OKD 4.14.0 failed to connect to the openshift console (1 master / 3 workers) #1823
-
|
Hello, I'm trying to install okd 4.14.0-0.okd-2023-11-14-101924 on bare metal server using Fedora Core & Red hat linux. Master: 32GB ram / 8 cpu The cluster is up on the redhat website. I attached some logs:
When opening a support case, bugzilla, or issue please include the following summary data along with any other requested information:
NAME READY STATUS RESTARTS AGE NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE NAME READY UP-TO-DATE AVAILABLE AGE NAME DESIRED CURRENT READY AGE NAME HOST/PORT PATH SERVICES PORT TERMINATION WILDCARD
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE NAME READY UP-TO-DATE AVAILABLE AGE NAME DESIRED CURRENT READY AGE NAME HOST/PORT PATH SERVICES PORT TERMINATION WILDCARD
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE NAME READY UP-TO-DATE AVAILABLE AGE NAME DESIRED CURRENT READY AGE
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE NAME DESIRED CURRENT READY UP-TO-DATE AVAILABLE NODE SELECTOR AGE NAME HOST/PORT PATH SERVICES PORT TERMINATION WILDCARD Does anyone have an idea what should be the problem ? |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 8 replies
-
Did must-gather produce the archive? That's the one we need |
Beta Was this translation helpful? Give feedback.
-
|
Thanks for your answer. I have one thing who I don't understand, on the redhat website, the authentication is down but when I saw pods status (with the oc commands), these ones are running =) I attached the new installation logs: |
Beta Was this translation helpful? Give feedback.
-
|
The console operator is unavailable because the ingress operator is degraded. The authentication operator is also degraded because of this. The ingress operator reports the canaries route checks are failing ( I would first look into resolving that issue. For example, is the wildcard |
Beta Was this translation helpful? Give feedback.
-
|
I use a local dns based on named/bind, I don't have a wildcard, all certificates are self-signed The reverse zone: I can ping and resolve the console: The first time who I installed the cluster, I installed it with 3 masters and I had access to the console and everything was working. You will find bellow the file configuration: I attached the haproxy configuration: I have no idea, normally if it works with 3 master, it should work with 1 master ? |
Beta Was this translation helpful? Give feedback.
-
|
From the haproxy configuration: From the must-gather: The cluster ingress is running on 192.168.22.6 and that node is not in the haproxy backend configuration. You can adjust the default IngressController As a sidenote, the SRV records are no longer needed since OKD 4.4. Please see User-provisioned DNS requirements for more on DNS settings in a UPI installation. |
Beta Was this translation helpful? Give feedback.
-
|
It finally works, I add the master and delete the compute nodes on the haproxy, like this: Small question before closing the thread, do I need to put the compute nodes on the haproxy backend next to the control plane ? |
Beta Was this translation helpful? Give feedback.
-
|
In this cluster — with only one control plane — the ingress operator will schedule the ingress pod on that control plane node by default. So you would not need to add the compute nodes to the haproxy backend. In a cluster with multiple control planes, the ingress controller default behaviour is different and will schedule the ingress pod(s) on the compute nodes. |
Beta Was this translation helpful? Give feedback.
From the haproxy configuration:
From the must-gather:
The cluster ingress is running on 192.168.22.6 and that node is not in the haproxy backend configuration. You can adjust the default IngressController
replicasandn…