I'm trying to run a pod with relatively small limits on a cluster with 3 nodes and it gives me the following error when I run a describe on the pod:
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Warning FailedScheduling 18s (x2 over 18s) default-scheduler 0/6 nodes are available: 1 Insufficient memory, 3 PodToleratesNodeTaints, 4 NoVolumeZoneConflict.In the meantime when I check the nodes it all seems fine:
NAME CPU(cores) CPU% MEMORY(bytes) MEMORY%
ip-172-20-101-66.eu-west-1.compute.internal 113m 2% 3353Mi 21%
ip-172-20-39-218.eu-west-1.compute.internal 364m 9% 7054Mi 44%
ip-172-20-60-117.eu-west-1.compute.internal 109m 2% 3162Mi 19%
ip-172-20-66-24.eu-west-1.compute.internal 183m 4% 3570Mi 22%
ip-172-20-76-91.eu-west-1.compute.internal 332m 8% 5011Mi 31%
ip-172-20-98-243.eu-west-1.compute.internal 579m 14% 6889Mi 43% The following is the limits on the pod:
resources:
limits:
cpu: 50m
memory: 0.5Gi
requests:
cpu: 50m
memory: 0.5GiI can't seem to find any indication that there's limits on nodes, so I'm kind of lost. Can someone please suggest why this might be happening?
The event message appears to describe only one of the nodes has insufficient memory. It also calls attention to Node Taints and Volume Zones.
Perhaps it's worth trying to submit a simpler Pod Spec with the same resources, but with no volumes attached. This won't resolve your issue but may help you in isolating it further.