[Bioperl-l] Newbie StandAloneBlast - too many files?

Jason Stajich jason at cgt.mc.duke.edu
Wed Feb 12 19:05:06 EST 2003


ugh - it was supposed to be fixed so you shouldn't have to call this
directly unless you've got an old version of bioperl.  Note there was a
bug in 1.2 with StandAloneBlast so it was missing
Bio::Tools::Run::WrapperBase in the distro so are you absolutely sure you
are running 1.2 on your machine or did you guys get the missing file too?

If you're running this on a webserver the module level DESTROY never gets
called until the server shuts down hence part of the problem, but this is
the same as a single script running blast multiple times.

I do suggest that if folks are writing scripts which run multiple queries
against the same DB rather than using StandAloneBlast you are better off
formatting the input as a single FASTA db and running that as one big
blastall run - you'll get better performance since the database will be
loaded into memory just once (assuming it fits in mem, usual caveats...).

-jason

On Wed, 12 Feb 2003, Keith Boroevich wrote:

> i am using the latest version of biperl (i have the administator upgrade it
> less than a week ago)... and i did notice that StandAloneBlast does have
> it's own destroy function but I never tried calling the cleanup function
> manually.  The files are deleted from the /tmp/ dirctory after each round of
> blasts (2 per loop).  I'll try the _io_cleanup.  currently i just reduced
> the number of sequences in each run to 20 and call the script multiple
> times...
>
> thanks
>
> ----- Original Message -----
> From: "Jason Stajich" <jason at cgt.mc.duke.edu>
> To: "Keith Anthony Boroevich" <kaboroev at sfu.ca>
> Cc: <bioperl-l at bioperl.org>
> Sent: Wednesday, February 12, 2003 5:14 AM
> Subject: Re: [Bioperl-l] Newbie StandAloneBlast - too many files?
>
>
> > This is a bug in old releases that should be fixed in bioperl-1.2.
> >
> > You can alternatively call (where $blastfactory is your StandAloneBlast
> > object):
> >
> > $blastfactory->io->_io_cleanup();
> >
> > before each blastall run to insure files are cleaned up on each round.
> >
> > -jason
> >
> > On Wed, 12 Feb 2003, Keith Anthony Boroevich wrote:
> >
> > > hi,
> > > i am running looped stand alone blast of 88 100pb sequences and around
> > > the 80th or so sequence i recieve the error...
> > >
> > >  >Uncaught exception from user code:
> > >  >Error in tempfile() using /tmp/XXXXXXXXXX: Could not create temp file
> > > /tmp/K6eorI51Yw: Too many open files at
> > > /usr/lib/perl5/site_perl/5.6.1/Bio/Root/IO.pm line 614
> > >  >Carp::croak('Error in tempfile() using /tmp/XXXXXXXXXX: Could not
> > > create temp...') called at /usr/lib/perl5/5.6.1/File/Temp.pm line 1052
> > >  >File::Temp::tempfile('UNLINK', 1, 'DIR', '/tmp') called at
> > > /usr/lib/perl5/site_perl/5.6.1/Bio/Root/IO.pm line 614
> > >  >Bio::Root::IO::tempfile('Bio::Root::IO=HASH(0x8c5989c)') called at
> > > /usr/lib/perl5/site_perl/5.6.1/Bio/Tools/Run/StandAloneBlast.pm line 706
> > >  >
> > >
> >Bio::Tools::Run::StandAloneBlast::_setinput('Bio::Tools::Run::StandAloneBla
> st=HASH(0x8c47024)', 'blastall', 'Bio::Seq=HASH(0x8f91528)') called at
> /usr/lib/perl5/site_perl/5.6.1/Bio/Tools/Run/StandAloneBlast.pm line 483
> > >  >
> > >
> >Bio::Tools::Run::StandAloneBlast::blastall('Bio::Tools::Run::StandAloneBlas
> t=HASH(0x8c47024)', 'Bio::Seq=HASH(0x8f91528)') called at ../bce.pl line 115
> > >  >
> > >  >[2]+  Exit 24                 ../bce.pl orangutan.fas
> > >  >orangutan.blast  (wd: ~/mager/primate)
> > >
> > > this however is only my testing set and i will "hopefully" be running
> > > hundreds to thousands of 100bp sequences through this.  Any suggestions
> > > would be gratly appriciated
> > >
> > > Keith
> > >
> > > _______________________________________________
> > > 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
> >
> >
>
>
> _______________________________________________
> 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


More information about the Bioperl-l mailing list