Search code examples
ibm-clouddevops-servicesactive-deploy

Active Deploy `begin` step fails after upgrade to devops toolchain

We recently upgrade our IBM Bluemix devops project to a toolchain as recommended by IBM and it doesn't deploy anymore. The pipeline configuration seems to have migrated over correctly, and the first step of the process deploy process even works, creating a new instance of the app. However when it gets to the active-deploy-begin step it fails with the error:

--- ERROR: Unknown status: 
--- ERROR: label: my-app_220-to-my-app_2 space: my-space routes: my-app.mybluemix.net 
phase: rampup start group: my-app_220 app (1) successor group: my-app_2 app (1) algorithm: rb 
deployment id: 84630da7-8663-466a-bb99-e02d2eb17a90 transition type: manual 
rampup duration: 4% of 2m test duration: 1s 
rampdown duration: 2m status: in_progress status messages: <none>

It appears to have started the build number from 1 instead of continuing from the previous number of 220. I've tried deleting the service at the app level from the Bluemix web interface to no avail. Any help or pointers will be much appreciated.

UPDATE:

Things I've tried:

  1. Deleting the app and running the build process to create a new instance. This worked the first time as it detected it was just the initial build. But then the second time it ran it failed with the same Unknown Status error.
  2. Deleting all the previous deployment records in the to eliminate the possibility that it was caused due to a deployment label name conflict. i.e. my-app_1-to-my-app_2

Also interestingly the active deploy command works from the cf command line using the active-deploy-create my-app_1 my-app_2 command. So it seems that the issue might be with the script that runs the active deploy commands for the pipeline.

Solution

  • This issue was reported also at https://github.com/Osthanes/update_service/issues/54. There you will find instructions how to get the issue fixed.