Migrating to a new hypervisor environment and having problems getting one of the packages install correctly. For some reason, Pre- and PostDeploy.ps1 doesn’t get picked up by the tentacle, for a specific pacakge, on the new environment. Works perfectly on the old environment (see attached screenshot). I first I thought it might have something to do with Powershell 3.0 being installed. However I have a different package, where Pre- and PostDeploy.ps1 gets picked up perfectly (See second screenshot).
- Windows 2008 Server r2, on both systems.
- Powershell 3.0 initially installed on the new system but uninstall again.
- They use the same package and PreDeploy.ps1 and PostDeploy.ps1 exists
- Octopus Deploy 184.108.40.2069
- One package seems to fails to register the files (DMS.KvkService) and another works correctly (DMS.Client)
I can e-mail a log file if needed.
Any information will be greatly appreciated!
Thanks for getting in touch and sorry you’re having this issue. Would it be possible to get a full deployment log?
You can email them to me directly (paul at octopusdeploy.com) if you prefer to keep them confidential.
Digged a little further and the new machines use v.2.4.5 of the tentacle, while the old ones use 2.3.3. Server version is 2.3.3. I downloaded the old 2.3.3 agent and installed it on the new host, and now the scripts are picked up…
So the easy answer would be just to upgrade to the newest version Octopus Server. Unfortunately, due to the new strategy regarding root only *deploy.ps1, we’re kind of pinned and cannot upgrade the 2.4 - got 48 package maintained by various teams and they need to be fixed, redeployed to production, so we have a rollback option on 2.4 and before we can upgrade. Bad (and breaking) design choice from our point of view…
So, I guess the agents are not backwards compatible and we cannot use the newest version of the tentacle with 2.3.3, correct?
Thanks for getting back to us about this!
You are correct, the Tentacle version needs to match the Octopus version. And I am glad that you found a way to get your scripts working again.
We are really sorry that the root deploy scripts was a breaking change for you and we do not have any easy solutions or work-arounds to help your situation.
I don’t know if you have seen the blog post about why we did it or not but here is a link: http://octopusdeploy.com/blog/rfc-root-most-scripts
Again really sorry it has had such a large impact on your deployments.
Yeah, I saw the blog post - We use Octopus in large scale and have almost 300 unique packages. So breaking changes like this is costly, from our point of view. Anyways, thanks for responding - I’ll close this issue