Are there a way to achieve something similar to maxSurge rolling update but for daemonsets? I’d like at least one pod for each node available during an update.
Currently using Metallb with local traffic policy together with Calico, whenever I do an update there will be downtime.
Would be great if the daemonset could run multiple pods on each node, support maxSurge rollout or if I can configure two daemonsets but that doesn’t roll its pod on the same node as the other daemonset.
Cluster information:
Kubernetes version: 1.19
Cloud being used: bare-metal