Is your feature request related to a problem? Please describe.
As a follow on to #1286, and in addition to #1287, once we have the GKE cluster and node pools, we will need to spin up the kubernetes resources to compose the testnet. A core part of this will be the penumbra fullnode(s). This node will participate in the p2p network with the validators but will not participate in consensus. It will additionally expose the interfaces to a public domain such as tendermint RPC and penumbra GRPC.
Describe the solution you'd like
Use a kubernetes Deployment or ReplicationController that will have a daemon container for pd and a second container for tendermint.
Use initContainers to handle startup tasks such as disk ownership and config population.
Two PVCs, one for the pd persistence (rocksdb) and one for tendermint persistence.
Kubernetes ingress for the exposed services (RPC and GRPC), using google managed certificates and reserved IPs on the load balancers. This will make the fullnode, or cluster of fullnodes, accessible behind a load balancer on a domain such as https://rpc.testnet.penumbra.zone.
Describe alternatives you've considered
N/A
Additional context
This is the strategy Strangelove uses to host fullnode infrastructure.
Is your feature request related to a problem? Please describe. As a follow on to #1286, and in addition to #1287, once we have the GKE cluster and node pools, we will need to spin up the kubernetes resources to compose the testnet. A core part of this will be the penumbra fullnode(s). This node will participate in the p2p network with the validators but will not participate in consensus. It will additionally expose the interfaces to a public domain such as tendermint RPC and penumbra GRPC.
Describe the solution you'd like
Deployment
orReplicationController
that will have a daemon container forpd
and a second container fortendermint
.initContainers
to handle startup tasks such as disk ownership and config population.https://rpc.testnet.penumbra.zone
.Describe alternatives you've considered N/A
Additional context This is the strategy Strangelove uses to host fullnode infrastructure.