APEX Cloud Platform for Red Hat OpenShift

本文档重点介绍 Dell Technologies 企业和数据中心基础架构产品的最低、建议和最新代码版本。戴尔会仔细评估和监控代码部署数据，以确保产品版本符合我们严格的期望，为客户提供可靠的产品和解决方案。

本文档重点介绍 Dell Technologies 企业和数据中心基础架构产品的最低、建议和最新代码版本。戴尔会仔细评估和监控代码部署数据，以确保产品版本符合我们严格的期望，为客户提供可靠的产品和解决方案。

The VM cannot be created via OpenShift Virtualization if no default StorageClass exists in OCP cluster. Because non-default StorageClass cannot be selected for rootdisk or for the user added disks.

If user set MTU for management traffic during cluster deployment or node addition process, corresponding MTU on physical switch ports needs to be set as well, otherwise network validation will fail.

Add worker nodes via Dell APEX Cloud Platform node addition function, the newly added nodes hardware information is missing on the Red Hat Web Console Node Details view page.

OCP upgrade post check error for node not ready or scheduling disabled due to VM configured with ReadWriteOnce(RWO) storage volume that cannot be live migrated.

OCP upgrade precheck failed for dryrun drain node error, due to some VM cannot be live migrated or some pod cannot be evicted.

An iDRAC issue causing cluster deploy failure, the firmware and hardware inventory information cannot be returned from iDRAC.

This KB introduces some potential reasons that can cause OCP nodes rebooting failure during updating the OCP ingress certificate and provides suggested prechecks to mitigate the potential failure.

Take VM snapshot via OpenShift Virtualization UI fails if there is IO running in the VM in OCP cluster.

Due to OpenShift Virtualization limitation, VM cannot be cloned if it has disks with "hotpluggable" option enabled.

One control plane node is non-gracefully shutdown, after it is powered on, it will go to "not ready" state and cause the cluster not operational.

In disconnected environment (cluster in restricted networks or without internet connection), when a user tries to remove a node via Dell APEX Cloud Platform web console plugin, the process fails at step "Check if the node had bare pods".

LCM failed during node reboot due to CSI controller pod and depot manager pod run into deadlock.

An OpenShift Container Platform issue causing cluster deploy failure, static pod status changed to completed.

Filesystem has less than 3% space left.

PersistentVolume is filling up.

PersistentVolumeInodes are filling up.

For a cluster with consolidated control plane, the control plane node has a control plane role as well as a worker role. The KB is used to convert consolidated control plane to dedicated control plane by running a script. The script will remove the worker role from control plane nodes.

Dell APEX Cloud Platform (ACP) provided plugin UI is capable of updating the OpenShift Container Platform (OCP) ingress certificate and automatically incorporating the trusted CA certificate into the OCP cluster. In addition to the UI approach, this knowledge base article details a manual procedure for adding or removing trusted CA certificates from the OCP cluster, which might be necessary when introducing new CA certificates or when existing certificates are nearing expiration.

An OpenShift Container Platform (OCP) issue causing cluster deployment failure, machine-config-operator degraded.

Filesystem has less than 3% inodes left.

Sustained high CPU utilization on a single control plane node, more CPU pressure is likely to cause a failover; increase available CPU.

Refer to this document to update the DNS settings in an OpenShift cluster.

Extremely High Individual Control Plane Memory

Cluster update failed to stage firmware on node when upgrading to 03.04.01.00.

This KB describe the procedure of manually rotate ACP ingress (gateway) certificate using CLI.

After updating the certificate in the Hub cluster, it does not take effect on the nodes of the HCP managed cluster.

Powerflex CSI node pods are "CrashLoopBackOff" after updating OCP ingress certificate

For an APEX Cloud Platform for Red Hat OpenShift Container Platform (ACP4OCP) cluster running on 03.01.00.00 release, day2 node addition failed at step "Approve CSR" due to hostname DNS lookup failure, when the nodes in cluster use the bond mode 4 "802.3ad" for management traffic.