proxmox input/output error

en4ble

Member
Feb 24, 2023
73
5
8
Hi,

Freshly installed os/new hardware. It was transported so it had to be powered down, after that I have noticed access to GUI stopped working on random occasions.

When trying grab any information such as /etc/network/interfaces or even looking at the default systemctl status commands getting input/output error for all of them.

Never encountered this issue before and I did install lots of pves. Is this looking like corrupted OS or hw related? Can't even run fsck which gives same error. Probably will try rebuild but thought I'll ask or atleast if someone has similar issue they can use this thread.

1722480106011.png

Thank You in advance for any feedback.
 
Your PVE cannot even load any programs anymore (no even read only), so there seems to be a problem with the root filesystem or the disk it's on.
I would check the drive and the cables to it.

Maybe try to mount it on other linux computer to see if the drive is still working?
 
Last edited:
Your PVE cannot even load any programs anymore (no even read only), so there seems to be a problem with the root filesystem or the disk it's on.
I would check the drive and the cables to it.

Maybe try to mount it on other linux computer to see if the drive is still working?
Thanks for reply. I mean if the issue was with the cable don't you think it wouldn't even boot instead being impaired?!
 
Thanks for reply. I mean if the issue was with the cable don't you think it wouldn't even boot instead being impaired?!
A bad cable doesn't mean necessarily that it fails continuously. It may work for some time and then drop the connection due to communication errors.

Fact is your root filesystem tries to mount itself as read-only because it detects an error. But in that state you should still be able to read and execute programs from it. This fails completely in your case.
 
As soon as this message shows up (Remounting filesystem read-only) the UI stops working and no info can be captured. Reseated cables, going with new install.
1722619491159.png
 
this is interesting, I have similar issues but on the VMs

I've installed a simple k3s ha cluster - nothing within it, so just control plane, etcd, master - after a period of time, usually within 24hrs the VMs show the IO error
1725952177756.png

all other VMs/LXCs work fine on the same server, even after a reboot, the k3s nodes work fine again for a period of time. They are all setup with cloud-init from an ubuntu 24.04.

the underlying disk is a Samsung 960 ssd or WD BlueSA510 nvme (on a different proxmox server) but the result is the same.
no idea how to troubleshoot the issue but I guess it has to do how pve handles the disk emulation.