Invalid enum value 'FailedStartingRole' cannot be deserialized...

Mysterious failures in Azure...

We ran into a problem yesterday when deploying our MVC API to Azure. At the time, we had code running in the Production slot and a deployment stopped in the Staging slot. When we kicked off a new build, set to deploy to the Staging slot, it failed with the following error:

A quick Google search brought up this Stackoverflow question. Our previous deployments weren't faulty but, nonetheless, we decided to try deleting the existing Staging deployment and kicking off a new build. I can't explain why, but this fixed the problem. Build succeeded and the deployment moved on.

EDIT: This happened again recently, and simply deleting the staging deployment did not fix the issue. At the time, there were a few unhealthy instances on a cloud service in Azure in the production slot. For whatever reason, if there are any unhealthy instances in production or staging, the deployment seems to get blocked. There are three options with varying degrees of impact. First, you could make a change to a config in the Azure management portal, which sometimes seems to get unhealthy instances up and running again. Second, you could scale down the production cloud service web/worker role until the unhealthy instances are removed. Third, you could delete the production and staging deployments.

Once production and staging are healthy or empty, a new deployment should work just fine.


portrait title

About Scott

I am a software engineer from Bozeman, MT enjoying the slightly warmer climate of Colorado. I think code can change lives. I think lives are worth changing. I write code.

You can find me on Twitter, LinkedIn, , and Github.