Aborting a pending deployment sequence in helm-service leads to failure until the aborted sequence finally finishes

[christian-kreuzberger-dtx]

This is with Keptn 0.9.2.

Current behaviour

I am deploying podtato-head and accidentally tried to deploy an image with a tag that does not exist.
I noticed my mistake, and aborted the delivery sequence:

However, when I triggered the next delivery sequence, I ended up getting the following error message:

Error when installing/upgrading chart pod-tato-head-dev-helloservice in namespace pod-tato-head-dev: another operation (install/upgrade/rollback) is in progress

This is to be expected, as helm-service is still working (or in other words, helm-service has not finished and sent out a deployment.finished event).

Eventually helm-service has finished in the aborted sequence, and I can deploy again.

Expected behaviour

Either helm-service or shipyard-controller need to react on the aborted sequence.

• When aborting, helm-service could abort the deployment process using helm cli and send out a deployment.finished event
• When aborting, shipyard-controller could wait until helm-service sends the deployment.finished CloudEvent (or times out) before starting the next sequence

[thisthat]

This ticket will be solved by adding this limitation to our documentation: Aborting a sequence does not aborts the running task

[bacherfl]

It seems like this behavior was already documented: https://keptn.sh/docs/0.10.x/reference/bridge/manage_sequences/#aborting-sequences