-
Notifications
You must be signed in to change notification settings - Fork 905
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Old replica sets not always scaling down after update from 1.6.6 to 1.7.1 #3762
Labels
bug
Something isn't working
Comments
Yup that is an issue it should be |
agaudreault
added a commit
to agaudreault/argo-rollouts
that referenced
this issue
Aug 12, 2024
6 tasks
zachaller
pushed a commit
that referenced
this issue
Aug 12, 2024
zachaller
pushed a commit
that referenced
this issue
Aug 13, 2024
meeech
pushed a commit
to CircleCI-Public/argo-rollouts
that referenced
this issue
Feb 10, 2025
…j#3762) (argoproj#3784) dco Signed-off-by: Alexandre Gaudreault <alexandre_gaudreault@intuit.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Describe the bug
After upgrading from
1.6.6
to1.7.1
we've noticed that occasionally when deploying and promoting releases the old ReplicaSet does not get scaled down. The old pods keep running indefinitely. Though they do not receive any traffic as active/preview services correctly point to a new ReplicaSetTo Reproduce
The issue is hard to replicate since it happens sporadically.
Version
1.7.1
Logs
controller-deployment.log
It looks like the controller sets
scale-down-deadline
annotation correctly to 30 seconds forward. Few lines below the ReplicaSet gets patched without changingscale-down-deadline
annotation. But then after few more lines it makes another patch operation withscale-down-deadline
set to an empty string, which I think is the problem.From this point of time, the controller keeps setting the deadline date and resetting it to the empty string.
I had a look at the recent changes to see what might cause this to happen. There was a new function added that is responsible for patching the replica set -
updateReplicaSetFallbackToPatch
I suspect the problem lies on the following line
argo-rollouts/rollout/controller.go
Line 990 in 708db68
It copies the annotation from
rs.Labels
instead ofrs.Annotations
Though, it's not clear to me what is the exact flow that triggers this issue intermittently.
Message from the maintainers:
Impacted by this bug? Give it a 👍. We prioritize the issues with the most 👍.
The text was updated successfully, but these errors were encountered: