Re: [Brahms-dev-l] proof problem

Hello.  Not exactly the same problem.  But, I noticed one clear bug(?) 
in the proof.  If you have empty histograms added to the output 
(fOutput) to be merged, the process takes very long time to merge.  
Particularly, this is particularly true if you have alot of empty 
histograms to be merged.  It looks like that the process just sit and do 
nothing after processing most (if not all) events.  The solution is not 
to add empty histograms (if you know they are empty).  Anyway, this is 
probably not your particular problem.  You may check the followings;  
your input file is corrupt.  Your output directory is full.  Your output 
file is larger than 2GB. Your proof (or proofserve at one of your slave 
machine) job is not killed/reniced.  Try running at different  
proofserver. etc..

Hiro


Johnson, Erik B wrote:

>Brahms,
>   This is a general question to everyone out there.  
>I have started a proof session and it processed most of the data except for the last 10000 events. Proof reports that the processing will finish in 0.5seconds, but it just hangs.  I run the same code in a root session over a liited number of events and it finishes fine with out error, so there isn't a bug in the code.  I have also ran over all of the data with a different piece of code and it finished, so it isn't the data.  Has anyone seen this before?  Does anyone have a solution or idea to what is going on here?
>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