Allow 1 restart per task

[matthdsm]

This should fix a concurrency issue with the CSI driver
cfr ceph/ceph-csi#3511 and hashicorp/nomad#15197

[abhi18av]

@matthdsm , would this trigger a nomad-native retry or is this expecting Nextflow to trigger this?

As of now, we have a model of 1 nextflow task -> 1 nomad job -> group -> task and therefore, error at nomad job level reflects error at nextflow level and the overall behaviour aligns with what Nextflow would expect.

How nomad-native retries would interact with the overall setup needs to be tested.

[jagedn]

If I'm not wrong, before to have this constants we're using the default value (3) and Nextflow wait correctly for the completion of a failed job, so maybe can be a good idea to use 1 as default

[abhi18av]

If I'm not wrong, before to have this constants we're using the default value (3) and Nextflow wait correctly for the completion of a failed job, so maybe can be a good idea to use 1 as default

It seems that 3 is the default for these attempts #82 (comment)

@jagedn , do you think its worth exposing the closure and including other fields like delay , interval etc?

[abhi18av]

Okay, merging this and tagging this realease as 0.2.0-edge3 to release a build

[jagedn]

@jagedn , do you think its worth exposing the closure and including other fields like delay , interval etc?

maybe we can implement all of them when required