-
Bug
-
Resolution: Done-Errata
-
Undefined
-
4.12.z
-
None
-
Moderate
-
No
-
False
-
-
N/A
-
Release Note Not Required
Description of problem:
The node selector for the console deployment requires deploying it on the master nodes, The node selector for the console deployment requires deploying it on the master nodes, while the replica count is determined by the infrastructureTopology, which primarily tracks the workers' setup. When an OpenShift cluster is installed with a single master node and multiple workers, this leads the console deployment to request 2 replicas as infrastructureTopology is set to HighlyAvailable. Instead, ControlPlaneTopology is set to SingleReplica as expected.
Version-Release number of selected component (if applicable):
4.16
How reproducible:
Always
Steps to Reproduce:
1. Install an openshift cluster with 1 master and 2 workers
Actual results:
The installation fails as the replicas for the console deployment is set to 2.
apiVersion: config.openshift.io/v1
kind: Infrastructure
metadata:
creationTimestamp: "2024-01-18T08:34:47Z"
generation: 1
name: cluster
resourceVersion: "517"
uid: d89e60b4-2d9c-4867-a2f8-6e80207dc6b8
spec:
cloudConfig:
key: config
name: cloud-provider-config
platformSpec:
aws: {}
type: AWS
status:
apiServerInternalURI: https://api-int.adstefa-a12.qe.devcluster.openshift.com:6443
apiServerURL: https://api.adstefa-a12.qe.devcluster.openshift.com:6443
controlPlaneTopology: SingleReplica
cpuPartitioning: None
etcdDiscoveryDomain: ""
infrastructureName: adstefa-a12-6wlvm
infrastructureTopology: HighlyAvailable
platform: AWS
platformStatus:
aws:
region: us-east-2
type: AWS
apiVersion: apps/v1
kind: Deployment
metadata:
annotations:
....
creationTimestamp: "2024-01-18T08:54:23Z"
generation: 3
labels:
app: console
component: ui
name: console
namespace: openshift-console
spec:
progressDeadlineSeconds: 600
replicas: 2
Expected results:
The replica is set to 1, tracking the ControlPlaneTopology value instead of hte infrastructureTopology.
Additional info:
- is cloned by
-
-
- Closed
-
- is depended on by
-
-
- Closed
-
- links to
-
RHEA-2024:0041
OpenShift Container Platform 4.16.z bug fix update
[OCPBUGS-27335] Installation fails with 1 master and 2 workers as the console deployment set the number of replicas based on the InfrastructureTopology rather than the ControlPlaneTopology
Setting the 'Affects Version' to 4.12.z since this issues is affecting previous versions as well.
Yanping Zhang
added a comment - Tested with payload 4.16.0-0.ci-2024-01-31-151542.
1. Launch a normal cluster with 1 master node and 2 worker nodes. Check console operator, deployment and pods, replica is set as 1 and console pod number is 1.
2. Launch hypershift cluster, set hosted cluster with one worker node, Check console operator, deployment and pods, replica is set as 1 and console pod number is 1.
The bug is fixed.
Yanping Zhang
added a comment - Tested with payload 4.16.0-0.ci-2024-01-31-151542.
1. Launch a normal cluster with 1 master node and 2 worker nodes. Check console operator, deployment and pods, replica is set as 1 and console pod number is 1.
2. Launch hypershift cluster, set hosted cluster with one worker node, Check console operator, deployment and pods, replica is set as 1 and console pod number is 1.
The bug is fixed. Hi yanpzhan1 can you send the result of the following in the case of hypershift (in the guest/hosted cluster)?
oc get infrastructure -o yaml
Yanping Zhang
added a comment - Using payload 4.16.0-0.nightly-2024-01-21-154905 to launch hypershift cluster, and configure hosted cluster only one worker node. Check on hosted cluster, the console operator is abnormal, and console deployment has replicas "2", but only one pod is in Running status.
# oc get node NAME STATUS ROLES AGE VERSION ip-10-0-142-37.ec2.internal Ready worker 61m v1.29.0+f629574 # oc get co console NAME VERSION AVAILABLE PROGRESSING DEGRADED SINCE MESSAGE console 4.16.0-0.nightly-2024-01-21-154905 True True False 60m SyncLoopRefreshProgressing: Working toward version 4.16.0-0.nightly-2024-01-21-154905, 1 replicas available # oc get node NAME STATUS ROLES AGE VERSION ip-10-0-142-37.ec2.internal Ready worker 63m v1.29.0+f629574 # oc -n openshift-console get pod NAME READY STATUS RESTARTS AGE console-544dc8c7d-klhc9 0/1 Pending 0 60m console-54bdd888cf-4wsc9 1/1 Running 0 61m console-5cdb687998-qg8p9 0/1 Pending 0 61m downloads-5bcd554dbf-nxvmx 1/1 Running 0 61m downloads-5bcd554dbf-wv4c4 0/1 Pending 0 61m # oc get deployment console -n openshift-console -ojsonpath='{.spec.replicas}' 2 # oc get deployment downloads -n openshift-console -ojsonpath='{.spec.replicas}' 2 # oc-n openshift-console get pod' NAME READY STATUS RESTARTS AGE console-544dc8c7d-klhc9 0/1 Pending 0 74m console-54bdd888cf-4wsc9 1/1 Running 0 75m console-5cdb687998-qg8p9 0/1 Pending 0 75m downloads-5bcd554dbf-nxvmx 1/1 Running 0 75m downloads-5bcd554dbf-wv4c4 0/1 Pending 0 75m [root@MiWiFi-RB03-srv ~]# oc get deploy -n openshift-console NAME READY UP-TO-DATE AVAILABLE AGE console 1/2 1 1 76m downloads 1/2 2 1 76m
The fix in pr is not suitable for hosted cluster, since console pods are deployed on worker nodes on this kind of cluster.
rhn-support-adistefa could you help to consider the fix for condition when cluster is hosted cluster?
Yanping Zhang
added a comment - Using payload 4.16.0-0.nightly-2024-01-21-154905 to launch hypershift cluster, and configure hosted cluster only one worker node. Check on hosted cluster, the console operator is abnormal, and console deployment has replicas "2", but only one pod is in Running status.
# oc get node
NAME STATUS ROLES AGE VERSION
ip-10-0-142-37.ec2.internal Ready worker 61m v1.29.0+f629574
# oc get co console
NAME VERSION AVAILABLE PROGRESSING DEGRADED SINCE MESSAGE
console 4.16.0-0.nightly-2024-01-21-154905 True True False 60m SyncLoopRefreshProgressing: Working toward version 4.16.0-0.nightly-2024-01-21-154905, 1 replicas available
# oc get node
NAME STATUS ROLES AGE VERSION
ip-10-0-142-37.ec2.internal Ready worker 63m v1.29.0+f629574
# oc -n openshift-console get pod
NAME READY STATUS RESTARTS AGE
console-544dc8c7d-klhc9 0/1 Pending 0 60m
console-54bdd888cf-4wsc9 1/1 Running 0 61m
console-5cdb687998-qg8p9 0/1 Pending 0 61m
downloads-5bcd554dbf-nxvmx 1/1 Running 0 61m
downloads-5bcd554dbf-wv4c4 0/1 Pending 0 61m
# oc get deployment console -n openshift-console -ojsonpath= '{.spec.replicas}'
2
# oc get deployment downloads -n openshift-console -ojsonpath= '{.spec.replicas}'
2
# oc-n openshift-console get pod'
NAME READY STATUS RESTARTS AGE
console-544dc8c7d-klhc9 0/1 Pending 0 74m
console-54bdd888cf-4wsc9 1/1 Running 0 75m
console-5cdb687998-qg8p9 0/1 Pending 0 75m
downloads-5bcd554dbf-nxvmx 1/1 Running 0 75m
downloads-5bcd554dbf-wv4c4 0/1 Pending 0 75m
[root@MiWiFi-RB03-srv ~]# oc get deploy -n openshift-console
NAME READY UP-TO-DATE AVAILABLE AGE
console 1/2 1 1 76m
downloads 1/2 2 1 76m
The fix in pr is not suitable for hosted cluster, since console pods are deployed on worker nodes on this kind of cluster.
rhn-support-adistefa could you help to consider the fix for condition when cluster is hosted cluster? Using payload 4.16.0-0.nightly-2024-01-21-154905 to launch a normal cluster with 1 master node and 2 worker nodes, the cluster is launched successfully. Check console operator, deployment and pods, replica is set as 1 and pod number is 1. They work as expected:
$ oc get co console
NAME VERSION AVAILABLE PROGRESSING DEGRADED SINCE MESSAGE
console 4.16.0-0.nightly-2024-01-21-154905 True False False 27m
$ oc get deployment console -n openshift-console -ojsonpath='
'
1
$ oc get pod -n openshift-console
NAME READY STATUS RESTARTS AGE
console-5ffbb8644-lpnt5 1/1 Running 0 25m
downloads-7597c9f7c-62sxs 1/1 Running 0 33m
Yanping Zhang
added a comment - - edited Using payload 4.16.0-0.nightly-2024-01-21-154905 to launch a normal cluster with 1 master node and 2 worker nodes, the cluster is launched successfully. Check console operator, deployment and pods, replica is set as 1 and pod number is 1. They work as expected:
$ oc get co console
NAME VERSION AVAILABLE PROGRESSING DEGRADED SINCE MESSAGE
console 4.16.0-0.nightly-2024-01-21-154905 True False False 27m
$ oc get deployment console -n openshift-console -ojsonpath='
{.spec.replicas}
'
1
$ oc get pod -n openshift-console
NAME READY STATUS RESTARTS AGE
console-5ffbb8644-lpnt5 1/1 Running 0 25m
downloads-7597c9f7c-62sxs 1/1 Running 0 33m Hi jhadvig@redhat.com, are you ok with backporting the fix for a few releases?
OpenShift Jira Bot
added a comment - Looks like this bug is far enough along in the workflow that a code fix is ready. Customers and support need to know the backport plan. Please complete the "Target Backport Versions" field to indicate which version(s) will receive the fix.
OpenShift Jira Bot
added a comment - Looks like this bug is far enough along in the workflow that a code fix is ready. Customers and support need to know the backport plan. Please complete the " Target Backport Versions " field to indicate which version(s) will receive the fix. 
Since the problem described in this issue should be resolved in a recent advisory, it has been closed.
For information on the advisory (Critical: OpenShift Container Platform 4.16.0 bug fix and security update), and where to find the updated files, follow the link below.
If the solution does not work for you, open a new bug report.
https://access.redhat.com/errata/RHSA-2024:0041