Fatal error acquiring package when no machines in Role

I think this is a bug.

I’m using Octopus Deploy 2.0.5.933, with tencales in listening mode. When I create and deploy a release using Octo.exe to an environment not containing any instances of a particular targeted role, I get an fatal error in the acquire packages stage - rather than just a warning.

The error relating to a package that has no associated machines in the role is:

“When comparing the hashes of the packages downloaded by each machine, one or more machines returned a different hash for the package. Deployment will not proceed since this would result in inconsistencies between the files installed on different machines.”

When using the UI it prevents me from getting into this scenario and marks the steps with missing machines in those roles as skipped. When triggering a deployment I would expect it to just ignore steps if there are no associated machines even if not set to skip (I beleive this is the behaviour in 1.0?).

Thanks,
Mike

Connecting to server…OK
Username: rasim6315
Authorizing…Failed!
Login error: Wrong password!

Here’s a copy of the log file indicating the problem. The steps without machines are flagged to be skipped, but it still proceeds with an attempt to acquire packages for those steps and throws a fatal error.

(Step_2 is the only one that requires Package_2, and is bound to the role NoMachinesInRole)

Log.txt (3 KB)

Thanks for the thorough report - yes, this is a bug. Issue added here: https://github.com/OctopusDeploy/Issues/issues/412

Regards,
Nick