-
Notifications
You must be signed in to change notification settings - Fork 453
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
How to control a huge vcluster #2396
Comments
vCluster can be deployed in HA configuration (docs), and you can change the resource requests and limits. Service is already used for the API traffic. |
@matskiv why As I said, at the beginning, one replica for control plane is enough, with time goes, how to change to multiple replica? That would be a great feature if |
vcluster project itself is about running a virtual kubernetes cluste, for managing a fleet of virtual clusters there is vCluster Platform or cluster-api-provider-vcluster
That is supported, just not with the SQLite as the backing store. Docs page about HA covers that. |
Here the doc says that when deploy, we can set multiple replicas, how about automatic expansion? I do not see those info, do you mean at the beginning, set 1 replica? While I do not find how to add replicase, could you help me? |
As I said above - |
Just double confirm, does it work if I set Or no need to set, I can update replicas as |
The replicas need to be changed in the configuration ( |
OK, I will test functions if I set |
Is your feature request related to a problem?
Thinking about this case, one user create a vcluster, as time goes by, the size of cluster becomes larger and larger.
In this time, this vcluster is very close to a true k8s cluster, while it is still controlled in one POD, no duplication can improve QoS, no HA. Now, we need a true control k8s plane which support expansion and HA.
Which solution do you suggest?
Support control plane split into multiple PODs, use service to do balancing, then each component can Dynamic expanse and contract.
Since we use SQLite as backend store, the split process(expanse) or unite process(contract) do not impact workload.
Which alternative solutions exist?
I only know change the codes by hand.
Additional context
No response
The text was updated successfully, but these errors were encountered: