sgissi
commented
1 year ago Hi @samos667, thanks for raising this. Indeed the initial timeout is fixed, I wasn't expecting a problem there as Vaultwarden is pretty lightweight. I'll take a look at adding an option to customize it across all charts.
Keep in mind that a warning event can show up on a temporary problem, just keep an eye on the "Count" and "Last seen". If it keeps happening and blocks the Pod from running it will be a problem. For example: if a PVC is unbounded because storage took a little longer to create the PV, you might get the "has unbound immediate PVC" a couple times but as soon as it completes, the deployment will continue. If the storage has an underlying issue (e.g. ran out of space to create a PV), the event count will keep increasing and the Pod never actually starts.
I'll update here once I get a chance to add the custom delay option.
samos667
Hi first thanks for your amazing chart for vaultwarden. I have 2 minors errors when i have deployed it with fluxcd on 1.24.x kubernetes cluster. I said minor because this not impact the service. The service was immedialty working on the initial deployment. But two event appair on my k8s dashboard Lens in the vaultwarden namespace:
One is about readiness failure on port 8080 of vaultwarden pod:
I have fix that on editing vaultwarden deployment and add delay for readiness, like this:
The second one is about a "ghost" PVC unbound, i didn't really understand:
The fix was to recreate pod by deleting or by editing the deployment(like fix readiness) for force recreate.
For the first I thing is possible to edit chart template for add inital delay for readiness. But for the second one i dont know at all.