Enable TLS on consul servers and drop WAN retry join #548
Reference in New Issue
Block a user
Delete Branch "benvin/consul-server-tls-fed"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Consul servers need TLS and a gRPC TLS port before mesh gateways can federate the datacenters. The WAN retry join is dropped ahead of federation; drw1 has no servers yet, so nothing depends on it.
tlsparameter to profiles::consul::server and set CA, cert and key from the vault PKI filesgrpc_tlson 8503auto_reload_configso cert rotation reloads without a restartjoin_remote_regions: falsein syd1 and drw1 to stop WAN retry jointls.defaults.cert_fileis inherited bytls.grpc; withports.grpc: 8502and noports.grpc_tls, Consul >=1.14 refuses to start ("theports.grpclistener no longer supports TLS ...ports.grpc_tlsis not defined"), so every server dies on its next puppet run → addgrpc_tls: 8503toprofiles::consul::server::ports(plaintext 8502 stays as is).enable_mesh_gateway_wan_federationis hardcoded for every region, but drw1 hiera keepsjoin_remote_regions: true, renderingretry_join_wan, which Consul rejects alongside WAN fed ("'retry_join_wan' is incompatible") → setjoin_remote_regions: falsein hieradata/country/au/region/drw1/infra/storage/consul.yaml or make the WAN-fed flag a param enabled only for syd1.join_remote_regions: falsealso set in syd1 (hieradata/country/au/region/syd1/infra/storage/consul.yaml:5), andauto_reload_config => trueadded (site/profiles/manifests/consul/server.pp:117) → list both, with why (syd1: retry_join_wan is rejected with wanfed; auto_reload: picks up the 90-day cert renewals without a restart)enable_mesh_gateway_wan_federationswitches cross-DC RPC/WAN gossip to gateway-only, but no mesh gateway exists until #549; syd1→drw1 traffic fails (ErrDCNotAvailable) from the first server restart → move the flag into #549 (after gateways are registered) or merge/roll #548+#549 together and say so in the rollout steps.Enable TLS and mesh gateway WAN federation on consul serversto Enable TLS on consul servers and drop WAN retry joinNo findings.
Pull request closed