Is there a way to rebuild a Hyperflex storage controller VM?

[smorgan_byte]

This is an older set of hardware in our lab, otherwise I would opened a TAC case, hoping someone may have some ideas.  One of nodes had a power issue and apparently the storage controller vm got hosed.  When it boots this message comes on the console.

boot disk not found fake mounted

Do I just need to completely remove the node from the cluster, wipe it, and expand the cluster?

[RedNectar]

Hi @smorgan_byte ,

It's a pity you can't open a TAC case for this.  Based on reading the Cisco HyperFlex Data Platform CLI Guide, 4.0, I note:

stcli node remove Command

IMPORTANT: Use the stcli node remove command to permanently remove a node from the cluster. Do not run the stcli node remove command if you intend to add the same node back into the same cluster later. Adding a node back into a cluster that it was already part of is not supported.

So your idea of "remove the node from the cluster, wipe it, and expand the cluster" may not work as well as intended.

[Kirk J]

TAC assists with similar requests that are usually triggered by boot drive failures, etc.

The high level steps involve:

• Reimage ESXi host with HX customized ISO
• Using version matched HX installer appliance (ova), spin up an installer appliance and run custom workflow/install, skipping UCSM, and only selecting configure Hypervisor, and deploy HXDP software.  Then put in offline node's IP info.
• There are TAC only workflows that normally need to be run to make sure rebuilt node is completely readded to cluster.

If TAC case is not an option, you might want to see what a cluster redeploy would take (i.e. alternate space to park your guestVMs temporarily, or backup/restore).

For redeploy, you would want to reimage all your nodes with Hyperflex customized ESXi ISO, use the HX installer ova appliance for regular deployment.

@RedNectar is right, trying to re-use same node, after doing node remove will not work.

Kirk...