We’ve upgraded to the recent version : 2018.6.12 and noticed that pre-defined steps for cloud service deployments no longer supports running the deployments on tentacles and forces it to be run through Octopus Server. Has this feature been removed intentionally or is this a bug?
Thanks for getting in touch. Could I just check which version you upgraded from? Cloud service deployments have run via server since 3.0, I believe.
Interestingly, with the introduction of Workers in 2018.7 the model has moved slightly back the other way again. A Worker lets you execute the deployment on a machine other than the server, but is setup slightly differently to a normal Tentacle.
I hope that helps and let me know if I can assist further.
We’ve upgraded from 2018.3.11. I’ve rechecked my history build to make sure my eyes did not deceive me and I’m positive this step has been executed at the tentacle level at that time
I’ve done some checking in the code and I think the 2 screenshots are actually both running on the server (the E:/Octopus/Work folder seems consistent). What I found is that a while back we changed the wording on that section banner, in preparation for Workers I believe.
The text you are seeing now is what you get when using the internal Worker. If you use an external Worker the text would change to Worker on behalf of... is my understanding.
You can use a Worker, as of 2018.7, if you do need to move the work off the Octopus server itself. Under the hood the Worker is actually a Tentacle, but the deployment runs once on a single Worker leased from a pool, rather than traditional Tentacle where it would try to execute on every Tentacle in an environment.
Hope that helps clear things up, if not please let me know.