[Bioperl-l] Re: "Too many files open" with bl2seq

[Jason Stajich]

Hmm - perhaps it is more of a problem with the BPlite object holding the
file handle on the report - so even though we told it to close from the
perspective of the Factory, perl is not cleaning up the filehandle because
the BPbl2seq object is not closing its filehandles until perl exits.
weird.  Something some can poke around with perhaps - this should be
managed by the Root::IO object.

Someday someone will get around to integrating bl2seq parsing into
SearchIO as well.  Matt Wiepert had some changes which worked I think
pre-Chervitz commits, I expect it is really not that hard though.

-jason
On Tue, 22 Apr 2003, Jamie Hatfield (AGCoL) wrote:

> I was having problems with the "too many files open" error using the
> bl2seq module also.  I think I found a suitable patch.  If you take the
> $report object returned from the $factory->bl2seq command and call
> $report->close() once you're done with it, you shouldn't get that error
> anymore.  By the way, oddly enough, this is only necessary when
> $report->sbjctName is defined.  I assume this is because this is the
> only time that something was actually returned.
>
> Of course, maybe I just don't have the latest files.  Either way, this
> should fix the problem for anyone who still has it... Maybe?
>
> ----------------------------------------------------------------------
> Jamie Hatfield                              Room 541H, Marley Building
> Systems Programmer                          University of Arizona
> Arizona Genomics Computational              Tucson, AZ  85721
>   Laboratory (AGCoL)                        (520) 626-9598
>
> > -----Original Message-----
> > From: bioperl-l-bounces at bioperl.org
> > [mailto:bioperl-l-bounces at bioperl.org] On Behalf Of Steven Cannon
> > Sent: Thursday, April 17, 2003 8:56 AM
> > To: Peter Schattner
> > Cc: bioperl-l at bioperl.org
> > Subject: [Bioperl-l] Re: "Too many files open" with bl2seq
> >
> >
> > Peter:
> >
> > Thanks. Actually, eval didn't help, at least as I applied it -- the
> > program keeps chugging (and ignores the errors), but the blasts still
> > don't give output after the ~250th hit. So I still wonder if
> > the bl2seq
> > problem might be the origin of the bioperl bl2seq errors.
> >
> > Steve
> >
> > On Thursday, April 17, 2003, at 10:12 AM, Peter Schattner wrote:
> >
> > > Steven Cannon wrote:
> > >
> > >> To second your '"Too many files open" with bl2seq'
> > question, I noticed
> > >> something similar last month when trying to use bl2seq via
> > bioperl. So,
> > >> apologies for re-posting, but maybe the extra details will
> > help someone
> > >> spot the glitch. I see a different set of errors (MSG:  No
> > hit object
> > >> found for bl2seq report  ... etc.) , but perhaps the
> > underlying cause
> > >> is
> > >> the same.
> > >>
> > >
> > > Steve
> > >
> > > Actually, these are two completely separate issues.  The
> > "MSG:  No hit
> > > object
> > > found for bl2seq report" crash is a bug in BPbl2seq.pm (I think).
> > > Could you
> > > submit this as a bug report?  I've seen this problem too and as a
> > > workaround
> > > you can just surround the StandAloneBlast call to bl2seq in
> > an "eval",
> > > ignore
> > > the error and (at least for me) everything works fine.
> > >
> > > The "Too many files open"  crash is something quite
> > different.  It has
> > > nothing to do with Bioperl since it happens when running
> > bl2seq from the
> > > command line as well.
> > >
> > > Peter
> > >
> > >
> >
> > _______________________________________________
> > Bioperl-l mailing list
> > Bioperl-l at bioperl.org
> > http://bioperl.org/mailman/listinfo/bioperl-l
> >
>
> _______________________________________________
> Bioperl-l mailing list
> Bioperl-l at bioperl.org
> http://bioperl.org/mailman/listinfo/bioperl-l
>

--
Jason Stajich
Duke University
jason at cgt.mc.duke.edu