I have configured EKS cluster (1.20) with 5 nodes and deployed a lot of helm charts, some applications needed some shared storage like EFS so I have configured it as well and mounted to the specific folders using PV and PVC which has been provided by efs-csi-controller.

everything worked very well for more than 6 months, but in the last days I noticed that two Nodes has SOMETIMES taints as NotScheduled/NotExecuted which means that node is not ready and new deployments are not working for me because nodes are not available during that time.

I found that only those two nodes have problems with executing command "df -h" (shell stuck) so I suppose that this is the problem with EFS which is not reachable, do you have any idea how to solve the issue?