This message was deleted.
# k3s
a
This message was deleted.
k
cc: @high-alligator-99144
c
If all three nodes were actually active in the etcd cluster, the loss of one node should be fine from a quorum perspective. Is it possible that you are short on resources, and the loss of that node has created capacity problems on the remaining nodes, and they are now overloaded?
Its hard to say with any certainty what’s going on without logs.