Trident fails to detach volume when running v25.10.0 or v25.06.2 while working in v24.10.1

[morete]

THIS BUG REPORT CORRESPONDS TO AN EXISTING SUPPORT CASE : #2010576550 but wanted to share findings with the community in case someone else is facing this.

Describe the bug
Trident fails to detach volumes with the following error:

level=error msg="Flush pre-check failed for the device." device=/dev/dm-123 error="exit status 1" logLayer=csi_frontend output="5296141.633595 | /dev/dm-123 is not a multipath map\n" requestID=9655d5e7-ebcd-40d6-aafd-54fc97dec322 requestSource=CSI workflow="node_server=unstage"

This blocks KubeVirt VM provisioning as CDI performs multiple volumes attach/detach operations in the process.

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

• Trident version: v25.10.0 or v25.06.2
• Trident installation flags used: Using helm chart with values tridentSilenceAutosupport: true, tridentDisableAuditLog: false
• Container runtime: containerd://1.7.27
• Kubernetes version: 1.32.8
• Kubernetes orchestrator: Vanilla Kubernetes
• Kubernetes enabled feature gates: No custom feature-gates
• OS: tried on 2 clusters, one on Ubuntu 22.04.5 and another on Ubuntu 24.04.3, in all cases with all the iscsi node preparations.
• NetApp backend types: ONTAP AFF-A90 with version 9.15.1P11
• Other:

To Reproduce
Steps to reproduce the behaviour:
Install Trident, use ontap-san-economy driver and create KubeVirt virtual machines.

Expected behaviour
Volumes should behave/work as they do with Trident v24.10.1, attach and detach without problems.

Additional context
Problem starts as soon as Trident upgrade is running and new VMs are created, problems gets resolved if you downgrade Trident.

[morete]

Testing Trident v25.06.3 at the moment.

[morete]

Problem solved in Trident v25.06.3

[morete]

Porting these fixes to 25.10.0+ would be great!

[morete]

Will reopen issue as this is not sorted on v25.10.x yet and that is our target.