[Biopython-dev] Local blast problem

[Jeffrey Chang]

Hmmm...  I'm not sure.  Is Python core-dumping, or just blast?  I've 
run into BLAST crashing before, but it usually results in a truncated 
stream and hasn't caused python to core-dump.

Some possible workarounds are to:
1) fork off a separate thread to run blast, so if it crashes, it 
won't take down your main application.  The MultiProc library might 
help here.
2) hack the blastall (or blatspgp) function so that it saves the 
output to a file, and then return a handle to that file.  This is 
technically a variation of the first solution, and might be more 
straightforward to implement.

Jeff


At 2:46 PM +1000 9/11/01, Sarah Kummerfeld wrote:
>Hi,
>
>Just wondering whether anyone know about an intermittent
>problem with locally run blast.
>
>I'm running lots of blast searches on a very small
>database. It will work for a while, or sometimes for
>the whole program. But other times it will core dump.
>
>Occasionally I get a python traceback which suggests
>that it tried to read a stream that was not there,
>but other times there is no error, it just dumps
>core.
>
>If I run blast on its own (not from my python program)
>it sometimes does the same thing. One time I rebooted
>my machine (linux) and found that a blast search I had
>run just before and that had crashed, now worked.
>
>I couldn't find anything helpful in the core file.
>
>I had thought it might be some new memory I put in,
>so I had it replaced -- but still have the problem.
>
>Any suggestions would be greatly appreciated!
>
>Sarah
>
>
>
>_______________________________________________
>Biopython-dev mailing list
>Biopython-dev at biopython.org
>http://biopython.org/mailman/listinfo/biopython-dev