These are chat archives for nextflow-io/nextflow

6th
Dec 2018
Raoul J.P. Bonnal
@helios
Dec 06 2018 09:20

I am experiencing this problem with PBS (NotPro)

$ pbs-config --version
4.2.5
N E X T F L O W  ~  version 18.10.1
Launching `main.nf` [distracted_coulomb] - revision: 1d8a541966
[warm up] executor > pbs
[6d/71054f] Submitted process > velocyto (1)
WARN: [PBS] queue (myq) status cannot be fetched > exit status: 2

dic-06 08:52:40.699 [Task monitor] WARN  n.executor.AbstractGridExecutor - [PBS] queue (myq) status cannot be fetched > exit status: 2

I saw a similar issue about lsf.

Paolo Di Tommaso
@pditommaso
Dec 06 2018 09:42
please open an issue on github including the .nextflow.log file
you may want to give a try to latest edge release
NXF_VER=18.11.0-edge nextflow run .. etc
Raoul J.P. Bonnal
@helios
Dec 06 2018 10:02
@pditommaso I will try the command and open the issue
Paolo Di Tommaso
@pditommaso
Dec 06 2018 10:02
:ok_hand:
Alexander Peltzer
@apeltzer
Dec 06 2018 12:03
Hm seeing this here:

java.text.ParseException: Unparseable date: "Tue, 10 Apr 2018 13:39:14 GMT"
        at java.text.DateFormat.parse(DateFormat.java:366)
        at nextflow.file.http.XFileSystemProvider.readHttpAttributes(XFileSystemProvider.groovy:392)
        at nextflow.file.http.XFileSystemProvider.readHttpAttributes(XFileSystemProvider.groovy:383)
        at nextflow.file.http.XFileSystemProvider.readAttributes(XFileSystemProvider.groovy:357)
        at java.nio.file.Files.readAttributes(Files.java:1737)
when doing this here: https://github.com/nf-core/ampliseq/blob/a584a06840afde3ee0c11d37ed3c367c5529b54c/main.nf#L407
Paolo Di Tommaso
@pditommaso
Dec 06 2018 12:08
weird, the format seems ok
Alexander Peltzer
@apeltzer
Dec 06 2018 12:09
Yeah and I know that this worked already :-)
Looks good to me as well ?
Paolo Di Tommaso
@pditommaso
Dec 06 2018 12:11
open an issue with a snippet replicating the error
I guess that just file('http://blah/blah').text should fail as well
Alexander Peltzer
@apeltzer
Dec 06 2018 12:18
I'll do that
Can replicate it , tried loading different ENV modules with java8, nothing changes
ok
Thanks a bunch!
Paolo Di Tommaso
@pditommaso
Dec 06 2018 12:20
:ok_hand:
Alexander Peltzer
@apeltzer
Dec 06 2018 12:26
Let me know if you need more
Paolo Di Tommaso
@pditommaso
Dec 06 2018 12:28
I'll try asap
Alexander Peltzer
@apeltzer
Dec 06 2018 12:30
No worries if not, I'll try the file(...).text way as well
Johannes Alneberg
@alneberg
Dec 06 2018 14:02

Hello! I'm trying to limit the number of jobs sent to aws batch at once by using the '-queue-size' parameter, but I can't seem to get it to work. It still sends off ~ 100 jobs. Can you spot any obvious errors in my command:

nextflow run -queue-size 10 ~/Sarek/germlineVC.nf -profile awsbatch -work-dir s3://mybucket/work --outDir s3://mybucket/results --sample s3://mybucket/Preprocessing/Recalibrated/recalibrated.tsv --genome iGRCh38 --genome_base s3://ngi-igenomes/igenomes/Homo_sapiens/GATK/GRCh38 --verbose --tools HaplotypeCaller,Manta,Strelka -resume

I've tried having -queue-size 10 last as well.

Paolo Di Tommaso
@pditommaso
Dec 06 2018 14:04
should be fine, check in the log file, it's reported at the beginning
Johannes Alneberg
@alneberg
Dec 06 2018 14:05
Seems to pick it up yes: ...poolSize:5, maxPoolSize:5, queueSize:10, maxRetries:10,...
But it should then not submit more than 10 processes and then wait until any of them finishes before continuing, am I right?
Paolo Di Tommaso
@pditommaso
Dec 06 2018 14:08
actually yes
Johannes Alneberg
@alneberg
Dec 06 2018 14:12
Well, I still get ~100 processes submitted at once
Paolo Di Tommaso
@pditommaso
Dec 06 2018 14:13
weird, create an issue with the log file
Johannes Alneberg
@alneberg
Dec 06 2018 14:14
Will do! Thank you!
Paolo Di Tommaso
@pditommaso
Dec 06 2018 14:14
:+1:
Johannes Alneberg
@alneberg
Dec 06 2018 16:25
A different question. I just discovered the disk directive for processes. Is this used with the awsbatch executor, and in that case how? This could be exactly what I've been missing when processes running on the same instances take up too much disk space combined together.
Paolo Di Tommaso
@pditommaso
Dec 06 2018 16:26
nope, all this area need to be improved
Johannes Alneberg
@alneberg
Dec 06 2018 16:27
Ah, sad. I guess it's a very good start that there is such a directive
Paolo Di Tommaso
@pditommaso
Dec 06 2018 16:27
unfortunately there isn;t a direct mapping with Batch API
Johannes Alneberg
@alneberg
Dec 06 2018 16:28
I'm completely lost when it comes to what governs what jobs will be run on the same instances. Is that something Nextflow can affect?
Paolo Di Tommaso
@pditommaso
Dec 06 2018 16:29
yes, but it's not so trivial, however it's definitively something I want to implement
Johannes Alneberg
@alneberg
Dec 06 2018 16:31
I can understand that. Sounds reassuring at least
Paolo Di Tommaso
@pditommaso
Dec 06 2018 16:31
LOL
Johannes Alneberg
@alneberg
Dec 06 2018 16:33
While I am at it. Is there a way for my AWS batch jobs to be able to share the input files if they are using the same? Without using the newly launched Lustre solution that is. Currently it seems like my jobs all copy over their own copy of bam files and reference files. Or am I misunderstanding something?
Paolo Di Tommaso
@pditommaso
Dec 06 2018 16:35
same problem where, there's no such concept in the Batch api AFAIK
you cannot specify where to run a job or what piece data it should use
however I had a chat with aws folks and they are aware of that, they plan to improve it
Johannes Alneberg
@alneberg
Dec 06 2018 16:38
I see. So I'm left to specifying a separate queue where the instance type limits the number of jobs to be run at the same time I guess
Paolo Di Tommaso
@pditommaso
Dec 06 2018 16:39
that sounds good
Johannes Alneberg
@alneberg
Dec 06 2018 16:40
Yes. That's probably what I'm aiming at as a solution now. We're not really at the point where we want to run aws in production yet. Only testing it to get a realistic pricing estimate
Paolo Di Tommaso
@pditommaso
Dec 06 2018 16:41
definitively makes sense
Johannes Alneberg
@alneberg
Dec 06 2018 16:41
Ok, thank you again for your help! I will let you know if it works out
Paolo Di Tommaso
@pditommaso
Dec 06 2018 16:41
welcome! and looking forward to reading that
Paolo Di Tommaso
@pditommaso
Dec 06 2018 17:06
@brwnj sorry, was missing your question; at this time trace data is not available from the script api
Joe Brown
@brwnj
Dec 06 2018 17:45
@pditommaso Thanks!
Paolo Di Tommaso
@pditommaso
Dec 06 2018 17:45
:+1: