Open massung opened 2 years ago
Bump. Just checking if there's any info or ideas on this. I only ask because currently none of our workflows can run on Tower.
Sorry for the late reply. Think we have identified the problem. We are going to release a patch in a couple of hours.
The patch was released. Please try again running your pipeline
Still failing to execute. Is there something I need to do to my pipeline or compute environment or should it "just work"?
Can you please include the nextflow output and log files?
Unfortunately it turns out it's glitch in the nextflow runtime. The only short term workaround is to add the setting nextflow.enable.dsl = 1
in the pipeline config either in the project repo or in the launch advanced settings (see below)
Yeah, that's making it work. At least I have a work-around. Thanks!
We released a new patch that does not require anymore the need for nextflow.enable.dsl = 1
definition in the launch config
I have several pipelines that have been running successfully (using AWS Batch) on Tower for months now. As of 2 days ago none of them will run. The compute environments are the same as are the pipelines.
They launch successfully and the initial spot instance (fleet instance?) is created successfully. Once provisioned and initialized the run ends within seconds. The system believes they ended successfully. It's as if I ran a pipeline with no processes. For example, here's the execution log for one of the latest (Apr 1):
Here's the output from the last time is successfully ran (Mar 30):
The compute environment ID being used for these is
3dQIgg5Q6qzNfD9pdgmwjX
.The only difference I'm noticing is the Nextflow version jump and that the newer one - in the output log - believes it's using DSL 2, which the
.nf
file does not use or specify.