These are chat archives for nextflow-io/nextflow

22nd
Aug 2018
Sander Bervoets
@Biocentric
Aug 22 2018 05:59
@pditommaso yeah, I'm also considering that its salmon. But indeed, the same script/command/main.nf works locally. Base image: Amazon ECS optimized. I recreated the image again to ensure i did it correctly
Sander Bervoets
@Biocentric
Aug 22 2018 07:04
@pditommaso If you mean base image for the container, it is nextflow/rnaseq-nf
In the job definition section of Batch, it shows up as nf-nextflow-rnaseq-nf. Not sure why. I'll create a different rnaseq docker image for debugging
Paolo Di Tommaso
@pditommaso
Aug 22 2018 09:36
@DoaneAS That error suggest that the compute node cannot properly execute the source activate command. Still possible the interactive environment (qlogin) is different from compute environment.
Sander Bervoets
@Biocentric
Aug 22 2018 10:24
@pditommaso @tbugfinder So I forked the RNA-Seq repo, changed some configs and made a new docker image which I pushed to my hub. Salmon is now v 0.11.2 and working fine on AWS Batch.
Paolo Di Tommaso
@pditommaso
Aug 22 2018 10:27
therefore was it a Salmon issue ?
Sander Bervoets
@Biocentric
Aug 22 2018 10:28
Guess so. But considering that it probably worked when you tried it couple of months ago, it makes me wonder what changed in the meantime.
Paolo Di Tommaso
@pditommaso
Aug 22 2018 10:33
I've seen bugged applications resulting in a segfaults error only in very specific mem configurations
Sander Bervoets
@Biocentric
Aug 22 2018 10:34
I tried several instance types, also increasing available memory. That didn't help
Alexander Peltzer
@apeltzer
Aug 22 2018 20:41
For the people struggling with AWSBatch setup with Nextflow, have a look here: https://apeltzer.github.io/post/01-aws-nfcore/