Open bobzoller opened 3 months ago
tl;dr; I think if you're having this issue it's not a problem with the argo helm chart
Long version:
Yesterday I had similar churn on this pod while using the community chart 7.4.3
.
I didn't realise it at the time but I had another cluster issue which was hiding the logs of the redis secret init pod, which might have been complicating things further.
Before I fixed my cluster (which I did with a combination of a minor update and rotating all the worker nodes), I made this work by using the argocli to generate the secret before installing the chart, and installing the chart with redisSecretInit.enabled: false
which seemed to make everything work. My hypothesis is that there was some issue creating that secret due to my cluster being broken, but pre-configuring it meant the helm chart could continue installing argo.
Since I fixed my cluster I'm not able to reproduce this issue, even after purging all argo resouces, the secret init container seems to work fine.
I think the secret init churn is a sign of other problems, rather than it being the cause, but like I said I didn't have logs at the time, and I can't reproduce the issue now that I have logs.
I expect the above will be a suitable workaround for anyone who has this issues, but I would once again caution that this seems to by a symptom of a problem, rather than the problem itself. For anyone who doesn't want to use the argo cli, I expect you can just kubectl apply the following yaml
apiVersion: v1
data:
auth: UjlpdVcyYktYaEFDdTcyUw== # <- replace this
kind: Secret
metadata:
name: argocd-redis
namespace: argocd
type: Opaque
I have tried to use helm chart version 7.5.2 with the below still the same issue persists: redisSecretInit.enabled: false
https://github.com/argoproj/argo-helm/pull/2928#issuecomment-2361789841
redisSecretInit: enabled: false
With this it will work. Thanks !
This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
Describe the bug
I'm seeing unexpected churn related to
argo-cd-argocd-redis-secret-init
(job, role, rolebinding, serviceaccount) each time we run ahelmwave diff
with the argo-cd helm chart. This was happening in version 7.3.6 and is still happening in 7.3.11. helmwave version 0.36.3.(I realize this could be a helmwave problem, or this could be exposing a helm problem? I'm a bit of a newb at both unfortunately, and I thought I'd file here first because y'all probably understand what the issue might be even if it's not an actual bug in the argo-cd helm chart itself... apologies in advance.)
I'm using the HA mode with autoscaling example.
relevant output:
Related helm chart
argo-cd
Helm chart version
7.3.11
To Reproduce
Expected behavior
no diff
Screenshots
No response
Additional context
No response