Re: /tmp

classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

Re: /tmp

Carson Hinton Holt
Re: [maker-devel] /tmp Sorry this took so long.  I’ve added debugging messages to the latest MAKER subversion repository.  They have the tag TMP_STAT on the front and they also list the process ID so you can see which messages go together.  If you run maker with the –qq option, then all other status messages should be suppressed, which will make it easy to look through only the TMP_STAT messages.

Thanks,
Carson

On 5/3/10 10:11 PM, "Aleksandr Levchuk" <alevchuk@...> wrote:

Dear Carson,

I upgraded to r369, but I still have temp files written to /tmp

In maker_opts.ctl I have:
TMP: /dev/shm/maker-alex/ #specify a directory other than the system default ...


but, I get nothing in /dev/shm/maker-alex and there are fresh files in:
  ls /tmp/maker_HdBcQF/
  te_proteins.mpi.40.6.fasta.dbformat.NFSLock.shared
  te_proteins.mpi.40.6.fasta.dbformat.NFSLock.tmp.4494.10096.7525
  te_proteins.mpi.40.6.fasta.phr
  te_proteins.mpi.40.6.fasta.pin
  te_proteins.mpi.40.6.fasta.psq

I think this is the output of NCBI's formatdb.


Last week when I did the earlier test runs, I was able to get away with:
  mount -o bind    /dev/shm/maker-alex   /tmp

It worked. Unfortunately, I shouldn't leave this bind mount
permanently on the compute cluster, so it would be very desired to
have the TMP option working.

Alex

On Fri, Apr 30, 2010 at 4:24 PM, Carson Holt
<carson.holt@...> wrote:
> Just incase it helps, I’ve made MAKER a little more aggressive in trying to
> establish access to the indicated TMP directory before reverting to /tmp.
>  You can update via subversion.
>
> Carson
>
>
> On 4/30/10 2:38 PM, "Aleksandr Levchuk" <alevchuk@...> wrote:
>
> Dear Maker Developers,
>
> Is there a way to configure Maker so that everything that it writes in
> /tmp would go somewhere else?
>
> In /tmp Maker puts the Blast formatdb files.
>
> Setting the "TMP:" parameter in maker_opts.ctl does not help.
>
> Alex
>
> --
> ---------------------------------------------------------------------------------
> Aleksandr Levchuk
> Administrator of Bioinformatic Systems and Databases
>
> Homepage: http://biocluster.ucr.edu/~alevchuk/
> Cell Phone: (951) 368-0004
> Lab Phone: (951) 905-5232
>
> Institute for Integrative Genome Biology
> University of California, Riverside
> ---------------------------------------------------------------------------------
>
> _______________________________________________
> maker-devel mailing list
> maker-devel@...
> http://box290.bluehost.com/mailman/listinfo/maker-devel_yandell-lab.org
>
>



--
---------------------------------------------------------------------------------
Aleksandr Levchuk
Administrator of Bioinformatic Systems and Databases

Homepage: http://biocluster.ucr.edu/~alevchuk/
Cell Phone: (951) 368-0004
Lab Phone: (951) 905-5232

Institute for Integrative Genome Biology
University of California, Riverside
---------------------------------------------------------------------------------


_______________________________________________
maker-devel mailing list
[hidden email]
http://box290.bluehost.com/mailman/listinfo/maker-devel_yandell-lab.org
Reply | Threaded
Open this post in threaded view
|

Re: /tmp

Aleksandr Levchuk
Great! Thanks!

I will test that today.

Alex

On Fri, May 28, 2010 at 12:47 PM, Carson Holt
<[hidden email]> wrote:

> Sorry this took so long.  I’ve added debugging messages to the latest MAKER
> subversion repository.  They have the tag TMP_STAT on the front and they
> also list the process ID so you can see which messages go together.  If you
> run maker with the –qq option, then all other status messages should be
> suppressed, which will make it easy to look through only the TMP_STAT
> messages.
>
> Thanks,
> Carson
>
> On 5/3/10 10:11 PM, "Aleksandr Levchuk" <[hidden email]> wrote:
>
> Dear Carson,
>
> I upgraded to r369, but I still have temp files written to /tmp
>
> In maker_opts.ctl I have:
> TMP: /dev/shm/maker-alex/ #specify a directory other than the system default
> ...
>
>
> but, I get nothing in /dev/shm/maker-alex and there are fresh files in:
>   ls /tmp/maker_HdBcQF/
>   te_proteins.mpi.40.6.fasta.dbformat.NFSLock.shared
>   te_proteins.mpi.40.6.fasta.dbformat.NFSLock.tmp.4494.10096.7525
>   te_proteins.mpi.40.6.fasta.phr
>   te_proteins.mpi.40.6.fasta.pin
>   te_proteins.mpi.40.6.fasta.psq
>
> I think this is the output of NCBI's formatdb.
>
>
> Last week when I did the earlier test runs, I was able to get away with:
>   mount -o bind    /dev/shm/maker-alex   /tmp
>
> It worked. Unfortunately, I shouldn't leave this bind mount
> permanently on the compute cluster, so it would be very desired to
> have the TMP option working.
>
> Alex
>
> On Fri, Apr 30, 2010 at 4:24 PM, Carson Holt
> <[hidden email]> wrote:
>> Just incase it helps, I’ve made MAKER a little more aggressive in trying
>> to
>> establish access to the indicated TMP directory before reverting to /tmp.
>>  You can update via subversion.
>>
>> Carson
>>
>>
>> On 4/30/10 2:38 PM, "Aleksandr Levchuk" <[hidden email]> wrote:
>>
>> Dear Maker Developers,
>>
>> Is there a way to configure Maker so that everything that it writes in
>> /tmp would go somewhere else?
>>
>> In /tmp Maker puts the Blast formatdb files.
>>
>> Setting the "TMP:" parameter in maker_opts.ctl does not help.
>>
>> Alex
>>
>> --
>>
>> ---------------------------------------------------------------------------------
>> Aleksandr Levchuk
>> Administrator of Bioinformatic Systems and Databases
>>
>> Homepage: http://biocluster.ucr.edu/~alevchuk/
>> Cell Phone: (951) 368-0004
>> Lab Phone: (951) 905-5232
>>
>> Institute for Integrative Genome Biology
>> University of California, Riverside
>>
>> ---------------------------------------------------------------------------------
>>
>> _______________________________________________
>> maker-devel mailing list
>> [hidden email]
>> http://box290.bluehost.com/mailman/listinfo/maker-devel_yandell-lab.org
>>
>>
>
>
>
> --
> ---------------------------------------------------------------------------------
> Aleksandr Levchuk
> Administrator of Bioinformatic Systems and Databases
>
> Homepage: http://biocluster.ucr.edu/~alevchuk/
> Cell Phone: (951) 368-0004
> Lab Phone: (951) 905-5232
>
> Institute for Integrative Genome Biology
> University of California, Riverside
> ---------------------------------------------------------------------------------
>
>



--
---------------------------------------------------------------------------------
Aleksandr Levchuk
Administrator of Bioinformatic Systems and Databases

Homepage: http://biocluster.ucr.edu/~alevchuk/
Cell Phone: (951) 368-0004
Lab Phone: (951) 905-5232

Institute for Integrative Genome Biology
University of California, Riverside
---------------------------------------------------------------------------------

_______________________________________________
maker-devel mailing list
[hidden email]
http://box290.bluehost.com/mailman/listinfo/maker-devel_yandell-lab.org
Reply | Threaded
Open this post in threaded view
|

Re: /tmp

Aleksandr Levchuk
Dear Carson,

To capture the /tmp problem with your TMP logging, I was burning a lot
of CPU for the last 2 days with maker-r400 (svn revision 400).

I got to reproduce the problem once but I lost the output file.
Once I have something definite - I will send it to you, but not yet.

For now, here is one of my output files:
  http://biocluster.ucr.edu/~alevchuk/maker-r400-out

Alex

On Fri, May 28, 2010 at 2:53 PM, Aleksandr Levchuk <[hidden email]> wrote:

>
> Great! Thanks!
>
> I will test that today.
>
> Alex
>
> On Fri, May 28, 2010 at 12:47 PM, Carson Holt
> <[hidden email]> wrote:
> > Sorry this took so long.  I’ve added debugging messages to the latest MAKER
> > subversion repository.  They have the tag TMP_STAT on the front and they
> > also list the process ID so you can see which messages go together.  If you
> > run maker with the –qq option, then all other status messages should be
> > suppressed, which will make it easy to look through only the TMP_STAT
> > messages.
> >
> > Thanks,
> > Carson
> >
> > On 5/3/10 10:11 PM, "Aleksandr Levchuk" <[hidden email]> wrote:
> >
> > Dear Carson,
> >
> > I upgraded to r369, but I still have temp files written to /tmp
> >
> > In maker_opts.ctl I have:
> > TMP: /dev/shm/maker-alex/ #specify a directory other than the system default
> > ...
> >
> >
> > but, I get nothing in /dev/shm/maker-alex and there are fresh files in:
> >   ls /tmp/maker_HdBcQF/
> >   te_proteins.mpi.40.6.fasta.dbformat.NFSLock.shared
> >   te_proteins.mpi.40.6.fasta.dbformat.NFSLock.tmp.4494.10096.7525
> >   te_proteins.mpi.40.6.fasta.phr
> >   te_proteins.mpi.40.6.fasta.pin
> >   te_proteins.mpi.40.6.fasta.psq
> >
> > I think this is the output of NCBI's formatdb.
> >
> >
> > Last week when I did the earlier test runs, I was able to get away with:
> >   mount -o bind    /dev/shm/maker-alex   /tmp
> >
> > It worked. Unfortunately, I shouldn't leave this bind mount
> > permanently on the compute cluster, so it would be very desired to
> > have the TMP option working.
> >
> > Alex
> >
> > On Fri, Apr 30, 2010 at 4:24 PM, Carson Holt
> > <[hidden email]> wrote:
> >> Just incase it helps, I’ve made MAKER a little more aggressive in trying
> >> to
> >> establish access to the indicated TMP directory before reverting to /tmp.
> >>  You can update via subversion.
> >>
> >> Carson
> >>
> >>
> >> On 4/30/10 2:38 PM, "Aleksandr Levchuk" <[hidden email]> wrote:
> >>
> >> Dear Maker Developers,
> >>
> >> Is there a way to configure Maker so that everything that it writes in
> >> /tmp would go somewhere else?
> >>
> >> In /tmp Maker puts the Blast formatdb files.
> >>
> >> Setting the "TMP:" parameter in maker_opts.ctl does not help.
> >>
> >> Alex
> >>
> >> --
> >>
> >> ---------------------------------------------------------------------------------
> >> Aleksandr Levchuk
> >> Administrator of Bioinformatic Systems and Databases
> >>
> >> Homepage: http://biocluster.ucr.edu/~alevchuk/
> >> Cell Phone: (951) 368-0004
> >> Lab Phone: (951) 905-5232
> >>
> >> Institute for Integrative Genome Biology
> >> University of California, Riverside
> >>
> >> ---------------------------------------------------------------------------------
> >>
> >> _______________________________________________
> >> maker-devel mailing list
> >> [hidden email]
> >> http://box290.bluehost.com/mailman/listinfo/maker-devel_yandell-lab.org
> >>
> >>
> >
> >
> >
> > --
> > ---------------------------------------------------------------------------------
> > Aleksandr Levchuk
> > Administrator of Bioinformatic Systems and Databases
> >
> > Homepage: http://biocluster.ucr.edu/~alevchuk/
> > Cell Phone: (951) 368-0004
> > Lab Phone: (951) 905-5232
> >
> > Institute for Integrative Genome Biology
> > University of California, Riverside
> > ---------------------------------------------------------------------------------
> >
> >
>
>
>
> --
> ---------------------------------------------------------------------------------
> Aleksandr Levchuk
> Administrator of Bioinformatic Systems and Databases
>
> Homepage: http://biocluster.ucr.edu/~alevchuk/
> Cell Phone: (951) 368-0004
> Lab Phone: (951) 905-5232
>
> Institute for Integrative Genome Biology
> University of California, Riverside
> ---------------------------------------------------------------------------------


--
---------------------------------------------------------------------------------
Aleksandr Levchuk
Administrator of Bioinformatic Systems and Databases

Homepage: http://biocluster.ucr.edu/~alevchuk/
Cell Phone: (951) 368-0004
Lab Phone: (951) 905-5232

Institute for Integrative Genome Biology
University of California, Riverside
---------------------------------------------------------------------------------

_______________________________________________
maker-devel mailing list
[hidden email]
http://box290.bluehost.com/mailman/listinfo/maker-devel_yandell-lab.org

maker-r400-out.gz (3K) Download Attachment