kubevirt/hyperconverged-cluster-operator

[release-1.10] Backport CDI 1.58.0 into 1.10 operator release

GingerGeek opened this issue · 2 comments

GingerGeek commented

Is this a BUG REPORT or FEATURE REQUEST?: Feature Request

/kind enhancement

What happened: Current operator release 1.10 includes version 1.57.0 of CDI and related packages

What you expected to happen: Ideally would like 1.58.0 of CDI so we have default storage class specifically for virtualization (kubevirt/containerized-data-importer#2913)

How to reproduce it (as minimally and precisely as possible): N/A

Anything else we need to know?:

#2670 pulled in the CDI bump via a bot in this repo onto the main branch, is it possible to ask the bot to update the release branch?
Not sure if a backport such as this is something you would usually do?

Environment:

  • HCO version (use oc get csv -n kubevirt-hyperconverged): kubevirt-hyperconverged-operator.v1.10.0
  • Kubernetes version (use kubectl version): 4.14.0-0.okd-2024-01-06-084517 v1.27.1-3474+5c56cc35e3edc1-dirty
  • Cloud provider or hardware configuration: Bare metal
  • Install tools:
  • Others:
tiraboschi commented

hi @GingerGeek,
is there any specific reason for this request?
As a general rule we try to keep parallel y-streams of the sibling projects.
So on HCO v1.10.z we will automatically bump v1.0.z releases of kubevirt/kubevirt, v1.57.z of CDI, v0.89.z of CNAO and so on.
We expect to consume CDI v1.58.z releases as part of HCO v1.11.z stream.

GingerGeek commented

Hi,

Thanks for the explanation. My main motivation was to pull in kubervirt/containerized-data-importer#2913 that allows you to tag a storage class as the default for VMs.

My default storage class within the cluster was a standard ODF-like volume, so our VMs gave warnings related to VirtualMachineCRCErrors. My solution currently is to switch the default cluster volume to a VM Storage class which contains "krbd:rxbounce" map option.

My preference would be not to have to modify cluster wide defaults, and indeed prior to installing HCO I didn't even have a default StorageClass within the cluster.

Totally understand it's your usual process to not pull in an upstream update, so I'll close this for now and await the next release!