Single Node Deployment (5.3.1) - "Disconnected Storage and Fed NICs detected"

[egoqed]

Hi community
A weird issue in a test environment:
We try to setup a single node (in an own cluster).
So, setup is set to not using stretch and the count is set to 1. My assumption is, that with this setting, the depl. manager ignores the storage and federation part. As of course there is no counterpart to connect to....

But this is not the case. We permanently get the error:

FAILURE: Disconnected Storage and Federation NIC(s) detected on host {IP_ADDRESS} ({MAC_ADDRESS}).
FAILED_DETAIL1: Verify the physical connection of the NIC(s) and make any necessary adjustments to continue.
FAILED_DETAIL2: Or, click Back and choose a different {NETWORK} NIC(s) at the Select Network Interface Cards screen.

As it is stated as failure the manager denies the setup. Also the messaged does not correlate with the PVCs in the vSphere Documentation as it somehow is a mix out of both (HostManagementNicValidator and HostStorageNicValidator).

Work around was to attach a DAC and link it to a 2nd host. Of course, the issue is solved but what happens if there is only one device available??
Also i know about the possibility to skip the depl pre test ...but again, I would like to have the reason why

thank you
egoqed

[Steven Clementi]

Storage and Federation networks are not ignored.  The NIC's need to have active links because the setup is going to assume that in the future, you will add more nodes so it needs to configure those services peroperly from the get-go.

Quick fix: If you have the adaptors to do it (and you should if you have 2 dual 10/25G cards, cross connect one port on one card and one port on the other card to provide the "link light".

[Sanika · Accepted solution]

Hello @egoqed 

The error happens because the deployment manager still checks for storage and federation NICs, even in single-node setups. It assumes future scalability or redundancy, even though you are not using stretch mode or multiple nodes.

Attaching a DAC cable and linking to another host solves it because it creates a valid connection, but I think that's not feasible if you have only one node.

Skipping the pre-test is an option but doesn't resolve the actual issue. It’s not recommended for production, even in single-node setups, unless you have manually validated all networking components.

If you have further doubts regarding this, I would recommend you to contact HPE support team for more assistance.

Regards,
Sanika.

If you feel this was helpful, please click the KUDOS thumb below. Also consider marking this as an "Accepted Solution", if the post has helped to solve your issue.