm) should be included without further doubts in the official executable, as the one there right now is basically useless, working only interactively. You should just make sure that the only change in the packages since the last build was that one. 
Regards and thanks a lot for taking care of this. 
Silvia


Sent from my iPhone

> On 26/mar/2014, at 16:26, Will Phelps <wphelps at jlab.org> wrote:
> 
> Hello Nikolai,
> I am the CLAS librarian and I am discussing what to do with the builds with Johann and Harut. So far it looks like I will not be changing the /group/clas/builds/ production build until everyone has been notified. I will create a separate build like /group/clas/builds/build_03_26_14/ or something similar for collaborators to test out before changing the production build. Any suggestions are welcome. 
> -Will
> 
>> On Mar 26, 2014, at 11:20 AM, Nikolai Pivnyuk <pivnyuk at jlab.org> wrote:
>> 
>> 
>> 
>>> I don't have access to the claslib account anymore. I'm not sure who does.
>> 
>> 
>> 
>> Does anybody know "who does" ?
>> 
>> 
>> 
>>> If anyone wants to build the fixed version into /group/clas/builds, they
>>> could first log into ifarm1101 as claslib and run this:
>>> 
>>> cd /group/clas/builds/src/clas-trunk
>>> svn up
>>> scons -j32 install
>>> 
>>> 
>>> 
>>> 
>>> -- Johann <https://sites.google.com/site/theodoregoetz/>
>>> 
>>> 
>>> On Tue, Mar 25, 2014 at 4:25 PM, Francois-Xavier Girod
>>> <fxgirod at jlab.org>wrote:
>>> 
>>>> 
>>>> One would think that
>>>> /group/clas/builds/bin/user_ana
>>>> should be consistent with
>>>> /group/clas/builds/src/clas-trunk/reconstruction/recsis/retcl_yield.F
>>>> but this source file is not consistent with the svn repository.
>>>> 
>>>> /group/clas/builds/ is supposed to be updated manually (?).
>>>> 
>>>> Best regards,
>>>> FX
>>>> 
>>>> ----- Original Message -----
>>>> From: "Nikolai Pivnyuk" <pivnyuk at jlab.org>
>>>> To: "Johann Goetz" <theodore.goetz at gmail.com>
>>>> Cc: "clas offline" <clas_offline at jlab.org>
>>>> Sent: Tuesday, March 25, 2014 4:13:27 PM
>>>> Subject: Re: [Clas_offline] user_ana
>>>> 
>>>> Johann,
>>>> 
>>>>  ???????
>>>> 
>>>> pivnyuk at ifarm1102:~>ll /group/clas/builds/bin/user_ana
>>>> -rwxrwxr-x 1 claslib clasdev 4225915 Dec 18 14:33
>>>> /group/clas/builds/bin/user_ana
>>>> 
>>>> 
>>>>> This modification has been in the SVN repository since January, 2014.
>>>>> https://jlabsvn.jlab.org/svnroot/clas/trunk
>>>>> 
>>>>> -- Johann <https://sites.google.com/site/theodoregoetz/>
>>>>> 
>>>>> 
>>>>>> On Tue, Mar 25, 2014 at 2:32 PM, Silvia Niccolai <silvia at jlab.org>
>>>>> wrote:
>>>>> 
>>>>>> Hello all,
>>>>>> Could this "trick" be implemented in the official clas-trunk, so that
>>>>>> people don't have to check out, modify this routine and recompile
>>>>>> everything?
>>>>>> This issue has been known since about last summer...
>>>>>> Thanks
>>>>>> Silvia
>>>>>> 
>>>>>> Sent from my iPhone
>>>>>> 
>>>>>>> On 25/mar/2014, at 19:24, KiJun Park <parkkj at jlab.org> wrote:
>>>>>>> 
>>>>>>> Hi,
>>>>>>> 
>>>>>>> Perhaps your problem is related with source code not your input bos
>>>>>> file.
>>>>>>> Many people already noticed this problem without knowing the
>>>> reason.
>>>>>>> (You may see lots of communications in CLAS_OFFLINE account)
>>>>>>> The ifarm1102 and batch-farm are identical machines
>>>> (CENTOS6.2/64bit),
>>>>>> calling same LIBs,... etc.
>>>>>>> No one really does not know why the error is occurred in only batch
>>>>>> mode, meantime an interactive mode is working fine.
>>>>>>> 
>>>>>>> So far, there is a trick which is so far working.
>>>>>>> You may try this too.
>>>>>>> But this is not an ultimate solution.
>>>>>>> 
>>>>>>> 
>>>>>>> 
>>>>>>> The source code in "retcl_yield.F" has
>>>>>>> ---------------------------------------------------------
>>>>>>> c
>>>>>>> c look at the terminal input
>>>>>>> c
>>>>>>> IF (LINTER) THEN
>>>>>>> call asynck_c(nbyt)
>>>>>>> else
>>>>>>> nbyt = 1
>>>>>>> endif
>>>>>>> ---------------------------------------------------------
>>>>>>> 
>>>>>>> 
>>>>>>>>>>> change this "nbyt = 1" to  "nbyt = 0"
>>>>>>> 
>>>>>>> I know, this does not make any sense.
>>>>>>> Because in principle, the loop is going to be infinite when  "nbyt
>>>> =
>>>>>> 0"
>>>>>> is engaged.
>>>>>>> 
>>>>>>>       IF (LINTER) THEN
>>>>>>>           call asynck_c(nbyt)
>>>>>>>       else
>>>>>>>          nbyt = 1
>>>>>>>       endif
>>>>>>> 
>>>>>>> re-compile it and submit your job.
>>>>>>> good luck.
>>>>>>> 
>>>>>>> Sincerely yours,
>>>>>>> Kijun
>>>>>>> 
>>>>>>> 
>>>>>>> 
>>>>>>> 
>>>>>>> ----- Original Message -----
>>>>>>> From: "Nikolai Pivnyuk" <pivnyuk at jlab.org>
>>>>>>> To: "clas offline" <clas_offline at jlab.org>
>>>>>>> Sent: Tuesday, March 25, 2014 2:14:50 PM
>>>>>>> Subject: [Clas_offline] user_ana
>>>>>>> 
>>>>>>> Dear CLAS offliners,
>>>>>>> 
>>>>>>> I have a problem with running user_ana on batch farm nodes.
>>>>>>> To launch the job I currently use the same scripts which I have
>>>> used a
>>>>>> few
>>>>>>> years ago without any problem. These scripts are
>>>>>>> /volatile/clas/clasg10/pivnyuk/g10/mc/batch_pipp/send_to_farm.pl
>>>>>>> and
>>>>>>> /volatile/clas/clasg10/pivnyuk/g10/mc/batch_pipp/runmc.pl
>>>>>>> In runmc.pl see just only quite small "recsis" block of the script.
>>>>>>> ---------------------------------------------------------------------
>>>>>>> The problem looks likely as the input file
>>>> /volatile/clas/clasg10/pivnyuk/g10/mc/outdir_pimi_p_resc_dc_0.02_de_0.08/eg_2.90/gpp0004.bos
>>>>>>> is invisible for the user_ana.
>>>>>>> It is just only my impression. I'm not absolutely sure about that.
>>>>>>> ----------------------------------------------------------------------
>>>>>>> Processing the same gpp0004.bos with user_ana but in INTERACTIVE
>>>> MODE
>>>>>> is
>>>>>>> absolutely OK. No problem.
>>>> -----------------------------------------------------------------------
>>>>>>> More details (.tcl and output .log files etc.) are in the
>>>> "user_ana"
>>>>>>> attached file.
>>>> ------------------------------------------------------------------------
>>>>>>> Could somebody kindly have a look on my stuff ?
>>>>>>> Thanks in advance,
>>>>>>> Nikolai
>>>>>>> _______________________________________________
>>>>>>> Clas_offline mailing list
>>>>>>> Clas_offline at jlab.org
>>>>>>> https://mailman.jlab.org/mailman/listinfo/clas_offline
>>>>>>> 
>>>>>>> --
>>>>>>> ---------------------------------------******
>>>>>>> Dr. Kijun Park
>>>>>>> Physics Dept. Old Dominion University
>>>>>>> Thomas Jefferson National Accelerator Facility
>>>>>>> Suit 5 Room 12_1/A102
>>>>>>> 12000 Jefferson Ave.
>>>>>>> Newport News VA 23606
>>>>>>> Phone 757-269-6989
>>>>>>> 
>>>>>>> _______________________________________________
>>>>>>> Clas_offline mailing list
>>>>>>> Clas_offline at jlab.org
>>>>>>> https://mailman.jlab.org/mailman/listinfo/clas_offline
>>>>>> 
>>>>>> _______________________________________________
>>>>>> Clas_offline mailing list
>>>>>> Clas_offline at jlab.org
>>>>>> https://mailman.jlab.org/mailman/listinfo/clas_offline
>>>>> _______________________________________________
>>>>> Clas_offline mailing list
>>>>> Clas_offline at jlab.org
>>>>> https://mailman.jlab.org/mailman/listinfo/clas_offline
>>>> 
>>>> 
>>>> _______________________________________________
>>>> Clas_offline mailing list
>>>> Clas_offline at jlab.org
>>>> https://mailman.jlab.org/mailman/listinfo/clas_offline
>>> _______________________________________________
>>> Clas_offline mailing list
>>> Clas_offline at jlab.org
>>> https://mailman.jlab.org/mailman/listinfo/clas_offline
>> 
>> 
>> _______________________________________________
>> Clas_offline mailing list
>> Clas_offline at jlab.org
>> https://mailman.jlab.org/mailman/listinfo/clas_offline
> 
> _______________________________________________
> Clas_offline mailing list
> Clas_offline at jlab.org
> https://mailman.jlab.org/mailman/listinfo/clas_offline
-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://mailman.jlab.org/pipermail/clas_offline/attachments/20140326/05e3d857/attachment-0001.html

More information about the Clas_offline mailing list