-
Notifications
You must be signed in to change notification settings - Fork 344
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
Error applying changes in ttlSecondsAfterFinished
with 1.13
#494
Comments
Do you think this deserves a patch release? |
@kimxogus which Kubernetes version are you using? |
@jpkrohling Neither EKS v1.12, nor EKS v1.13 have the needed feature gate active and there is no way for end-users to flip the switch, as it is a managed service offering. Besides, See also: https://kubernetes.io/docs/concepts/workloads/controllers/ttlafterfinished/ |
Thanks for confirming, @headcr4sh. Would any of you be able to provide a patch? We are currently working on another issue at the moment, and we might need to do a patch release once that issue is fixed. If we have a PR for this one here, we'll include it in the patch release. |
I am also facing the similar issue : Operator version - 1.13.0 Operator Logs :
Spec :
This is working fine with |
@jpkrohling As far as I can tell, the following patch brought in the changes that cause trouble: A viable solution might be to set |
Possible solution (untested, because I am not [yet] able to compile/run the jaeger-operator myself): Replace all occurences of
with
|
A similar change would be made anyway as part of #466, so, I'd be OK with a change like that. I just pushed a Docker image with this fix and I'd appreciate if one of you could confirm that it works: Image: In the meantime, I'm writing the required tests and will send a PR soon. |
For the record, this works with edit: "this works" means that no such failure is seen, but I'm not sure the feature itself is working (ie, that batch job pods are cleaned up after N seconds) |
Unfortunately, I can confirm that the Docker image you provided does still lead to the exact same errors:
Too bad,... :-( |
Are you using the same Jaeger instance that you had before? |
I tried both a new and an existing Jaeger CR manifest. Both lead to the same error, I am afraid. |
Alright, I'll try to get hold of an AWS account to reproduce the problem and come up with a patch. |
I am pretty sure, that minikube with the correct feature gates could also be used to reproduce the situation: Just adding |
Looks like the I just pushed a new image making the property noop. Once one of you confirm that this fixes the problem, I'll send a PR and get it reviewed/merged. Image: |
After upgrading jaeger operator
1.13
, error below occurredIt seems
ttlSecondsAfterFinished
should not be set by default, otherwiseTTLAfterFinished
feature gate must be enabled in cluster.The text was updated successfully, but these errors were encountered: