Request for FAQs about running FoundationDB in production

We are planning on a big push to extend the documentation on running FDB in production over the next few months, and we would like to get some feedback from you all about areas that you think we should add or extend in this documentation.

It would be nice to cover ability to run in containerized environments such as Kubernetes and issues around hostnames (Proposal: Don't identify a coordinator based on its IP). I don’t think it’s clear whether there are workarounds (host networking) or general feasibility / advice.

I think that having good documentation around monitoring FDB clusters would be very helpful:

  • Description for fields in status json
  • Structure and interpretation of different kind of events in fdb logs
  • Types of files that FDB creates (with explanation for those).

One of the first things we do is to migrate a lot of data to fdb. Only after that we try to optimize cluster based on production conditions (regions/read vs write traffic etc).

Based on our current understanding, a certain cluster configuration is good for faster migration of data and another configuration of cluster is good for production.

It would be great if we can have a FAQ for “cluster setup for loading bulk data”

A FAQ on “How to perform zero/minimum downtime upgrades?” would be really helpful.

2 Likes

Yeah, we can definitely provide some workarounds and general guidance, while we work to make it more seamless out of the box.

These are all great suggestions!