search

kxr / ocp4_setup_upi_kvm

Script to Setup an OpenShift 4 UPI Cluster on KVM. Based on this guide: https://kxr.me/2019/08/17/openshift-4-upi-install-libvirt-kvm/
52 stars 57 forks source link

GET result: Internal Server Error on installation of worker node #16

Closed Cheny-chen closed 4 years ago

Cheny-chen commented 4 years ago

Fedora 32 Linux kernel 5.8.9-200.fc32.x86_64 Red Hat OpenShit Version = 4.5.9 Red Hat CoreOS Version = 4.5.6

Install command ./ocp4_setup_upi_kvm.sh --cluster-name ocp4 --ocp-version 4.5.latest --pull-secret pull-secret --vm-dir /mnt/vms2/images

During installation processing, it never check worker1 and worker2. Two workers console keep showing repeatedly

Get result: Internal Server Error
GET https://api-int.ocp4.local:22623/config/worker

The boostrap console show repeatedly

 SELinux: mount invalid. Same superblock, different security settings for (dev mqueue, type mqueue)

If check installation processing log, It never asked for workers if workers Ready or not.

 --> Node master-1.ocp4.local NotReady
  --> Node master-2.ocp4.local NotReady
  --> Node master-3.ocp4.local NotReady
............
  --> Node master-1.ocp4.local Ready
  --> Node master-2.ocp4.local Ready
  --> master-3.ocp4.local Ready

[root@localhost ~]# oc login -u kubeadmin -p ELTqR-mDxtD-2cqbG-DtLEz https://api.ocp4.local:6443 Login successful.

You have access to 57 projects, the list has been suppressed. You can list all projects with 'oc projects'

Using project "default". [root@localhost ~]# oc get nodes NAME STATUS ROLES AGE VERSION master-1.ocp4.local Ready master,worker 50m v1.18.3+2cf11e2 master-2.ocp4.local Ready master,worker 50m v1.18.3+2cf11e2 master-3.ocp4.local Ready master,worker 50m v1.18.3+2cf11e2

kxr commented 4 years ago

Yeah I think in the recent versions I have noticed that sometimes worker nodes are not properly added to the cluster by the installer.

Check the Pending CSR requests by running oc get csr and approve any pending requests by oc adm certificate approve <csr-name>. You will have to do it twice for each node.

Cheny-chen commented 4 years ago

I've tried to re-install again There are not pending issue. works are still not up.

However, version 4.2 is working properly , but not 4.5

[root@localhost auth]# oc get csr NAME AGE SIGNERNAME REQUESTOR CONDITION csr-8tc4c 52m kubernetes.io/kubelet-serving system:node:master-2.ocp4.local Approved,Issued csr-fzsrt 52m kubernetes.io/kube-apiserver-client-kubelet system:serviceaccount:openshift-machine-config-operator:node-bootstrapper Approved,Issued csr-l76qn 52m kubernetes.io/kubelet-serving system:node:master-3.ocp4.local Approved,Issued csr-ph7n6 52m kubernetes.io/kube-apiserver-client-kubelet system:serviceaccount:openshift-machine-config-operator:node-bootstrapper Approved,Issued csr-wz4m9 52m kubernetes.io/kubelet-serving system:node:master-1.ocp4.local Approved,Issued csr-z9kz5 52m kubernetes.io/kube-apiserver-client-kubelet system:serviceaccount:openshift-machine-config-operator:node-bootstrapper Approved,Issued

kxr commented 4 years ago

@Cheny-chen Hello I have added the logic in the script to wait for the CSR approvals and worker nodes to get ready before the script finishes.