SQL Server 2022 database corruption

You'll be pleased to know Windows Server 2022 and SQL server is indeed stable (so far) on the virtstor drivers for 0.1.271 - cloned the machine, tested offline with new "old" drivers and everything stable and no issues even when running maintenance tasks, hammering the databases and set off new replication and backup tasks.

I wonder if as this was stable for many months until triggering SQL database checker issues whether some Windows 2022 update triggered it to work in a way like the already known 2025 issues?

How on earth is this faulty driver still available as the default latest version over 12 months old!
It seems the original signed driver are stale.

Fix was included in https://github.com/virtio-win/kvm-guest-drivers-windows/releases/tag/mm315

"pre-release and untested binary available"
 
Do we need our own virtio signed drivers distribution?
Someone has to maintain it, which means proper testing and release management. If the goal is Microsoft certification/WHQL, that also means establishing the appropriate relationship with Microsoft, going through their certification process, and potentially involving legal and contractual resources. It is certainly possible with an appropriate investment of resources.

If this is simply an unofficial distribution, you can absolutely clone the repository and distribute your own signed builds. However, that introduces a number of risks for users, as you can imagine: trust, signing, update management, compatibility, supportability, and potentially liability.

So the short answer is: sure, it would be nice to have more actively maintained driver. The bigger question is who is willing to own, test, sign, and support the development, or sponsor the resources to do so?


Blockbridge : Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox
 
You'll be pleased to know Windows Server 2022 and SQL server is indeed stable (so far) on the virtstor drivers for 0.1.271 - cloned the machine, tested offline with new "old" drivers and everything stable and no issues even when running maintenance tasks, hammering the databases and set off new replication and backup tasks.
That's good news, but remember that you are using an unsupported storage stack that will sooner or later give you some issue that no one will be willing to investigate / fix. Take your time to rebuild your storage to a proper ZFS setup, which might be as easy as configuring the RAID controller in JBOD/IT mode and creating a RAIDz2 natively with ZFS. Or just replacing the RAID card with an HBA, which should not be expensive.
 
just as suggestion - i am running the following setup:

VM ubuntu 2024.04 LTS with microsoft SQL 2025
disk 1 - for ubuntu only
disk 2 - dedicated data disk for databases
disk 3 - for internal backups, generated from scripts

full backups:
1) every hour internal on the PVE node on nvme ssd
2) every night to external PBS

setup SQL on linux is a little more work, but once ready you get a rock stable high performance SQL server.