These are chat archives for nextflow-io/nextflow

24th
Nov 2017
Phil Ewels
@ewels
Nov 24 2017 15:39
Hi @pditommaso! Maybe I've asked you this before, but is there any way to know about processes that failed and were ignored?
We have a few non-critical processes set like this, so that they don't trip up the entire pipeline if something goes wrong (eg. FastQC: https://github.com/SciLifeLab/NGI-RNAseq/blob/master/conf/base.config#L36-L38)
The current behaviour is fine, except that I'd like to be able to throw some kind of warning to the user in the final report that not all processes were successful
Paolo Di Tommaso
@pditommaso
Nov 24 2017 15:40
are they not included in the html report ?
Phil Ewels
@ewels
Nov 24 2017 15:41
Probably, but you still get a big green box at the top saying that everything was fine
Instead of fail / success, I'd like to have fail / warn / success
Paolo Di Tommaso
@pditommaso
Nov 24 2017 15:41
oohh
:)
Phil Ewels
@ewels
Nov 24 2017 15:42
Perhaps I can check through all of the task.status values to look for anything that isn't CACHED or SUCCESS or somethintg
but that's tricky, as something may fail once but succeed on the second attempt
Paolo Di Tommaso
@pditommaso
Nov 24 2017 15:43
I don't know, maybe we could collect warning messages in metadata property
Phil Ewels
@ewels
Nov 24 2017 15:44
could be nice... or have a counter for ignored errors or something
Paolo Di Tommaso
@pditommaso
Nov 24 2017 15:44
make makes sense, a kind of tasks status
number of successfully completed, failed, ignored, etc
Phil Ewels
@ewels
Nov 24 2017 15:45
yeah exactly
in this case then it would be fine if we have failed tasks that were resubmitted and succeeded
as it's the ignored tasks that we want to warn about
Paolo Di Tommaso
@pditommaso
Nov 24 2017 16:08
ok, please open a issue for that on GH
Phil Ewels
@ewels
Nov 24 2017 16:11
Will do :+1:
nextflow-io/nextflow#534
Paolo Di Tommaso
@pditommaso
Nov 24 2017 16:14
:clap: