Nutanix AHV update fails from LCM

Are you running Community Edition? If so, read this: https://next.nutanix.com/discussion-forum-14/do-not-upgrade-ce-to-ahv-10-3-and-higher-45028

hi ​@mmusialek 
Before going further, a couple of quick questions that will help narrow this down.

Is this environment running-

1/Nutanix Community Edition
2/ Nested AHV (inside VMware / VirtualBox / KVM)
3/Or a physical single-node cluster?

The error you're seeing sometimes happens depending on how the system disk is presented to AHV, so knowing the setup will help identify the cause.

Thanks 
Selvamani.S

Hi all, 

Thanks for your responses. It seems that issue with version 10.3 will stop my plans to upgrade to newest version. For the record it’s community edition built on VMWare virtual machine. 

By the way, do you have any ideas what to do with Prism deployment issue i’ve mentioned above?

Ideally I’d like to build it as a separate VM on my VMWare environment, though as far as I know it’s not possible on CE, right?

hi ​@mmusialek 

Since this is Nutanix Community Edition running nested on VMware, what you’re seeing is fairly common.

For Prism Central, CE often shows “No compatible versions for deployment” in Prism Element. The compatibility check mainly targets production AOS builds and licensed environments, so CE clusters frequently don’t return a deployable PC version.

Running Prism Central as a separate VM on VMware and attaching it to a CE cluster is generally not supported either. In CE labs, when PC is used it’s normally deployed from Prism Element, which creates the VM inside the AHV cluster.

About the AHV upgrade error:

No symlink matching */md/* or */disk/by-id/* found for block device /dev/sda

This usually happens in nested CE setups on VMware. During the AHV upgrade the LCM reimage process expects the system disk to have a persistent path (like /dev/disk/by-id). Some VMware disk/controller configurations don’t expose that, so the upgrade stops at the prepare_reimage step.

On the AHV host you can check it with:

ls -l /dev/disk/by-id

If the system disk doesn’t appear there, the installer can’t map it correctly.

In VMware-based CE labs this is often fixed by:

• using a SCSI controller (LSI Logic SAS or VMware PVSCSI)
• avoiding SATA disks
• keeping the AHV system disk as /dev/sda

Redeploying the node with a SCSI controller usually resolves the LCM upgrade issue in nested environments.

These docs may also help

https://portal.nutanix.com/page/documents/details?targetId=Nutanix-Community-Edition-Getting-Started:top-community-edition-limitations-c.html

https://portal.nutanix.com/page/documents/details?targetId=Prism-Central-Guide:prism-central-deployment-overview-pc-c.html

Thanks
Selvamani.S

Please ignore the AI bot above. 

Prism Central is a virtual machine you install inside the nutanix cluster. You can deploy Prism Central on a CE cluster without any problem. Running CE 2.1 with the default version should give you the option to install Prism Central. I will record a video on this later.