Is there any documentation available for clustering in v2 to achieve HA?
They've said that they're moving HA support to their enterprise edition. They removed the storeconfig script that helped load a config into a KV store, but they've also stated that they'll have support for KV stores again. I went from using consul and a distributed setup in 1.0 to using static configs for 2.0 because of this. Now I'm waiting for them to re-enable some KV providers.
Clustering is still possible in 2.0 if you use the same static config on each pod, so... Technically it's possible?
I was thinking this over and I realized I should be a bit more specific. You can set up clustering and still use the dynamic providers in traefik 2.0, you just need to configure each instance of traefik in the same way and provide it with some certs. Keep in mind that you'll hit your limit with letsencrypt if you have multiple instances constantly requesting tls certs. To get around this, I got wildcard cert for my domain manually and gave that to each of my traefik instances to use. I'm running on kubernetes, so I use the kubernetes ingressRoute provider. I defined entrypoints, log options and providers as arguments to the traefik docker image in my deployment.yaml, and created an IngressRoute.yaml for each of the services that I wanted exposed. I then started traefik up as a replicaset and it's working fine for the time being. Not exactly a thorough solution, but we needed tcp passthrough support and for us it's working for the time being.
I too am feeling the rub of the consul/etcd KV store provider absence in Traefik2.0. I have a bare-metal K8s cluster that spans two data-centers, with Traefik deployments pinned to specific worker nodes in each data-center by a nodeSelector, and a traefik service with a metalLB load balanced IP from its data-center directing traffic to its data-center region-specific traefik deployments. Previously, like you, I had my certificates stored in etcd KV store, but now when I load up my Kubernetes ingress-routes, each of the four traefik deployments bum rush the LE servers for certificates, and pow - I've hit cert request cap. I've worked around it for now by manually adding the LE wildcard cert as a secret, but it's a real PITA if I need to do this every 90 days. First world IT problems, amirite?
I was planning to do the same via Docker swarm, would sharing the folder of "letsencrtypt" i.e the certificates with the pods would? (all other configs for traefik being the same)