Recent searches
Search options
Incredibly stupid question potentially, but does Kubernetes have sort of a "look at current Pods and potentially reschedule on other nodes" trigger? I've just had the situation that after a sequential node restart, I had two nodes completely full, up to having some Pods pending, while the last node to be rebooted still had CPU free. But to make use of that, some Pods on the full nodes would need to be rescheduled.
1/2
I know that I could manually delete some Pods from the full nodes and k8s would likely reschedule them on the open node, but that would be a manual process.
So my question: Is there a way to trigger k8s to look at all pods and potentially reschedule some of them, potentially including moving pods from one node to another?
2/2
@badnetmask@hachyderm.io@mmeier @jana @taavi
I read about it here once. I hope this helps adding more colors.
https://learnk8s.news/@learnk8s/112094754208296603
@badnetmask @jana @taavi Yupp, it did give a nice overview - definitely looks like this does exactly what I need, the ability to rebalance pods after a node restart.