I'm working with Longhorn for managing storage across multiple zones. In my setup, I have two main zones for storage nodes, along with a smaller zone dedicated to a tiebreaker monitor for Rook Ceph. I want to know if Longhorn operates similarly, especially in terms of handling scenarios where I might lose half of my worker nodes. Specifically, if I have 4 replicas and only 2 are available, will my volume still be writable?
3 Answers
There's a difference between regular storage volumes that can be spread across all workers and logically replicated volumes that you might limit for performance. As long as at least one copy of a logical volume is available, it remains writable, but you'll get a warning that it’s operating in a degraded state.
It should still be writable! But just to be sure, you might want to test it out. Try running Longhorn and shutting down or destroying 2 out of your 3 nodes that hold the volume replicas, especially under heavy load. From my tests, the applications managed to keep reading and writing data to the volumes without issues.
Longhorn doesn't have a built-in tiebreaker like Rook Ceph. For it to keep the volume writable, you need a quorum, which means more than half of your replicas must be online. So, if you've only got 2 out of 4 replicas available, the volume won't stay writable. It's good to plan for redundancy to avoid data loss!
Related Questions
Can't Load PhpMyadmin On After Server Update
Redirect www to non-www in Apache Conf
How To Check If Your SSL Cert Is SHA 1
Windows TrackPad Gestures