These are chat archives for nextflow-io/nextflow

7th
Oct 2018
Ashley S Doane
@DoaneAS
Oct 07 2018 01:20
@pditommaso thanks!

question on error strategy, if I have something like this:

memory { 4.GB * task.attempt }
errorStrategy { task.exitStatus == 140 ? 'retry' : 'terminate' }

How would I change this behavior to start with 8 GB and add 2 GB if it fails, something like:

memory {8.GB + 2.GB * task.attempt }

thanks

Paolo Di Tommaso
@pditommaso
Oct 07 2018 11:04
@aunderwo what makes you think it should use more ?
Anthony Underwood
@aunderwo
Oct 07 2018 11:06

By this I mean that in the trace file all the processes which are high mem usage report a rss value of 1G exactly

I've run the same SPAdes assembly workflow locally and get the same process reporting exactly 1Gb rss no matter what the sample. So I conclude this is probably not a Nextflow thing but perhaps a feature of SPAdes which defines exactly 1024Mb of memory to be used . I still find it a little bit of a black box.

Paolo Di Tommaso
@pditommaso
Oct 07 2018 11:07
NF just pass the mem limit on the docker command line
Diogo Silva
@ODiogoSilva
Oct 07 2018 12:21
We've been using spades a lot and the memory usage fluctuates, but mostly above 1Gb
Diogo Silva
@ODiogoSilva
Oct 07 2018 12:27
Spades does have a parameter do limit memory usage (-m), though