[SOLVED] Virtio drivers Windows 2019 crashing

M

mrhill741

New Member
May 30, 2024
7
3
3
I'm running a Veeam VM using Windows Server 2019 with virtio drivers 0.1.271 in Proxmox 8.4.17. About 2 weeks ago the vm started rebooting after a bug check while backups were running causing a heavy i/o load. The server has been stable until recently. I have 3 PVE hosts all on the same version that I have migrated this vm to see if it was a host problem, but it happens on any host. Hosts are Dell PowerEdge R550 using HBA and SAS drives in ZFS RAID10.

memoru.dmp shows:
FILE_IN_CAB: MEMORY.DMP

BUGCHECK_CODE: 7a

BUGCHECK_P1: ffffeec247063388

BUGCHECK_P2: ffffffffc0000185

BUGCHECK_P3: 2000a957e880

BUGCHECK_P4: ffff848e0c671130

ERROR_CODE: (NTSTATUS) 0xc0000185 - The I/O device reported an I/O error.

IMAGE_NAME: dump_vioscsi.sys

MODULE_NAME: dump_vioscsi

FAULTING_MODULE: fffff8007b170000 dump_vioscsi

DISK_HARDWARE_ERROR: There was error with disk hardware

I can't find any known issues with the version of virtio drivers I'm running on this vm that show a problem under heavy load. I've seen this mentioned on older versions of the virtio drivers. Are there any adjustments I can make to prevent the vm from rebooting under heavy load? I haven't had a problem with this vm until recently. It's been very stable until now.
 
Yes, Veeam backup 13 uses postgresql but Veeam One installed on the same server uses MSSQL Express.
 
Did you already try an older version of the VirtIO drivers? .226 is more stable for 2019.
 
Downgraded to 0.266 and it seems to be stable during heavy i/o. I'll have to see what happens over the next week. Thanks for the info. I was seeing that 0.271 was stable too but this is the first server to upgrade the virtio drivers on.
 
_gabriel said:
HDD or SSD ?
How many per host ?
Click to expand...
Dell PE R550 with 8 SAS drives - ZFS RAID10. I have 15 Windows VM's all using virtio drivers v0.266. The server having issues is the only one on 0.271, and being a Veeam backup server, does get pretty high disk and network i/o. The host shows very low cpu i/o delay and no other vm's are affected on the same host. So far after downgrading drivers to v0.266 I haven't seen the crazy high spikes or reboots yet.
 
I think I can close this out. Downgrading virtio drivers from 0.271 to 0.266 seems to have solved the main issues of random reboots/crashing.

Another side effect is the network transfers are more stable over the full Veeam backup window. I had not realized the network transfers were slowing over time before now they are solid for the entire backup window.
 
  • Like
Reactions: cwt and SteveITS
You must log in or register to reply here.
Top Bottom
Back