vm essentials migration tool after os boot failed

[jskim88]

We are currently conducting a customer proof of concept in a VME 9.0.1 environment.

The customer’s Linux VM is running Rocky Linux 8.9. Following the migration guide, we installed the VirtIO drivers and the QEMU Guest Agent.

The migration completed successfully without any errors. However, when the VM starts, it fails to boot the operating system and enters the UEFI Interactive Shell instead.

We are not sure what we may have missed or which areas we should check again.

We would greatly appreciate any assistance or advice from anyone who has experienced a similar issue.

Thank you.

1.png

[SergeMe · Accepted solution]

It would appear from the screenshot that this disk doesn't have a recognisable partition table on it. It only shows the raw disk itself (BLK0:).

What you could try at this point is validate if there is a valid partition table (such as MBR).

dblk blk0: 0 1

That dumps the raw bytes at LBA0. A valid partition table (MBR or GPT's protective MBR) ends with the signature 55 AA at the very end of that sector.

Then try:

dblk blk0: 1 1

That dumps LBA1 — a genuine GPT header starts with the ASCII text EFI PART near the beginning if one exists.

If both come back as zeros or garbage, the disk this VM is booting from has no partition table the firmware can see at all, which points at something wrong with the migration itself. 

At that point go back to the host side and check the actual backing disk file for this VM with qemu-img info on the HVM host. Look at its allocated/used size and whether that matches with anything like a populated OS install or close to empty. The latter would tell us whether the migration job's result was accurate or not.

If either dump does show a valid signature, that would be a more unusual case. Let us know.

[jskim88]

I resolved the issue as follows.

In our test environment, both the VME and VMware servers were using the same DNS server. However, in the customer’s POC environment, the VME and VMware servers were configured to use different DNS servers.

Since the VMware Cloud connection appeared normal in the VME Manager UI, and the ESXi hosts and VMs were displayed correctly, I did not initially suspect a DNS-related issue.

The migration proceeded successfully only after I added the IP addresses and FQDNs of the vCenter Server and ESXi hosts to the /etc/hosts files on both the VME hosts and the VME Manager.

Thank you to everyone who shared their suggestions and insights.

[Thaufique_Mod]

Hello @jskim88,

Thank you for sharing the resolution. It is very helpful.