Windows Server 2022 reports disk errors on ceph volume

[lifeboy]

We installed a new Windows server 2022 on a cluster that uses an SSD-based ceph volume. All seems to be going well, when suddenly windows event log reports:

"An error was detected on device \Device\Harddisk0\DR0 during a paging operation"

It's Windows error # 51

There are other Windows servers (but older 2016 versions) running on the cluster without a single report of this happening. They're also using the same ceph cluster.

The machine is configurated with scsi-virtio single and the driver are installed, as well as the KVM-guest-agent.

Could this be a peculiarity with Window Server 2022? Any other ideas on why this would be happening?

PS. I checked the https://pve.proxmox.com/wiki/Windows_2022_guest_best_practices page and updated the Virtio driver to the latest version now, which might have been the cause. I was on 0.1.229 and now have 0.1.240. I'll update if that solved this.

 

[lifeboy]

We installed a new Windows server 2022 on a cluster that uses an SSD-based ceph volume. All seems to be going well, when suddenly windows event log reports:

"An error was detected on device \Device\Harddisk0\DR0 during a paging operation"

It's Windows error # 51

There are other Windows servers (but older 2016 versions) running on the cluster without a single report of this happening. They're also using the same ceph cluster.

The machine is configurated with scsi-virtio single and the driver are installed, as well as the KVM-guest-agent.

Could this be a peculiarity with Window Server 2022? Any other ideas on why this would be happening?

PS. I checked the https://pve.proxmox.com/wiki/Windows_2022_guest_best_practices page and updated the Virtio driver to the latest version now, which might have been the cause. I was on 0.1.229 and now have 0.1.240. I'll update if that solved this.

The driver update made no difference.

However, the scsi-virtio driver allows thin provisioning of the disk volume, which is what we use, so Windows starts a defrag (Edit: it's actually an "optimization") once a week by default and it uses 75% of the available RAM. When some user also log on, the system runs out of RAM.

Are there known issues with thin-provisioned storage systems and Windows defrag?

 

[bbgeek17]

We installed a new Windows server 2022 on a cluster that uses an SSD-based ceph volume

so Windows starts a defrag once a week by default

What a throwback to early 90ies.. Defragging SSD is backwards boondoggle, doing so on a disk that sits on top of Ceph - doubly so.
I am not aware of a modern Windows doing "defrag by default". I dont think even Windows 95 did it by default. If its something you added, I suggest removing it.

The "page" error is an indication disk problem, possibly that your storage is too slow, or, given the virtualization layers and memory issues, perhaps overall system performance is not great.

---

Blockbridge : Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox

 

[lifeboy]

What a throwback to early 90ies.. Defragging SSD is backwards boondoggle, doing so on a disk that sits on top of Ceph - doubly so.
I am not aware of a modern Windows doing "defrag by default". I dont think even Windows 95 did it by default. If its something you added, I suggest removing it.

Sorry, it's now called retrim, which differs from defrag. However, it still kills the machine when doing it on a 1TB drive.

The "page" error is an indication disk problem, possibly that your storage is too slow, or, given the virtualization layers and memory issues, perhaps overall system performance is not great.

It's NVMe storage, can't be too slow. It doesn't happen on any other machine, only on the new WS 2022 installation.

---

Blockbridge : Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox

 

[bbgeek17]

https://forum.proxmox.com/threads/windows-vm-requires-lots-of-ram-to-trim-larger-hard-disks.96288/

---

Blockbridge : Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox

 

[IT Signaturit]

Hi, we are using proxmox since 1.8 version (many years!)
now we are using proxmox 7.4.3 and we have the problem reported by lifeboy in this case is happening to us with Windows Server 2019 virtual machines.

We check physical raid disks and all of them are ok, it seems happening something at virtual hard drives

Help!!