Hi,
I have gone through your documentation about Service Fabric and I was trying to setup OnPrem Service Fabric deployment. In the process I discovered that both Service Fabric step templates available in Octopus (running version 2018.5.6) are not relevant to me because they work only when deploying to Azure cluster.
Do you have any guidance for deploying to standalone non Azure server?
PI plan to use custom PowerShell scripting, however I thought they may be a better way…
Many Thanks
Emil
Hi Emil,
The step name and our documentation may be a bit misleading in this instance as it should be possible to use the service fabric steps with an on-prem cluster since our steps don’t know any of the specifics related to Azure. The only authentication mode we don’t support at this point in time is active directory.
The documentation would still be relevant to some extent. You would need to create your target and set up the connection details and choose the security mode and you should be good to go. We don’t have official guidance for how to get things such as the certificates setup for on-premise so you would, unfortunately, be on your own in that regard.
Please let us know if there is any way we can help though.
Regards,
Shaun Marx
Hi Shaun,
Yes it is misleading indeed, however I did try running those steps before I posted my original message and they both failed with an error message:
‘The machine HQSERVER0001 will not be deployed to because it is not an Azure Service Fabric Cluster target.’
and I assumed they are Azure specific.
Anyway my question now is: How does the SF deployment gets executed with those steps? I think I read no tentacle is needed - instead you need the Service Fabric SDK installed on the Octopus Server and the deployment is handled via Service Fabric…
I don’t think this will work for us, because our target server is behind a very tight firewall and that’s why we love Octopus with the tentacle model. I guess you don’t have any documentation on using the tentacle deployment model with Service Fabric?
Hi Emil,
Do you get that error message given the role maps to your configured service fabric target?
Yes, that is correct. No tentacle is required at this point as all of our steps run on the server. The workers feature may be of interest to you as that would allow these steps to run on other machines such as those in restricted environments. We are currently rolling this feature out in phases where the most recent 2018.6.0 release added support for a single external worker and worker pool. It may still not suite your needs at this point as the single external worker limitation would mean that all tasks would execute on that specific worker. Once we lift this limitation you will be able to choose where these tasks run on a case by case basis.
Unfortunately using a tentacle for deployment with Service Fabric would not be possible until we fully release external workers.
Regards,
Shaun
HI Shaun,
Yes, I get this error given the role maps to my configured service fabric target!
I have tried this with unsecured service fabric cluster and I was hoping to work, but this error doesn’t mean much…? There is lack of detailed documentation about this scenario and I am stuck - please help!
I have also setup secure service fabric cluster using self signed certificates and I cannot work out how to use the Service Fabric steps because there is nothing about client certificates…?
Emil,
Please see my reply to your other post.
A Service Fabric target will need to be created and this is where you would configure the certificate.
Regards
Ben