Deployment slow/never completing

Hi,

Our deployments typically take 4-5 minutes from Octopus Deploy Cloud. We use a docker image as the for the deployment. However this morning, it’s taken more than 30 minutes to pull the image to worker.

I’ve stopped and restarted the deploy, the same worker was leased. Perhaps it’s an issue with that node?
Leased worker 21-10-01-0342-rz1ok from pool Default Worker Pool (lease Leases-2836078).

Any help would be greatly appreciated as it’s holding up our CI/CD pipeline.

Update: It took 44 minutes to pull the docker image, and then happily handled the deployment in 4 minutes.

The next deployment picked up the same worker and as it had the docker image, it’s jumped right on to the deployment.

Hi Simon,

Thanks for getting in touch!

I have come across this recently too, typically a Worker will cache these images once they’ve been downloaded once and then just update them as needed during deployments.
I’m going to check with our Cloud engineers to see if we’ve made any changes to the Workers, such as recycling them more often, that could explain the behaviour.

Out of curiosity, is the docker image you’re using one of our worker-tools images? And if so, was it the windows or linux one?

Regards,
Paul

Hi Paul,

Yes, it is one of the worker-tools images, the windows one:
octopusdeploy/worker-tools:3.0.0-windows.ltsc2019

However, rather than pulling it from Dockerhub, I have copied it into a private AWS repo as I have seen Dockerhub report the images as unavailable at times, and so wanted to eliminate this dependence.

Regards, Simon

I figured as much.

I did a test pulling the image directly from Docker outside of Octopus and it did take around the same amount of time.

Depending on what steps you’re running on the image, using the Ubuntu one might reduce the time it takes, as the image is a quarter of the size.