[Bioperl-l] Too many links at LoadHelper.pm

Dan Kortschak dan.kortschak at adelaide.edu.au
Sat Feb 6 05:11:23 UTC 2010


When I was working through those issues (I was having problems with
Bio::DB::Sam causing tempfiles to be released before they were finished
with) I paid quite close attention to having a clean /tmp after running
tests, so I pretty confident that my modules are working with
Bio::Root::IO properly and that Bio::Root::IO is working properly with
them.

Sorry I can't be more helpful.
Dan

On Fri, 2010-02-05 at 23:49 -0500, Mark A. Jensen wrote:
> Dan has been partial to temp in Root::IO (bowtie); I scurried off to File::Temp 
> (sam)--
> any issues with tempfile handles, Dan?
> ----- Original Message ----- 
> From: "Chris Fields" <cjfields at illinois.edu>
> To: "Florent Angly" <florent.angly at gmail.com>
> Cc: <bioperl-l at lists.open-bio.org>
> Sent: Friday, February 05, 2010 11:41 PM
> Subject: Re: [Bioperl-l] Too many links at LoadHelper.pm
> 
> 
> > The only place I can think of tempfiles being generated offhand are 
> > Bio::SeqFeature::Collection (actually, tied handles), but as you mention this 
> > may not be related.  There are calls to tempfile from 
> > Bio::Assembly::IO::sam/bowtie (one using Root::IO, one using File::Temp 
> > directly).
> >
> > chris
> >
> > On Feb 5, 2010, at 9:24 PM, Florent Angly wrote:
> >
> >> It can't be, Chris, because I use the workaround mentioned here: 
> >> http://bugzilla.open-bio.org/show_bug.cgi?id=2577#c5
> >>
> >> It seems like temporary files keep accumulating assembly after assembly. 
> >> Looking at the lsof command, I get a lot of files described as "(deleted)". I 
> >> am thinking that maybe some filehandle is not closed or something.
> >>
> >> I'll investigate...
> >>
> >> Florent
> >>
> >> On 06/02/10 13:13, Chris Fields wrote:
> >>> Florent,
> >>>
> >>> The Bio::Assembly error is documented in Bugzilla:
> >>>
> >>> http://bugzilla.open-bio.org/show_bug.cgi?id=2577
> >>>
> >>> Switching to a Bio::SeqFeature::CollectionI capable of distinguishing 
> >>> features by seq_id (such as Bio::DB::SeqFeature::Store) may be the best way 
> >>> to go.
> >>>
> >>> chris
> >>>
> >>> On Feb 5, 2010, at 9:06 PM, Florent Angly wrote:
> >>>
> >>>
> >>>> I've run into this issue very recently, albeit working on something 
> >>>> unrelated to loading GFF files: generating many small independent 
> >>>> assemblies. Maybe there's a bug in the tempdir() code, or its documentation 
> >>>> is not clear. I will take a look at it when I have some time.
> >>>> Florent
> >>>>
> >>>>
> >>>> On 05/02/10 22:27, Mark A. Jensen wrote:
> >>>>
> >>>>> The tempfiles are not being unlinked before there are too many for the OS 
> >>>>> to handle.
> >>>>> Sounds like this may be your job, fungazid, if you're not employing the 
> >>>>> script (but
> >>>>> not sure about that)
> >>>>> ----- Original Message ----- From: "Fungazid"<fungazid at yahoo.com>
> >>>>> To:<Bioperl-l at lists.open-bio.org>
> >>>>> Sent: Thursday, February 04, 2010 9:52 PM
> >>>>> Subject: [Bioperl-l] Too many links at LoadHelper.pm
> >>>>>
> >>>>>
> >>>>>
> >>>>>> Bioperl hi,
> >>>>>>
> >>>>>> I have multiple gff3 files that look like:
> >>>>>>
> >>>>>> --------------------------------------------------------
> >>>>>> ##gff-version 3
> >>>>>> # file: temp1.gff3
> >>>>>> contig453278    hs    contig    1    699    .    +    . 
> >>>>>> Name=contig453278
> >>>>>> contig453278    hs    alignment    1    194    .    +    . 
> >>>>>> Name=hs.ENST00000258455
> >>>>>> contig453278    hs    alignment    252    699    .    +    . 
> >>>>>> Name=hs.ENST00000258455
> >>>>>> contig453278    hs    mi2    194    252    .    +    .    Name=mi2
> >>>>>> contig453278    hs    transcript    104    183    .    +    . 
> >>>>>> Name=hs.ENST00000258455;ID=transcript-hs.ENST00000258455;Note=CD,4
> >>>>>> contig453278    hs    transcript    1    9    .    +    . 
> >>>>>> Name=hs.ENST00000258455;ID=transcript-hs.ENST00000258455;Note=CD,1
> >>>>>> contig453278    hs    transcript    73    103    .    +    . 
> >>>>>> Name=hs.ENST00000258455;ID=transcript-hs.ENST00000258455;Note=CD,3
> >>>>>> contig453278    hs    transcript    10    72    .    +    . 
> >>>>>> Name=hs.ENST00000258455;ID=transcript-hs.ENST00000258455;Note=CD,2
> >>>>>> ---------------------------------------------------------
> >>>>>>
> >>>>>>
> >>>>>>
> >>>>>> I uploaded gff3 files to mysql and view the tracks in gBrowse like this:
> >>>>>>
> >>>>>> --------------------------------------------
> >>>>>> Bio::DB::SeqFeature::Store->new(...)
> >>>>>> $loader=Bio::DB::SeqFeature::Store::GFF3Loader->new(...)
> >>>>>> for($i=0; $i<$#files+1; $i++){
> >>>>>> $loader->load($file[$i]);
> >>>>>> }
> >>>>>> --------------------------------------------
> >>>>>>
> >>>>>> Eventually after uploading ~15000 small gff3 files I got the error:
> >>>>>>
> >>>>>> Error in tempdir() using /tmp/XXXXXXXXXX: Could not create directory 
> >>>>>> /tmp/9bbpdE5ibf: Too many links at 
> >>>>>> /usr/local/share/perl/5.8.8/Bio/DB/SeqFeature/Store/LoadHelper.pm line 49
> >>>>>>
> >>>>>>
> >>>>>>
> >>>>>> I am trying to see what went wrong, maybe you can rescue me,
> >>>>>> Avi
> >>>>>>
> >>>>>>
> >>>>>>
> >>>>>>
> >>>>>> _______________________________________________
> >>>>>> Bioperl-l mailing list
> >>>>>> Bioperl-l at lists.open-bio.org
> >>>>>> http://lists.open-bio.org/mailman/listinfo/bioperl-l
> >>>>>>
> >>>>>>
> >>>>>>
> >>>>> _______________________________________________
> >>>>> Bioperl-l mailing list
> >>>>> Bioperl-l at lists.open-bio.org
> >>>>> http://lists.open-bio.org/mailman/listinfo/bioperl-l
> >>>>>
> >>>> _______________________________________________
> >>>> Bioperl-l mailing list
> >>>> Bioperl-l at lists.open-bio.org
> >>>> http://lists.open-bio.org/mailman/listinfo/bioperl-l
> >>>>
> >>>
> >>
> >> _______________________________________________
> >> Bioperl-l mailing list
> >> Bioperl-l at lists.open-bio.org
> >> http://lists.open-bio.org/mailman/listinfo/bioperl-l
> >
> >
> > _______________________________________________
> > Bioperl-l mailing list
> > Bioperl-l at lists.open-bio.org
> > http://lists.open-bio.org/mailman/listinfo/bioperl-l
> >
> > 





More information about the Bioperl-l mailing list