Issue Summary:

When deploying a pipeline on Spinnaker 1.15.2., the "reset source to original size" step never seems to complete. On five separate runs, it stalled at the "wait for capacity match" task. (Each time, we aborted the job after 25 minutes.)

When deploying the exact same pipeline on Spinnaker 1.14.14, the same task is virtually instantaneous, taking less than a second.

We suspect there may be a regression with the Clouddriver module.

Cloud Provider(s):

AWS, Kubernetes, Cloud Foundry

Environment:

Spinnaker itself is deployed within an Amazon Web Services Kubernetes cluster; however when deploying the desired artifact itself, we're pushing to open sourced Cloud Foundry.

Feature Area (if this issue is UI/UX related, please tag @spinnaker/ui-ux-team):

Pipelines

Description:

Error as it's shown in v. 1.15.2:

Success as it's shown in v. 1.14.14:

Steps to Reproduce:

1. Execute any pipeline, confirm success.
2. Upgrade to 1.15.2. Execute any pipeline, note failure. (All pipelines stall on the same task.)
3. Via Halyard, create new deployment:

• copy existing configuration file
• set version to 1.14.14
• update override URLs in the security section
• update location within the deployment section to point to a different Kubernetes namespace
• retain all other values

4. Within new deployment, confirm Spinnaker loads pipelines from shared persistent storage.
5. Execute any pipeline, confirm success.