SGE -Y flag

[Charles Murphy]

When running NF processes on SGE it uses the the -Y qsub flag. How critical is it that NF has that flag set? I ask because there is something odd going on with my cluster setup. Where if I use -Y flag when submitting jobs the job fails.

I' am asking this really just out of curiosity. Though I' am working with my cluster admins to solve this be it is unclear so far.

[Paolo Di Tommaso]

I'm not aware of any -Y option added by NF. 

Are you sure that you are not adding in the `clusterOptions` setting ? 

p

On Tue, May 30, 2017 at 8:05 PM, Charles Murphy <murphy....@gmail.com> wrote:

When running NF processes on SGE it uses the the -Y qsub flag. How critical is it that NF has that flag set? I ask because there is something odd going on with my cluster setup. Where if I use -Y flag when submitting jobs the job fails.

I' am asking this really just out of curiosity. Though I' am working with my cluster admins to solve this be it is unclear so far.

--
You received this message because you are subscribed to the Google Groups "Nextflow" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nextflow+unsubscribe@googlegroups.com.
Visit this group at https://groups.google.com/group/nextflow.
For more options, visit https://groups.google.com/d/optout.

[Charles Murphy]

Apologies, I mean the -V option. Here is the first portion of an example .command.run script:

#!/bin/bash
#$ -wd /athena/elementolab/scratch/chm2059/from_dat02/chm2059/work/66/9812c97d235ca019d1d7527a5aaadc
#$ -N nf-test_(3)
#$ -o /athena/elementolab/scratch/chm2059/from_dat02/chm2059/work/66/9812c97d235ca019d1d7527a5aaadc/.command.log
#$ -j y
#$ -terse
#$ -V
#$ -notify
#$ -l h_vmem=1G

# NEXTFLOW TASK: test (3)




On Tuesday, May 30, 2017 at 2:29:27 PM UTC-4, Paolo Di Tommaso wrote:

I'm not aware of any -Y option added by NF. 

Are you sure that you are not adding in the `clusterOptions` setting ? 

p

On Tue, May 30, 2017 at 8:05 PM, Charles Murphy <murphy....@gmail.com> wrote:

When running NF processes on SGE it uses the the -Y qsub flag. How critical is it that NF has that flag set? I ask because there is something odd going on with my cluster setup. Where if I use -Y flag when submitting jobs the job fails.

I' am asking this really just out of curiosity. Though I' am working with my cluster admins to solve this be it is unclear so far.

--
You received this message because you are subscribed to the Google Groups "Nextflow" group.

To unsubscribe from this group and stop receiving emails from it, send an email to nextflow+u...@googlegroups.com.

[Paolo Di Tommaso]

That option export the current environment to the job. See here.

It looks very strange that is a problem in your cluster. Is this the first time you run NF on this cluster? Have you tried to ask help to your sysadmins ?

p

To unsubscribe from this group and stop receiving emails from it, send an email to nextflow+unsubscribe@googlegroups.com.

[Charles Murphy]

I been running NF on the same cluster for a year or two now, but was recently switched to a new master/head node to submit my jobs (but still to the same cluster). I’ve been in contact with my system admins, but no solutions yet. There is just something strange with how my environment is setup.

But this issue is not specific to NF. If I create a job outside NF and include the -V option then the job fails.

You received this message because you are subscribed to a topic in the Google Groups "Nextflow" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/nextflow/4YJsyEgIhpA/unsubscribe.
To unsubscribe from this group and all its topics, send an email to nextflow+u...@googlegroups.com.

[Paolo Di Tommaso]

I see. Let me know how this progress. 

p

[Emilio]

Hi Charles,

are you exporting LD_LIBRARY_PATH or similar in your environment? I know that latest versions of SGE do not allow this variable when using the `-V` option for security reasons.

Maybe your problem is related to this.

Cheers,

Emilio

[Paolo Di Tommaso]

I've noticed that NF is exporting the environment only for the SGE,PBE and NQSII executors. I'm starting to think that it would be better to remove it for consistency with all batch executors. 

p

[Charles Murphy]

Emilio, I do see the LD_LIBRARY_PATH environment variable, though I also see it defined on the machine I used to run NF off of as well (where NF does work for me). Both machines are running SGE 6.2u5, which does not seem to be the most recent version.

I’ve made a post about my issue on stack overflow if you wish to understand it more: https://unix.stackexchange.com/questions/367625/job-environment-not-being-setup-correctly-on-sge

Paolo, yeah if it is not too much trouble to exclude the -Y option. Because I currently cannot make use of my cluster right now because of it. Though, thankfully, the master/head node is also meant for interactive work (has 56 cores and 500GB memory), so I’ am using NF there.

You received this message because you are subscribed to a topic in the Google Groups "Nextflow" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/nextflow/4YJsyEgIhpA/unsubscribe.
To unsubscribe from this group and all its topics, send an email to nextflow+u...@googlegroups.com.

[Paolo Di Tommaso]

I've created a ticket to follow this issue 

https://github.com/nextflow-io/nextflow/issues/362

p

[Charles Murphy]

Great!