Hi,
I’m using Azure scaleset and Octopus Deploy to provide scalable solution for auto scaling web front layer.
This project is pretty simple. Each machine in scaleset is described using ARM template with Octopus Deployment extension added.
After few tests I’ve encountered an issue with the deployments not beeing triggered on newly created machines.
This works fine when only 1 machine is created during auto scaling out. Then Octopus Server detects new deployment target and triggers the deployment.
However when multiple machines are being created by azure scaleset, Octopus Server does not trigger deployments to all of them.
This case is illustrated on the picture below. Machines 001Z and 001X are correctly detected and application deployment is triggered. But machines 001Y and 0020 are detected by Octopus Server are healthy but the deployment is not triggered.
I’ve looked into the problem and I suspect that you have hit this bug which matches the symptoms that you are reporting. If you are able to confirm if you are in the affected version range and if the logging message match that would be greatly appreciated.
I’ve been having a think about your scenario as well to see if there is anyway to workaround this issue, and unfortunately nothing is coming to mind as you can’t control when the machines get scaled out. I’ve checked and while we haven’t started work on this issue yet it is high in the backlog so we should start work on it in the not too distant future.
That does look like the probably issue then. I’ve checked in with the engineering team and this bug is very high on our backlog so we should start working on it shortly.
