[Clas_offline] user_ana

[Silvia Niccolai]

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