che
che copied to clipboard
eclipse-che
Devworkspace Operator keeps changing public ports
Describe the bug
If creating a devworkspace from a devfile that has more than 3 public endpoints ingress hostnames keep changing for quite a while.
This will cause endpoints displayed by checode/theia to point to the wrong service. It seems that after a while this is stopped. I haven't figgered out the exact time it takes to settle down.
Che version
next (development version)
Steps to reproduce
I've created a demo repository that causes the issue and has 5 public endpoints using grafana as a service.
https://github.com/nils-mosbach/devfile-io-changing-ports
Devworkspace Operator logs can be found here: https://github.com/nils-mosbach/devfile-io-changing-ports/blob/main/devworkspace-operator-logs.json
Expected behavior
Ingress endpoints should not change after they are created.
Runtime
Kubernetes (vanilla)
Screenshots
No response
Installation method
chectl/next
Environment
Linux
Eclipse Che Logs
# Che Dashboard Logging
Initializing DevWorkspace
Preparing networking
Preparing services
Preparing ingresses
Networking ready
Preparing ingresses
Networking ready
Waiting for DevWorkspace ServiceAccount
Preparing ingresses
Waiting for workspace deployment
Preparing ingresses
Networking ready
Waiting for workspace deployment
Networking ready
Waiting for workspace deployment
Preparing ingresses
Networking ready
Waiting for workspace deployment
Networking ready
Waiting for workspace deployment
Preparing ingresses
Networking ready
Waiting for workspace deployment
Preparing ingresses
Networking ready
Waiting for workspace deployment
Preparing ingresses
Waiting for workspace deployment
Preparing ingresses
Waiting for workspace deployment
Preparing ingresses
Waiting for workspace deployment
Preparing ingresses
Waiting for workspace deployment
Preparing ingresses
Networking ready
Waiting for workspace deployment
Preparing ingresses
Networking ready
Waiting for workspace deployment
Preparing ingresses
Additional context
No response
I was able to reproduce this with only two public endpoints as well. It seems to settle down after a couple of seconds, but the editor port extension still catches one of the invalid states and therefore points to the wrong service. Specially on Che-Code which loads quite fast.
Reloading the editor at least fixes wrong endpoint-service-mappings.
Issues go stale after 180 days of inactivity. lifecycle/stale issues rot after an additional 7 days of inactivity and eventually close.
Mark the issue as fresh with /remove-lifecycle stale in a new comment.
If this issue is safe to close now please do so.
Moderators: Add lifecycle/frozen label to avoid stale mode.
