Brad Davidson

We're planning to write a tool to move the pending releases into failed state, and then trigger an upgrade or reinstall of the chart. RKE2 currently runs into this most...

Just run kine (either via the binary or the docker image), and then point your Kubernetes distro of choice at it as the etcd datastore. Or use k3s, which already...

Hmm, yeah that is an interesting question. I suspect that perhaps the current RBAC is set up such that SA has access to all namespaces, instead of just the one...

K3s embeds etcd, the core Kubernetes control-plane components, and several additional controllers (including the one from this project) in the main k3s binary.

that is correct. On recent releases, you can start k3s with `--disable-helm-controller` if you want to turn it off.

Implementing storage.Interface directly would mean that we would have to build modified versions of the apiserver to support Kine. By implementing a subset of etcd v3 GRPCs, you can point...

As the project readme states, kine currently only supports dqlite, sqlite, Postgres, MySQL and NATS JetStream. dqlite support is deprecated. Porting it to redis would be difficult, as redis does...

I think it wouldn't perform very well if you could get it working at all, but I certainly encourage you to give it a try. I think redis only supports...

Kine assumes that there will be no gaps in the auto-increment sequence - that is to say, they are strictly sequential. Any clustering configuration that violates that constraint will most...

This is a design constraint due to how we use the auto-increment pk as the MVCC revision counter. If you need an active-active backend you're probably better off going with...