5

IMPORTANT UPDATE: I tried to use the oldest VM I have, 275GB in size. As soon as I resize it to 500GB to make more space (it needs it), it will show the "Error loading operating system" now. What could cause this? Can I shrink it back?

-- Original --

We are running a Windows 2003 Server Standard in a VM because of a legacy software that we need to run (VMware ESXi, 6.5.0, 4564106). Lately, the server crashed (ESXi purple screen) and since then the machine doesn't boot anymore.

To solve the problem we restored a backup from a day earlier, but weirldy enough, the problem doesn't go away. Even when we go several days back, it still won't boot.

The problem in detail is the following: First, I get 'Error loading operating system' on booting. I can get it to a black cursor in the top left corner, that doesn't blink, by booting with the CD ISO and doing fixboot, fixmbr and bootcfg /rebuild. But I can't get it to boot properly. I tried the following:

  • Changing the disk controller in vsphere from SCSI to IDE, redoing the trinity (fixboot, fixmbr, bootcfg /rebuild)
  • I looked at the data using a debian live, everything is there. autoexec.bat and config.sys are empty (0 bytes), boot.ini is ok.
  • We tried to install Debian on a second disk, tried to boot with GRUB, same thing (black screen, white cursor, not blinking)
  • We tried more RAM, less RAM, fewer CPUs, more CPUs, using VM Version 10 (6.0), but no luck
  • Running it on another host
  • Running the Windows Server 2003 restore (where it overwrites over several system files), same result
  • Running chkdsk /P /R

I don't know anymore what else to try. I recently extended the VHD from 250GB to 500GB (and resized the partition to 500GB in compmgmt.msc), that's the only thing I did to it. The VHD used to be run in vmware Player before we converted it and ran it on vsphere.

Any ideas?

LueTm
  • 101
  • 1
  • 8
  • Try detaching any superfluous hardware from the virtual machine definition (eg: sound cards, USB controller, multiple PS2, etc). – shodanshok Sep 27 '17 at 15:05
  • Did, no improvement :( – LueTm Sep 27 '17 at 15:20
  • Did you try booting your windows server in safety mode? – Mr Zach Oct 02 '17 at 09:01
  • @MrZach Unfortunately we do not get that far into the boot sequence - spamming F8 does nothing, so I guess the system fails before we get to this point. – LueTm Oct 02 '17 at 09:06
  • 1
    I have never seen the black screen with the cursor in a VM but I have seen it on a physical server. The fix, in that case, was the boot order was not correct. – SpiderIce Oct 02 '17 at 13:20
  • 1
    Option 2: would create a new VM server load a 2k3 os and do a full system restore from a working time. That would bypass any VM Bios issues too. – SpiderIce Oct 02 '17 at 13:41
  • @SpiderIce Tried Option 1, did not change anything unfortunately. Will try Option 2 asap. – LueTm Oct 02 '17 at 13:43
  • @SpiderIce How would you do the system restore? Just cp -R everything in a Linux live environment? – LueTm Oct 03 '17 at 09:52
  • @LueTm it depends on the config of the server. If you have a simple config you could just dump the data back once your apps are loaded. If it's a complex config then you should load your backup software and do a full restore and overwrite everything. If you're doing VMware level backups when you restore pick to restore it to a new VM. – SpiderIce Oct 03 '17 at 12:57
  • @SpiderIce We do VM Level backups. No backup we have available (-30d) is working. – LueTm Oct 04 '17 at 05:56
  • Wild guess here: It sounds like the boot mode got flipped from legacy BIOS to UEFI or vice versa. No idea how a host crash would do that though. – Brandon Xavier Oct 04 '17 at 11:56
  • How can I change that up? – LueTm Oct 04 '17 at 14:59
  • Settings, VM Options tab, Boot Options, Firmware (that's from a vCenter 6 - similar path from the standalone client or other versions) – Brandon Xavier Oct 05 '17 at 01:07
  • Ah, that one got my hopes up. Unfortunately it didn't work (switching to EFI from BIOS) . I'm trying to copy all files from a recent backup into a newly installed OS. – LueTm Oct 05 '17 at 05:49
  • 1
    Ah well, it was worth a shot. I fully understand your need to get your system back in service now by restoring into a fresh OS install. However, if you have time for post-mortem troubleshooting you might try fiddling with the BIOS settings with a fresh restore from backup - it's not impossible some of the previous troubleshooting steps might have "fixed" things in the wrong mode (EFI vs. BIOS). Good luck whichever way you go. – Brandon Xavier Oct 05 '17 at 12:03
  • I've ran into this recently as well, but it was caused after an update was applied and got stuck in the middle of applying. The only way I could get it back was to rekick the machine. – hiiambo Oct 08 '17 at 11:47
  • Important Update, see original! – LueTm Oct 11 '17 at 12:27
  • I got it to boot! http://www.hirensbootcd.org/download/ I will post back with more info as soon as I found out more. – LueTm Oct 12 '17 at 07:54

1 Answers1

0

I was able to boot it with Hiren's Boot CD. The server is now usable. I think the MBR and bootsector are currupt and I need to spend some time trying to fix it. But maybe I will just boot the machine with the CD a few times a year.

LueTm
  • 101
  • 1
  • 8