I’m configuring a small private bare-metal Kubernetes cluster.
I wanted to try the Portworx features so I can replace the current open-source storage driver with something more advance and easier for maintenance, but the installation failed.
The system: one master + one worker node in two mini-PCs. OS: RHEL 9.2 (the current latest version). The master taint is configured to serve as worker. I know it’s below the minimal requirements. I need to buy more hardware soon…
I scaled down the csi-driver, stork and stork-scheduler from 3 workers to 1 worker, but the portworx pod does not start.
The major error I found in the logs is:
Failed to load PX filesystem dependencies for kernel 5.14.0-284.11.1.el9_2.x86_64
The current version 2.13 supports RHEL 8.7. On the node a portworx.service was installed automatically, but it’s failing as well. The exit status code is 10.
The portworx api pod is failing:
Readiness probe failed: Get “http://127.0.0.1:11001/status”: dial tcp 127.0.0.1:11001: connect: connection refused
As I see the portworx does not support the latest Kubernetes and RHEL versions.
My question is: Is it possible to configure Portworx so It is less dependent on the node OS kernel?. The worker integrated storage is already mounted to a specific path, so no complex update in the node configurations is required.