Re: [XMLPipeDB-developer] gmb2b33-debugging stuck in loop

[John D. N. D. <do...@lm...>]

I looked through all three log files, plus the code that generated the  
log statements, and it looks like this is actually a case where the  
act of logging (since it was much more verbose than before) ended up  
being a bottleneck on the export's progress. Things actually seemed to  
be going OK, albeit extremely slowly because every relationship that  
was generated wrote 4 lines to the log file (disk-related activities  
take 100-1000 times longer to finish, so it's not too farfetched to  
surmise that the generation of 4 lines per relationship would slow  
things down that much).

Thus, I've released a non-debug (i.e., less verbose) version of  
GenMAPP Builder to SourceForge.  Since the logging level is the only  
change, I've kept the version at 2.0b33, but without the "-debug"  
suffix now.

Give this one a shot, and keep me posted...the amount of information  
in gmbuilder.log should now be more in line with 2.0b32 and prior.

John David N. Dionisio, PhD
Assistant Professor, Computer Science
Loyola Marymount University


On Aug 6, 2009, at 11:45 AM, Dahlquist, Kam D. wrote:

> Hi,
>
> I let the gmb2b33-debugging export of P. auregenosa go overnight.   
> It was still stuck at the same place this morning, so I killed it.
>
> I'm attaching the last log file and a screenshot of the console  
> taken right before I killed it (assuming that the list can handle  
> attachments).  Each log is ~5 MB, so I'll have to attach the other  
> ones in separate e-mails.  I'll see how this one is handled first.
>
> Unfortunately, the Gene Ontology tables were still empty as before,  
> although that is not where it seems to have gotten stuck.  It seems  
> to have gotten stuck building the relationship tables.
>
> Kam