Windows Server 2025 Black Screen

H

h4rm0n!0u54

New Member
Jul 14, 2025
8
0
1
Hi All,

We have a server that after a reboot just went to a black screen within Proxmox 9.1 I've since moved the server over thinking it was something flaky with that cluster since the NFS SAN kept disconnecting and we were having issues with cameras not connecting after an immediate shutdown due to the loss. Its not in our non-production environment for IT consumption since we have to have a few apps on premise but wanted something for Dev/Testing too - anywho

So far i've done the following:

  • The Proxmox boot logo appears, and Windows Recovery sometimes loads, but normal Windows startup results in a black screen.
  • VM configuration: OVMF/UEFI, Q35 9.0 machine type, x86-64-v2 CPU, 8 cores, and approximately 32 GB RAM.
  • Storage uses VirtIO SCSI Single, with a 120 GB OS disk and a 10.5 TB data disk on Tintri NFS storage.
  • Tried both Standard VGA and SPICE/QXL display settings—no change.
  • Detached the 10.5 TB data disk to isolate the OS disk—the VM still does not boot successfully.
  • Checked boot order: the 120 GB OS disk (scsi0) is first.
  • Checked UEFI Boot Manager: Windows Boot Manager is present and points to \EFI\Microsoft\Boot\bootmgfw.efi.
  • Accessed Windows Recovery and worked through loading the VirtIO SCSI driver and assigning drive letters to access the disks.
  • Checked for the Startup Repair log (SrtTrail.txt), but it does not exist.
  • Inspected the EFI partition’s BCD store. The installed Windows Server loader is listed as {default}, with device and osdevice pointing to C: and the loader path set to \Windows\system32\winload.efi.
  • Confirmed that the separate recovery loader has its own GUID.
  • Attempted to force Safe Mode through BCDEdit but encountered “Element not found.” Identified {default} as the correct installed Windows entry; successful Safe Mode boot has not yet been confirmed.

qemu config:

qm config 100
agent: 1
bios: ovmf
boot: order=scsi0;ide0;ide2
cores: 8
cpu: x86-64-v3
efidisk0: tintri:100/vm-100-disk-0.qcow2,efitype=4m,ms-cert=2023k,pre-enrolled-keys=1,size=528K
ide0: tintri:iso/SW_DVD9_Win_Server_STD_CORE_2025_24H2.6_64Bit_English_DC_STD_MLF_X24-02024.ISO,media=cdrom,size=6625184K
ide2: tintri:iso/virtio-win-0.1.285.iso,media=cdrom,size=771138K
machine: pc-q35-9.0
memory: 32000
meta: creation-qemu=9.2.0,ctime=1759238479
net0: virtio=BC:24:11:6D:67:4D,bridge=vmbr0
net1: virtio=BC:24:11:FD:8C:1F,bridge=vmbr0,link_down=1,tag=46
numa: 0
ostype: win11
scsi0: tintri:100/vm-100-disk-1.qcow2,iothread=1,size=120G
scsihw: virtio-scsi-pci
smbios1: uuid=55c8dc40-a351-4e62-b958-e6d2ccd3495a
sockets: 1
spice_enhancements: videostreaming=all
tpmstate0: tintri:100/vm-100-disk-0.raw,size=4M,version=v2.0
unused0: tintri:100/vm-100-disk-2.qcow2
vmgenid: 533ca41b-cda7-493d-b551-81520ddb43e4
 
h4rm0n!0u54 said:
Hi All,

We have a server that after a reboot just went to a black screen within Proxmox 9.1 I've since moved the server over thinking it was something flaky with that cluster since the NFS SAN kept disconnecting and we were having issues with cameras not connecting after an immediate shutdown due to the loss. Its not in our non-production environment for IT consumption since we have to have a few apps on premise but wanted something for Dev/Testing too - anywho

So far i've done the following:

  • The Proxmox boot logo appears, and Windows Recovery sometimes loads, but normal Windows startup results in a black screen.
  • VM configuration: OVMF/UEFI, Q35 9.0 machine type, x86-64-v2 CPU, 8 cores, and approximately 32 GB RAM.
  • Storage uses VirtIO SCSI Single, with a 120 GB OS disk and a 10.5 TB data disk on Tintri NFS storage.
  • Tried both Standard VGA and SPICE/QXL display settings—no change.
  • Detached the 10.5 TB data disk to isolate the OS disk—the VM still does not boot successfully.
  • Checked boot order: the 120 GB OS disk (scsi0) is first.
  • Checked UEFI Boot Manager: Windows Boot Manager is present and points to \EFI\Microsoft\Boot\bootmgfw.efi.
  • Accessed Windows Recovery and worked through loading the VirtIO SCSI driver and assigning drive letters to access the disks.
  • Checked for the Startup Repair log (SrtTrail.txt), but it does not exist.
  • Inspected the EFI partition’s BCD store. The installed Windows Server loader is listed as {default}, with device and osdevice pointing to C: and the loader path set to \Windows\system32\winload.efi.
  • Confirmed that the separate recovery loader has its own GUID.
  • Attempted to force Safe Mode through BCDEdit but encountered “Element not found.” Identified {default} as the correct installed Windows entry; successful Safe Mode boot has not yet been confirmed.

qemu config:

qm config 100
agent: 1
bios: ovmf
boot: order=scsi0;ide0;ide2
cores: 8
cpu: x86-64-v3
efidisk0: tintri:100/vm-100-disk-0.qcow2,efitype=4m,ms-cert=2023k,pre-enrolled-keys=1,size=528K
ide0: tintri:iso/SW_DVD9_Win_Server_STD_CORE_2025_24H2.6_64Bit_English_DC_STD_MLF_X24-02024.ISO,media=cdrom,size=6625184K
ide2: tintri:iso/virtio-win-0.1.285.iso,media=cdrom,size=771138K
machine: pc-q35-9.0
memory: 32000
meta: creation-qemu=9.2.0,ctime=1759238479
net0: virtio=BC:24:11:6D:67:4D,bridge=vmbr0
net1: virtio=BC:24:11:FD:8C:1F,bridge=vmbr0,link_down=1,tag=46
numa: 0
ostype: win11
scsi0: tintri:100/vm-100-disk-1.qcow2,iothread=1,size=120G
scsihw: virtio-scsi-pci
smbios1: uuid=55c8dc40-a351-4e62-b958-e6d2ccd3495a
sockets: 1
spice_enhancements: videostreaming=all
tpmstate0: tintri:100/vm-100-disk-0.raw,size=4M,version=v2.0
unused0: tintri:100/vm-100-disk-2.qcow2
vmgenid: 533ca41b-cda7-493d-b551-81520ddb43e4
Click to expand...
I've seen similar issues on both Microsoft Hyper-V and Proxmox VE and am not sure what it's exactly related to, but it frequently (not always; sometimes it's even Out of the Box; it seems like a lottery) happens after installing a Microsoft Update. Booting into safe mode and removing the update usually causes the virtual machine to work again (if it's right after install, the install is f'ed). We've done this on both Skylake Server and Cascade Lake Server (where the Cascade Lake processors are supported) processors and have this issue. It's not with all Microsoft Windows Server 2025 virtual machines, only with some (and some of them do not have heavy I/O and have very simple workloads like a basic file server). I've also tried various versions of the VirtIO (0.271, 0.285 and 0.302) tools (even without), but it doesn't seem to matter.

Microsoft Windows Server 2025 is just not a very reliable operating system in general in my opinion. I've re-installed most of these virtual machines with Microsoft Windows Server 2022 and all issues were gone.

See: https://www.reddit.com/r/WindowsServer/comments/1l2rizi/black_screen_after_login/ and https://www.reddit.com/r/sysadmin/comments/1lvo8uz/server_2025_guest_on_hyperv_black_screen_on_boot/ and many other Reddit threads and internet posts for details.
 
Updating the post that we saved our bacon by exporting the .cow2 file and unzipping it. Thankfully all the data needed to re-install the software was on there with the license information. Now we have a replacement server up with our new backup system on it.
 
You must log in or register to reply here.
Top Bottom
Back