juicer on PBS, first job terminated then remaining jobs are orphaned

[aidenlab]

juicer on PBS, first job terminated then remaining jobs are orphaned

[mzhibo]

It might be due to your test fastq file size being smaller than the settings for multi-threads setting. see below:
if [ "$fastqsize" -gt "2592410750" ]
then
threads=16
fi

When i build the dependencies, I set the AlnWrp to work after the fastq splitting and mv jobs.
in AlnWrp job submission: #PBS -W depend=afterok:${jID_splitmv}
The job splitting and mv jobs is only placed when this if [ -n "$threads" ] && [ -z "$skipsplit" ] test is ture.

So, it will be fine if you work with fastq files larger than 2592410750

You can add the following lines between the hash tag lines
############################
if [ -n "$threads" ] && [ -z "$skipsplit" ]
then
waitstring_alnwrp="#PBS -W depend=afterok:${jID_splitmv}"
fi
echo "waitstring_alnwrp is ${waitstring_alnwrp}"
#####################
in front of the ALINWRAP job script (right before the following block.)
timestamp=$(date +"%s" | cut -c 4-10)
qsub <<ALIGNWRAP
#PBS -S /bin/bash
#PBS -q $queue
#PBS -l $walltime
#PBS -l nodes=1:ppn=1:AMD
#PBS -l mem=6gb
#PBS -o ${logdir}/${timestamp}alnwrap${groupname}.log
#PBS -j oe
#PBS -N AlnWrp${groupname}
$waitstring_alnwrp
${EMAIL}
#PBS -m a

I will also upload the new updated version. You can download the new juicer.sh launch_stats.sh and postprocessing.sh files from the github once the admin accepts the pull request.
I also fixed another bug this morning in postprocessing.sh.
Zhibo
@jstansfield0

[mzhibo]

Also check the log file in logs directory. If my previous guess is true, then the ${logdir}/${timestamp}alnwrap${groupname}.log file should tell you something about qsub -W error.
Zhibo

[jstansfield0]

From the log file: [190]stansfieldjc@merlot/home/sequencing/juicer/logs % cat 6519501_alnwrap_C19501.log qsub: Bad UID for job execution jID_cntlig 1 id is /var/spool/PBS/mom_priv/jobs/207474.merlot.bis.vcu.edu.SC: line 64: * 5 : syntax error: operand expected (error token is "* 5 ") qsub: Bad UID for job execution qsub: Bad UID for job execution

…

On Thu, Sep 28, 2017 at 3:21 PM, mzhibo ***@***.***> wrote: Also check the log file in logs directory. If my previous guess is true, then the ${logdir}/${timestamp}*alnwrap*${groupname}.log file should tell you something about qsub -W error. Zhibo — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#39 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ATJE8EGMtqIVcwKX0KgF_tzsQV9eZAgoks5sm_G6gaJpZM4PnnVB> .

[mzhibo]

@jstansfield0
it seems to be a problem that your cluster does not allow jobs to be submitted from computer nodes other than the login node.
The cluster in our school allows us to submit jobs from all the computer nodes.
I would suggest you to contact your culster team to see if then have any good suggestions.
I also just started to use linux and PBS, sorry I can't help more.
this link might be related to your issue.
https://lists.sdsc.edu/pipermail/npaci-rocks-discussion/2002-August/000211.html
Zhibo

[mzhibo]

The AlnWrp job was submitted and run on a computer node. This job will submit several other jobs. From your log file, it looks like the AlnWrp job was started successfully, but the qsub commands in AlnWrp job failed to run due to bad user ID error. This is usually due to the set-up of cluster that does not allow user to submit jobs from computer nodes other than the qlogin node.

[theaidenlab]

Let us know if this has been resolved.

[jstansfield0]

We have not been able to resolve the issue. Our cluster does not allow for jobs to be submitted from a node so the script does not work.

[nchernia]

Closing as we won't be able to resolve since we don't have access to PBS