VMware Cloud Community
izblah
Contributor
Contributor
Jump to solution

ESXi 4.1 no longer boots...

I've had an ESXi 4.1 host up and running for many many years now. (Well, it started as an ESXi 3.5 host, and has been updated to 4.1)

Nothing - I repeat - nothing has been changed/updated/modified on the ESXi host since it was successfully updated to ESXi 4.1 many months ago.

I power on this server, usually run it for a few days, and shut it down. The host runs multiple Windows VMs in an AD lab environment.

Yesterday, this server all-of-the-sudden no longer boots into the hypervisor.

The system itself POSTs succesfully, the Adaptec 5805 controller loads successfully, no issues with any logical disks.

The problem is that once the "Loading VMWare Hypervisor" screen starts, the progress bar at the bottom halts near the end and never goes any further. At this point, the system is non-responsive, keyboard presses do not register, etc.

If I hit ESC during this loading screen (before it freezes), I see many modules load. The last item displayed before it freezes is:

Booting: MBI=0x000100f0, entry=0x00400256

Searching the internet and forums for this only turned up people having issues trying to install ESXi on non-supported or non x86-64 capable CPUs. This is obviously not the case in my situation, as this host has been successfully running ESXi for many years now.

So - maybe some system files got corrupted some how?

Is there a way to reload the host's ESXi OS without wiping any of the VMFS partitions? Something like an OS "refresh" or a re-load on-top? I need to maintain all of the VMFS datastores...

Any other suggestions that I can try to revive this host so it boots up?

Reply
0 Kudos
1 Solution

Accepted Solutions
DSTAVERT
Immortal
Immortal
Jump to solution

You can do a repair install. It should repair the ESXi install and leave the VMFS partition. I never like doing a repair install without having my VMs backed up.

Assuming you have had updates along the way you can try to boot the most previous update. Reboot the ESXi host and when the initial ESXi boot screen appears press SHIFT + R It will prompt you for confirmation.

You can install ESXi to a USB stick. You will need to follow the following to get booted. http://kb.vmware.com/kb/1035107

-- David -- VMware Communities Moderator

View solution in original post

Reply
0 Kudos
6 Replies
DougParker
Contributor
Contributor
Jump to solution

Is the server listed on the HCL?  Are you using local storage or a SAN?

Reply
0 Kudos
izblah
Contributor
Contributor
Jump to solution

The HCL isn't relevant. The server has ran ESXi for many years now. The server is a custom build.

The host uses local storage only on an Adaptec 5805 controller.

If I reload the ESXi OS ontop of my current install, will the VMFS datastores and my ESXi config be wiped out?

Reply
0 Kudos
DSTAVERT
Immortal
Immortal
Jump to solution

You can do a repair install. It should repair the ESXi install and leave the VMFS partition. I never like doing a repair install without having my VMs backed up.

Assuming you have had updates along the way you can try to boot the most previous update. Reboot the ESXi host and when the initial ESXi boot screen appears press SHIFT + R It will prompt you for confirmation.

You can install ESXi to a USB stick. You will need to follow the following to get booted. http://kb.vmware.com/kb/1035107

-- David -- VMware Communities Moderator
Reply
0 Kudos
a_p_
Leadership
Leadership
Jump to solution

An installation of ESXi will wipe the installation disk, so installing with a VMFS datastore on the installation disk is not an option. A repair installation might also not be an option because the partition layout changed from 3.5 to 4.1 and the repair installation might destroy the VMFS partition.

Which update did you apply the last time (from which version/build to which version/build)? Depending on this, pressing Shift-R during boot to revert back to the prior firmware version may work for you.

Another option could be to attach a USB stick to your host and install ESXi on it!?

André

DSTAVERT
Immortal
Immortal
Jump to solution

The KB article is somewhat drastic. This will get you going althhough it isn't permanent http://www.vm-help.com/esx41/file_system_UUID.php

-- David -- VMware Communities Moderator
Reply
0 Kudos
izblah
Contributor
Contributor
Jump to solution

Okay...

So, I went ahead and tried the SHIFT-R recovery option...

I first installed ESXi 3.5, then updated to 4.0, then to 4.0u1, then to 4.1 (I'm pretty sure, at least)

SHIFT-R told me that it would revert back to 4.0.0-2.19.33

I accepted, and quickly it told me it was done...

I rebooted the host, and the original behavior noted in POST1 didn't happen. Instead, I was greeted with a PSOD and the following error:

No pages allocated to Node 1 -- big mismatch between BIOS and SRAT memory maps, or MTRR error. Try checking memory or upgrading BIOS.

Well, at least that is more helpful!

The screen states that I am now running ESXi 4.0.0 build-332073

So, I ran memtest86+ and was greeted with all errors.

So anyways, I begin by upgrading my BIOS to the latest version on Tyan's site, and then pulled out all eight DIMMs, and blowing the case and sockets out with canned air (The server was pretty clean already, as I air it out a few times a year).

Anyways, that seemed to do the trick... My ESXi host booted up! (in Maintenance Mode, presumably because I reverted to a previous ESXi version)

Besides a few other little things (like my Logical Disk 0 doing a verification/rebuild), all seems well...

I'll let it bake in, then upgrade ESXi to 4.1 then to 4.1u1

Thanks for the responses and suggestions!

Reply
0 Kudos