Problems with 2.6.32 - system totally frozen

[xkuba]

Hi,

we've tried to upgrade our Proxmox systems to 2.6.32. But we've experienced total system hangups - even the console was not working. Nothing in kern.log. Only hard reset worked.

The systems in question were/are happily running 1.5/1.6 + 2.6.18 without any problems for months. One machine hangs up two times in 15 minutes after reboot. After second hangup we've changed the kernel back to 2.6.18 and it's running well again. Second machine hangs up two times as well but in few hours after reboot. Again, we've changed the kernel back to 2.6.18 and it's running well again.

Some info about machines:

Thare are dozen of OpenVZ virtual machines and at most one KVM. We experienced one hangup when no KVM machine was running at all. The systems are not overloaded.

The boxes are using Intel motherboards DG41TY and DQ45CB, Intel Core 2 Quad proccessors.

Proxmox is installed on dedicated harddisk from 1.5 ISO. There are also other harddisks in the machines which are in software RAID. I know software RAID is not supported. Could this be the issue even when Proxmox is not installed on SW RAID and the SW RAID partitions are bind-mounted into VMs?

We though about some HW problems first but now it doesn't seem likely to us when two machines have problems with 2.6.32...

How can we debug this issue?

Thanks,

Kuba

P.S. This was with 2.6.32 few days ago - not the new version. We'll retest...

 

[tom]

pls always post your 'pveversion -v' to show the detailed versions.

 

[xkuba]

First system with hangups in 15 minutes with 2.6.32 (now it's not installed)

pve-manager: 1.6-2 (pve-manager/1.6/5087)
running kernel: 2.6.18-2-pve
proxmox-ve-2.6.18: 1.5-5
pve-kernel-2.6.18-2-pve: 2.6.18-5
qemu-server: 1.1-18
pve-firmware: 1.0-7
libpve-storage-perl: 1.0-14
vncterm: 0.9-2
vzctl: 3.0.24-1pve4
vzdump: 1.2-7
vzprocps: 2.0.11-1dso2
vzquota: 3.0.11-1
pve-qemu-kvm-2.6.18: 0.9.1-5

Second system, now being upgraded to new 2.6.18 version.

pve-manager: 1.6-2 (pve-manager/1.6/5087)
running kernel: 2.6.18-3-pve
proxmox-ve-2.6.18: 1.6-6
pve-kernel-2.6.18-3-pve: 2.6.18-6
pve-kernel-2.6.18-2-pve: 2.6.18-5
qemu-server: 1.1-18
pve-firmware: 1.0-7
libpve-storage-perl: 1.0-14
vncterm: 0.9-2
vzctl: 3.0.24-1pve4
vzdump: 1.2-7
vzprocps: 2.0.11-1dso2
vzquota: 3.0.11-1
pve-qemu-kvm-2.6.18: 0.9.1-5

 

[tom]

you need to send the versions with problems.

pveversion -v should show this:
http://pve.proxmox.com/wiki/Downloads#Update_a_running_Proxmox_Virtual_Environment_to_1.6

I suggest you try the update/upgrade again on a test environment.