1
[ my public key: https://keybase.io/alixaxel; my proof: https://keybase.io/alixaxel/sigs/lmoZ4-X5LFvrwZ9dviP6ZyRfdG_slBMQctzlYELH9OI ]
db866a
The only project that I know that's trying to address this is Flynn.io, they try to solve it with a state machine that keeps data replicated but even trivial mishaps (like the cluster restarting at the same time) will render the cluster permanently offline.
Are there are self hosted PaaS that are able to provide high-availability and failover support for the most common stateful services (databases and files)? Are the big guys (Kubernetes, OpenShift, Rancher) trying to address this somehow?