Multi-TiB csi.trident pvc is detected with wrong size by kubevirt guest

[arnongilboa]

Describe the bug
Attaching kubevirt VM a multi-TiB pvc on trident-nfs storage class (csi.trident.netapp.io provisioner), it is detected with wrong size by the guest. For example 90TiB pvc was initially detected (lsblk/journalctl) as 5.2T and slowly increased to 11.5T during the following 3 days. The behavior is not reproduced with other storage classes.

Environment
Provide accurate information about the environment to help us reproduce the issue.

• Trident version: 24.10
• Trident installation flags used: --debug
• Container runtime: cri-o://1.30.6-3.rhaos4.17.git49b5172.el9
• Kubernetes version: 1.30.4
• Kubernetes orchestrator: OpenShift v4.17.1
• Kubernetes enabled feature gates: ?
• OS: Red Hat Enterprise Linux CoreOS 417.94.202410090854-0
• NetApp backend types: ?
• Other:

To Reproduce

1. Create a kubevirt VM with 90 TB trident-nfs PVC
2. Boot the VM
3. Check the detected disk size using journalctl and lsblk
4. 5.2 TB - 11.5 TB disk is detected

Expected behavior
90 TB disk is detected

Additional context

[sjpeeris]

hi @arnongilboa , We need to investigate further to identify the root cause. Can you please open a NetApp support ticket, so they can help collect gather the required info to investigate further ?

[arnongilboa]

@sjpeeris completed?

[sjpeeris]

Hi @arnongilboa, were you able to open a support case, so that our support team can help collect the required logs and investigate more ? If you already did, we can work with our support teams for RC analysis.