bammds-users
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Bammds-users] ASD file error


From: Ole Tange
Subject: Re: [Bammds-users] ASD file error
Date: Sun, 7 Sep 2014 16:52:21 +0200

Victor's suggestion is along the same lines as mine: Try to put the
'tmp' dir on a local file system and not on a NFS-mounted filesystem.

I have just been able to reproduce your error:

  cd ~/data
  seq 1 | TMPDIR=. bammds_parallel echo
  Error unlinking file ./eKEo4j8p2o using unlink0 at
/usr/local/bin/bammds_parallel line 5150.

And it is clearly not an error that exists in GNU Parallel < 20131222,
so it is likely fixable.

Until I get that fixed, do not use a NFS-mounted dir as the tmpdir.

/Ole

On Sun, Sep 7, 2014 at 4:16 PM, José Víctor Moreno Mayar
<address@hidden> wrote:
> Hey Morten,
>
> Can you try running this on watt before running bammds?
>
> export TMPDIR=$HOME/scratch
>
> Víctor
>
> On Sun, Sep 7, 2014 at 4:07 PM, Morten Allentoft
> <address@hidden> wrote:
>>
>> Thanks Ole
>>
>> It turned out to be some weird access issue, I think. When getting the
>> error message I was running bammds on watt, whereas it worked fine on eos.
>>
>> Any idea why this might be the case?
>>
>> Cheers,
>> Morten
>>
>> On Sun, Sep 7, 2014 at 12:54 PM, Ole Tange <address@hidden> wrote:
>>>
>>> On Sun, Sep 7, 2014 at 1:56 AM, Morten Allentoft
>>> <address@hidden> wrote:
>>> > Hi guys
>>> >
>>> > I'm getting an annoying error message that you perhaps can help me
>>> > with?
>>> >
>>> > When I go:
>>> > bammds MA840_L1_S2_trim15_PMDS_filter3.bam
>>> > /k/genotypes/HGDP_hg19_one_allele.txt
>>> >
>>> > I get this:
>>> > Error unlinking file tmp/EFD6SGwO2i using unlink0 at
>>> > /usr/local/bin/bammds_parallel line 5150.
>>>
>>> This should never happen.
>>>
>>> It has most likely nothing to do with your data, but may have
>>> something to do with your Linux environment.
>>>
>>> I ran:
>>>
>>>   cd /tmp
>>>   bammds
>>> /home/meallentoft/data/Skhul/mapped_15bp/MA840_L1_S2_trim15_PMDS_filter3.bam
>>> /k/genotypes/HGDP_hg19_one_allele.txt
>>>
>>> and that works.
>>>
>>> So that leads me to believe that the filesystem on
>>> /home/meallentoft/data/Skhul/mapped_15bp is teasing you. If I look
>>> right now, I see no 'tmp' dir there - and that is definitely bad, as
>>> bammds creates that and uses that intensively. When finished
>>> /home/meallentoft/data/Skhul/mapped_15bp/tmp should contain the same
>>> as: /scratch/tange/tmp
>>>
>>> Can you do this by hand:
>>>
>>>   cd /home/meallentoft/data/Skhul/mapped_15bp
>>>   mkdir tmp
>>>
>>> and run it again? Do you get the same error?
>>>
>>>
>>> /Ole
>>
>>
>>
>>
>> --
>> Morten E. Allentoft
>> Postdoc, PhD
>>
>> Centre for GeoGenetics
>> Natural History Museum
>> Øster Voldgade 5-7
>> DK-1350 Copenhagen
>>
>> +45 35 32 12 91 (office)
>> +45 29 82 46 34 (mobile)
>> http://geogenetics.ku.dk
>>
>> Sign up for our newsletter!
>
>
>
>
> --
> José Víctor Moreno Mayar
> Centre for GeoGenetics
> University of Copenhagen



reply via email to

[Prev in Thread] Current Thread [Next in Thread]