12 ms·

I believe that in newer k8s versions, running ‘cordon’ is redundant. This doc seems a little outdated.

• aflag 4y ago
  If one node has a problem and you don't want anything else scheduling there as you debug, what should you use instead?
  • dw2072 4y ago
    The article lays it out into two steps, with "Step 1" being to `cordon` and "Step 2" being to `drain`. My understanding was that `cordon` used to be required to (safely) drain a node, hence why so many tutorials still show it, but that on newer versions of K8s, draining a node will automatically cordon it as well. Of course, if you want to keep the node up and running, but evict the pods, then absolutely that is when `cordon` _should_ be used.
    • Pirate-of-SV 4y ago
      If only one node is being drained then yes. If you want to drain 3 nodes as fast as possible it's best to start by cordoning all 3 of them and after that run drain on each respectively. This will cause minimal unnecessary interruption of pods. What we want to avoid is that new pods start on nodes that we soon want to remove/drain.