Azure Service Fabric 2018.05.02

Hi,

We’re attempting to set up an Azure Service Fabric deploy with the latest release and the changes in the Azure infrastructure handling. We seem to have gotten a connection to the Cluster and it shows up as healthy (using certificate based authentication), however the moment we create a release and try to deploy a package there using the “Deploy a Service Fabric App”, it blows up with “Object reference not set to an instance of an object.”. The only place I’ve found that in googling related to Octopus has to do with packages. However I spun up a step to just deploy so I could validate the package I was referencing could be used, and that worked.

Your documentation still appears to reference the old way of doing things, and I’m not quite clear any more on whether or not I have to install the certificate on the local Octopus server or not. Since it was referenced in the connection I was hoping I wouldn’t (but I installed the cert on the server anyways). Still no luck. And the cluster still shows healthy in the Infrastructure section. I did validate that I could hit the port we are using on the cluster from this machine (so it’s definitely not a firewall issue either, unless there is something else under the covers going on).

Octopus server logs don’t show much.

----- The problem with typing up this while troubleshooting ----
We made modifications
We downgraded the SF SDK. We’ve had other folks in the company run into issues with 3.1.274.9494.
We upgraded to PS 5.0 (The step template mentions WMF3.0 not 5.0 FWIW, but one of the blog posts mentions 5.0)
DotNet we were already fine on (4.7.x)

Turns out none of that helped though. Even after a reboot, necessitated by the WMF install, still get the same error

Hi,

Thanks for getting in touch. Sorry to see you have had some trouble with the new Azure Service Fabric target.

I have been unable to recreate the issue you are seeing, would you be able to provide some further information such as the full task log and details about your deployment such as process (you can download the deployment process as a JSON file in the overflow menu on the process overview).

Regards
Ben

Not a whole lot there.

I validated that the package exists. I looked through the Octopus logs. I looked through Event Viewer logs, not much information there to give me an indication of what the problem is. Other things I tried because of things I read online. (Package Cache issues)

ServerTasks-48808.log.txt (1.01 KB)

visualobjects-process.json (5.04 KB)

Hi,

Thanks for the additional information.

You should not need to manually install the SF certificate on the server anymore. As long as you select the certificate from your Octopus library, we will automatically install it. If your target has successfully completed a health check, then that certificate has all been installed successfully.

While we were not able to reproduce, we did find the source of a potential null reference error which was fixed in version 2018.5.6. Could you please upgrade to the latest version and let us know if the problem still occurs for you?

Cheers
Mark

So we did this on another server, and then on our production server and so far I have not been able to reproduce the problem that we saw on the original. So I’m not quite sure what is going on here.

-Mike