Calico / Tigera-operator tolerations leading to pods do be scheduled on draining node
Summary
Observed on https://gitlab.com/sylva-projects/sylva-core/-/jobs/7243615488 (upgrade of kubeadm management cluster)
We see that cp-2 machine fails to be deleted:
sylva-system mgmt-1357074163-kubeadm-capm3-virt-control-plane-fjxgr mgmt-1357074163-kubeadm-capm3-virt mgmt-1357074163-kubeadm-capm3-virt-management-cp-0 metal3://sylva-system/mgmt-1357074163-kubeadm-capm3-virt-management-cp-0/mgmt-1357074163-kubeadm-capm3-virt-cp-eb3f4d18dc-zx4jp Running 55m v1.28.9
sylva-system mgmt-1357074163-kubeadm-capm3-virt-control-plane-h7c6c mgmt-1357074163-kubeadm-capm3-virt mgmt-1357074163-kubeadm-capm3-virt-management-cp-2 metal3://sylva-system/mgmt-1357074163-kubeadm-capm3-virt-management-cp-2/mgmt-1357074163-kubeadm-capm3-virt-cp-eb3f4d18dc-5dgcz Deleting 95m v1.28.9
sylva-system mgmt-1357074163-kubeadm-capm3-virt-control-plane-mwcpb mgmt-1357074163-kubeadm-capm3-virt mgmt-1357074163-kubeadm-capm3-virt-management-cp-1 metal3://sylva-system/mgmt-1357074163-kubeadm-capm3-virt-management-cp-1/mgmt-1357074163-kubeadm-capm3-virt-cp-eb3f4d18dc-nfx9q Running 95m v1.28.9
sylva-system mgmt-1357074163-kubeadm-capm3-virt-md0-ltd6j-xgw5p mgmt-1357074163-kubeadm-capm3-virt Provisioning 39m v1.28.9
CAPI keeps evicting tiger-operator pods
I0702 11:04:38.836420 1 machine_controller.go:910] "evicting pod tigera-operator/tigera-operator-795545875-njmzp\n" controller="machine" controllerGroup="cluster.x-k8s.io" controllerKind="Machine" Machine="sylva-system/mgmt-1357074163-kubeadm-capm3-virt-control-plane-h7c6c" namespace="sylva-system" name="mgmt-1357074163-kubeadm-capm3-virt-control-plane-h7c6c" reconcileID="f124c124-e3d4-472c-8376-a3048d23ad39" KubeadmControlPlane="sylva-system/mgmt-1357074163-kubeadm-capm3-virt-control-plane" Cluster="sylva-system/mgmt-1357074163-kubeadm-capm3-virt" Node="mgmt-1357074163-kubeadm-capm3-virt-management-cp-2"
I0702 11:04:38.836473 1 machine_controller.go:910] "evicting pod tigera-operator/tigera-operator-795545875-mxxhq\n" controller="machine" controllerGroup="cluster.x-k8s.io" controllerKind="Machine" Machine="sylva-system/mgmt-1357074163-kubeadm-capm3-virt-control-plane-h7c6c" namespace="sylva-system" name="mgmt-1357074163-kubeadm-capm3-virt-control-plane-h7c6c" reconcileID="f124c124-e3d4-472c-8376-a3048d23ad39" KubeadmControlPlane="sylva-system/mgmt-1357074163-kubeadm-capm3-virt-control-plane" Cluster="sylva-system/mgmt-1357074163-kubeadm-capm3-virt" Node="mgmt-1357074163-kubeadm-capm3-virt-management-cp-2"
I0702 11:04:38.836161 1 machine_controller.go:910] "evicting pod tigera-operator/tigera-operator-795545875-k9rxl\n" controller="machine" controllerGroup="cluster.x-k8s.io" controllerKind="Machine" Machine="sylva-system/mgmt-1357074163-kubeadm-capm3-virt-control-plane-h7c6c" namespace="sylva-system" name="mgmt-1357074163-kubeadm-capm3-virt-control-plane-h7c6c" reconcileID="f124c124-e3d4-472c-8376-a3048d23ad39" KubeadmControlPlane="sylva-system/mgmt-1357074163-kubeadm-capm3-virt-control-plane" Cluster="sylva-system/mgmt-1357074163-kubeadm-capm3-virt" Node="mgmt-1357074163-kubeadm-capm3-virt-management-cp-2"
I0702 11:04:38.836667 1 machine_controller.go:910] "evicting pod tigera-operator/tigera-operator-795545875-ppq5l\n" controller="machine" controllerGroup="cluster.x-k8s.io" controllerKind="Machine" Machine="sylva-system/mgmt-1357074163-kubeadm-capm3-virt-control-plane-h7c6c" namespace="sylva-system" name="mgmt-1357074163-kubeadm-capm3-virt-control-plane-h7c6c" reconcileID="f124c124-e3d4-472c-8376-a3048d23ad39" KubeadmControlPlane="sylva-system/mgmt-1357074163-kubeadm-capm3-virt-control-plane" Cluster="sylva-system/mgmt-1357074163-kubeadm-capm3-virt" Node="mgmt-1357074163-kubeadm-capm3-virt-management-cp-2"
Plenty of tigera-operator pods are scheduled on node-2
tigera-operator tigera-operator-795545875-hhtst 0/1 Terminating 0 9m36s <none> mgmt-1357074163-kubeadm-capm3-virt-management-cp-2 <none> <none>
tigera-operator tigera-operator-795545875-hjfh7 0/1 Terminating 0 7m18s <none> mgmt-1357074163-kubeadm-capm3-virt-management-cp-2 <none> <none>
tigera-operator tigera-operator-795545875-j2fvw 0/1 Terminating 0 92s <none> mgmt-1357074163-kubeadm-capm3-virt-management-cp-2 <none> <none>
tigera-operator tigera-operator-795545875-j4m9h 0/1 Terminating 0 13m <none> mgmt-1357074163-kubeadm-capm3-virt-management-cp-2 <none> <none>
tigera-operator tigera-operator-795545875-j887f 0/1 Terminating 0 27m <none> mgmt-1357074163-kubeadm-capm3-virt-management-cp-2 <none> <none>
tigera-operator tigera-operator-795545875-jzr2b 0/1 Terminating 0 29m <none> mgmt-1357074163-kubeadm-capm3-virt-management-cp-2 <none> <none>
tigera-operator tigera-operator-795545875-k27t7 0/1 Terminating 0 28m <none> mgmt-1357074163-kubeadm-capm3-virt-management-cp-2 <none> <none>
tigera-operator tigera-operator-795545875-k9rxl 0/1 Terminating 0 11m <none> mgmt-1357074163-kubeadm-capm3-virt-management-cp-2 <none> <none>
tigera-operator tigera-operator-795545875-kfr7v 0/1 Terminating 0 25m <none> mgmt-1357074163-kubeadm-capm3-virt-management-cp-2 <none> <none>
tigera-operator tigera-operator-795545875-kg7b9 0/1 Terminating 0 6m54s <none> mgmt-1357074163-kubeadm-capm3-virt-management-cp-2 <none> <none>
tigera-operator tigera-operator-795545875-kghf6 0/1 Terminating 0 21m <none> mgmt-1357074163-kubeadm-capm3-virt-management-cp-2 <none> <none>
tigera-operator tigera-operator-795545875-khchw 0/1 Terminating 0 16m <none> mgmt-1357074163-kubeadm-capm3-virt-management-cp-2 <none> <none>
tigera-operator tigera-operator-795545875-ksq8j 0/1 Terminating 0 24m <none> mgmt-1357074163-kubeadm-capm3-virt-management-cp-2 <none> <none>
tigera-operator tigera-operator-795545875-kv2mj 0/1 Terminating 0 115s <none> mgmt-1357074163-kubeadm-capm3-virt-management-cp-2
This seems to be related to https://github.com/projectcalico/calico/issues/6136
Edited by Francois Eleouet