Errors seen in rancher logs when group memberships are refreshed on a rancher upgrade

[anupama2501]

Rancher Server Setup

• Rancher version: v2.5.14 upgraded to v2.6-head
• Installation option (Docker install/Helm Chart): Docker install
  • If Helm Chart, Kubernetes Info: k3s
    Cluster Type (RKE1, RKE2, k3s, EKS, etc):
    Node Setup: 1 node
    Version: v1.21.3+k3s1
• Proxy/Cert Details: self-signed

Information about the Cluster

• Kubernetes version: v1.20.15
• Cluster Type (Local/Downstream): Downstream infrastructure provider 3 worker, 1 etcd, 1 cp
RKE1

Describe the bug
Refreshing the group memberships of the users, returns an error when azure AD auth is enabled.

To Reproduce

1. Install rancher v2.5.14
2. Enable azure AD auth
3. Login as users from azure AD and create a few downstream clusters
4. Upgrade the rancher server to the latest v2.6 head version
5. Refresh group membership from Users and auth >> Users >> Refresh group memberships or from Users and auth >> groups >> Refresh group memberships

Result
Verify the rancher logs and notice the following error logs:

2022/06/23 00:16:17 [ERROR] Error refreshing token principals, skipping: graphrbac.UsersClient#Get: Failure responding to request: StatusCode=400 -- Original Error: autorest/azure: Service returned an error. Status=400 Code="Unknown" Message="Unknown service error" Details=[{"odata.error":{"code":"Request_BadRequest","message":{"lang":"en","value":"Bad request. Please fix the request before retrying."}}}]

Expected Result
No error logs

Additional context
Does not happen on a fresh install of rancher on v2.6-head and azure AD enabled.

[anupama2501]

Verified on rancher upgraded from v2.5.15 >> v2.6-head 07ee596 and fresh install v2.6.7-rc2

Enabled azure AD >> Added two groups from Users & auth , groups >> refresh group membership
No error logs seen.