Crashing Host - 3.4.9

[BenB]

Hi,

I am runnnig a three node cluster built from the PROXMOX ISO on Dell R320.

root@pve0:~# pveversion
pve-manager/3.4-9/4b51d87a (running kernel: 2.6.32-39-pve)

One of the nodes is crashing intermitently, 3 times in 48 hours. In the syslog data prior to the crash I see the atached. Please advise on the best way forwards, should I be looking at firmware and BIOS updates? Upgrading to latest version 4.1? Is there an easy upgrade path, these are production servers that I can not afford extended downtime on, can I shuffle VMs around in cluste while half the cluster is 4.1 and the rest is still 3.4. Ok here is the log data attached for the two crashes today.

TIA

Ben

 

[manu]

Hi
In the short term, I would advise you to to install the pve-kernel-3.10.0-15-pve as this somes many problem with KVM.
In the middle term, you should start palnning an upgrade to 4.0, as 3.4 maintenance will be dropped at some point.

 

[spirit]

also, do you have done all bios updated on your poweredge ? (bios updates include cpu microcode updates)