Reloader

Reloader copied to clipboard

I have deployed the reloader and using and auto-configuration in deployment.yaml but when i am editing any value or secret it's showing me the above error. also, i would like to bring to your notice that i have 2 namespace eg, develop and runner and i have deployed your application in develop namespace so if i have configmap in develop namespace reloader is working fine but if i have configmap in runner namespace its not updating anything .. and giving the below error.

full Error:

time="2020-06-16T13:20:52Z" level=error msg="Update for 'abc' of type 'Deployment' in namespace 'develop' failed with error Operation cannot be fulfilled on deployments.apps "abc": the object has been modified; please apply your changes to the latest version and try again"

env : GCP K8 cluster

I'm seeing the same behaviour. When you bounce the reloader pod, new instance will look for outdated pods and fire a restart signal on outdated deployments.

Reloader should re-try on failure and shouldn't leave anything in outdated state.

Reloader doesn’t store any of the events to retry as of now, it works as a controller loop that watches the events as these occurs and acts upon them. The concurrent modification error occurs usually when more than one user/system is trying to update the resource at the same time. So it can happen when more than one reloader replicas are running or another user/system is trying to change the resource at the same time. Currently, reloader doesn't support the retrying of failed restarts but any PR and contribution is always welcome :)

FYI, We have the same problem with Kubernetes 1.23 (EKS) and the combo FluxCD + SealedSecrets. Sometime pods are not restarted with this same log.