Re: [Brahms-dev-l] Flow Norm Weights

From: Hironori Ito <hito@rcf.rhic.bnl.gov>
Date: Tue Nov 29 2005 - 22:18:12 EST
Hello.  I just check the standard output of one of your un-finished 
job.  It is not related to i/o problem in RCF since it is running at 
full CPU.  Now, it might be related to our disk problem rmine003 and 
rmine004 have been rebooted in a last few days.  Do you have always have 
a problem with particular sets of runs?  If so, it might be file 
corruptions?  Otherwise, I suggest to resubmit. (Or, check the code.)

Hiro


Johnson, Erik B wrote:

>Hiro,
>  Thanks for that info.  That will help a lot in the future.  As for the problems I have been having with these final runs, it seems that processing is just very very very slow?   I'm not sure why, but it might have somehitng to do with these i/o issues we have had in the past?  
>Erik
>
>
>
>-----Original Message-----
>From: brahms-dev-l-bounces@lists.bnl.gov on behalf of Hironori Ito
>Sent: Sun 11/27/2005 3:03 PM
>To: Brahms Dev
>Subject: Re: [Brahms-dev-l] Flow Norm Weights
> 
>Hello.  To find out what is wrong in your program with condor, it is 
>easiest to see a error log and output of that job.  To do this, you need 
>to log on to that machine.  (use condor_q or condor_status)  If you 
>submit jobs to rcas, you can log on to the machine that is actually 
>running your job. (Don't kill (condor_rm) your job since that would 
>remove any temporary file.)  Then, go to 
>/home/condor/local/brahms/execute directory.  You should see a directory 
>like dir_XYZABC.  One of them (if more than one job is running) 
>corresponds to your job.  There, you should be able to see your job 
>output (_condor_stdout_XYZ) and error (_condor_stderr_XYZ) file.
>
>Hiro
>
>Johnson, Erik B wrote:
>
>  
>
>>Hiro,
>> I decided to not wait anymore.  The norm weights for all the runs are now set.  There were 13 runs that I'm still waiting for to finish.  I don't know why these runs are taking so long to finish, but I just set their weighs to the weigts from the previous run. 
>>Erik
>>
>>
>>________________________________
>>
>>From: Hironori Ito [mailto:hito@rcf.rhic.bnl.gov]
>>Sent: Sat 11/26/2005 5:29 PM
>>To: Sanders, Stephen J
>>Cc: Johnson, Erik B; Brahms Dev
>>Subject: Re: [Brahms-dev-l] Flow Norm Weights
>>
>>
>>
>>Did you also change bdst class?  Is that necessary?
>>
>>Hiro
>>
>>Stephen Sanders wrote:
>>
>> 
>>
>>    
>>
>>>Yes, The official brat should be updated.  Otherwise the new AuAu 
>>>cent calibration
>>>will not be incorporated in the lastest dsts.
>>>..steve
>>>On Nov 25, 2005, at 2:05 PM, Johnson, Erik B wrote:
>>>
>>>   
>>>
>>>      
>>>
>>>>Flemming,
>>>>  I haven't changed anything in brat for about a couple weeks  now. 
>>>>So as long as the offical brat has been updated before that  then we
>>>>are good.
>>>>I know Hiro had a question about the flow code at one point and it 
>>>>looked like he didn't have the newest version.  I'm assuming that  he
>>>>resolved this issue.
>>>>Erik
>>>>
>>>>
>>>>-----Original Message-----
>>>>From: Flemming Videbaek [mailto:videbaek@rcf.rhic.bnl.gov]
>>>>Sent: Fri 11/25/2005 12:56 PM
>>>>To: Johnson, Erik B; Brahms Dev
>>>>Subject: Re: [Brahms-dev-l] Flow Norm Weights
>>>>
>>>>Hi Erik,
>>>>
>>>>Should the official brat be updated before the DSTs for the auau  are
>>>>processed?
>>>>Let us have an answer before processing starts.
>>>>
>>>>Flemming
>>>>
>>>>--------------------------------------------
>>>>Flemming Videbaek
>>>>Physics Department
>>>>Bldg 510-D
>>>>Brrokhaven National Laboratory
>>>>Upton, NY11973
>>>>
>>>>phone: 631-344-4106
>>>>fax:        631-344-1334
>>>>e-mail: videbaek @ bnl.gov
>>>>----- Original Message -----
>>>>From: "Johnson, Erik B" <ebj@ku.edu>
>>>>To: "Brahms Dev" <brahms-dev-l@lists.bnl.gov>
>>>>Sent: Friday, November 25, 2005 1:52 PM
>>>>Subject: [Brahms-dev-l] Flow Norm Weights
>>>>
>>>>
>>>>     
>>>>
>>>>        
>>>>
>>>>>Hiro,
>>>>> Here is an update on the normalization weights.  All but 16 runs 
>>>>>have been completed.  I don't know what's been happening with
>>>>>these last sixteen runs, but I have changed the method a little 
>>>>>bit.  Instead of processing 1billion events, I limited it to
>>>>>10million.  The results should not change significantly, but the 
>>>>>weight will be calculated faster.  I will keep a closer eye on
>>>>>them.  As for the rest of the runs, I have checked them and I have 
>>>>>committed them to the database.  I will keep a closer eye on
>>>>>the last 16 runs over the day and weekend.
>>>>> If you want to start processing the DSTs there is a list of runs 
>>>>>that DO NOT have the norm weight committed to the database.
>>>>>9541 - 9546
>>>>>9701 - 9711
>>>>>9836 - 9836
>>>>>9840 - 9840
>>>>>9844 - 9844
>>>>>9845 - 9845
>>>>>9847 - 9849
>>>>>9930 - 9930
>>>>>9945 - 9945
>>>>>10051 - 10054
>>>>>10456 - 10456
>>>>>10458 - 10458
>>>>>10626 - 10626
>>>>>10855 - 10875
>>>>>10948 - 10952
>>>>>11117 - 11129
>>>>>These runs are NOT calibrated!!!
>>>>>
>>>>>  I have rechecked the flow modules and they seem to be working 
>>>>>correctly.
>>>>>
>>>>>  I will send you an update when more runs are finished.
>>>>>Erik
>>>>>_______________________________________________
>>>>>Brahms-dev-l mailing list
>>>>>Brahms-dev-l@lists.bnl.gov
>>>>>http://lists.bnl.gov/mailman/listinfo/brahms-dev-l
>>>>>
>>>>>       
>>>>>
>>>>>          
>>>>>
>>>>_______________________________________________
>>>>Brahms-dev-l mailing list
>>>>Brahms-dev-l@lists.bnl.gov
>>>>http://lists.bnl.gov/mailman/listinfo/brahms-dev-l
>>>>     
>>>>
>>>>        
>>>>
>>>_______________________________________________
>>>Brahms-dev-l mailing list
>>>Brahms-dev-l@lists.bnl.gov
>>>http://lists.bnl.gov/mailman/listinfo/brahms-dev-l
>>>   
>>>
>>>      
>>>
>>
>>
>> 
>>
>>    
>>
>
>_______________________________________________
>Brahms-dev-l mailing list
>Brahms-dev-l@lists.bnl.gov
>http://lists.bnl.gov/mailman/listinfo/brahms-dev-l
>
>  
>

_______________________________________________
Brahms-dev-l mailing list
Brahms-dev-l@lists.bnl.gov
http://lists.bnl.gov/mailman/listinfo/brahms-dev-l
Received on Tue Nov 29 22:19:12 2005

This archive was generated by hypermail 2.1.8 : Tue Nov 29 2005 - 22:19:25 EST