BUG? VE3.2 kernel 2.6.32-28-pve breaks libnl/netlink in host and VMs

Jisi

New Member
Apr 11, 2014
1
0
1
Hello Forum

Basically all the information are already in the header. I just updated a Proxmox installation from VE 3.1to 3.2. I noticed that our software crashed because it was unable to initialize libnl. After some digging I found the reasons is the new proxmox kernel. This can be easily reproduced by installing the libnl-utils package on the host or in a VM, which will also fail.

With the current kernel:

Code:
[COLOR=#4C2F2D][FONT=Courier]# uname --all[/FONT][/COLOR]
[COLOR=#4C2F2D][FONT=Courier]Linux proxmox 2.6.32-28-pve #1 SMP Tue Mar 18 15:21:47 CET 2014 x86_64 GNU/Linux[/FONT][/COLOR]
[COLOR=#4C2F2D][FONT=Courier]# nl-qdisc-list [/FONT][/COLOR]
[COLOR=#4C2F2D][FONT=Courier]Error: Unable to allocate link cache: Input data out of range[/FONT][/COLOR]
[COLOR=#4C2F2D][FONT=Courier]# nl-link-list [/FONT][/COLOR]
[COLOR=#4C2F2D][FONT=Courier]Error: Unable to allocate link cache: Input data out of range[/FONT][/COLOR]

When I rebooted the machine with the previous kernel

Code:
[COLOR=#4C2F2D][FONT=Courier]# uname --all[/FONT][/COLOR]
[COLOR=#4C2F2D][FONT=Courier]Linux proxmox 2.6.32-24-pve #1 SMP Fri Sep 13 07:29:30 CEST 2013 x86_64 GNU/Linux[/FONT][/COLOR]
[COLOR=#4C2F2D][FONT=Courier]# nl-qdisc-list [/FONT][/COLOR]
[COLOR=#4C2F2D][FONT=Courier]qdisc pfifo_fast dev eth1 id none parent root bands 3[/FONT][/COLOR]
[COLOR=#4C2F2D][FONT=Courier]qdisc pfifo_fast dev eth0 id none parent root bands 3[/FONT][/COLOR]
[COLOR=#4C2F2D][FONT=Courier]# nl-link-list [/FONT][/COLOR]
[COLOR=#4C2F2D][FONT=Courier]lo loopback <loopback,up,running,lowerup> [/FONT][/COLOR]
[COLOR=#4C2F2D][FONT=Courier]eth1 ether d8:5d:4c:80:bc:6d <broadcast,multicast,up> [/FONT][/COLOR]
[COLOR=#4C2F2D][FONT=Courier]eth0 ether 1c:6f:65:38:2c:49 <broadcast,multicast,up,running,lowerup> [/FONT][/COLOR]
[COLOR=#4C2F2D][FONT=Courier]vmbr0 ether 1c:6f:65:38:2c:49 <broadcast,multicast,up,running,lowerup> [/FONT][/COLOR]
[COLOR=#4C2F2D][FONT=Courier]vmbr1 ether d8:5d:4c:80:bc:6d <broadcast,multicast,up,running,lowerup> [/FONT][/COLOR]
[COLOR=#4C2F2D][FONT=Courier]venet0 void <broadcast,pointopoint,noarp,up,running,lowerup> [/FONT][/COLOR]
[COLOR=#4C2F2D][FONT=Courier]veth100.0 ether 82:dc:83:cc:c9:d7 <broadcast,multicast,up,running,lowerup> [/FONT][/COLOR]
[COLOR=#4C2F2D][FONT=Courier]veth104.0 ether 96:13:ba:a7:3d:6b <broadcast,multicast,up,running,lowerup> [/FONT][/COLOR]
[COLOR=#4C2F2D][FONT=Courier]veth101.0 ether 16:fd:5d:08:5f:4f <broadcast,multicast,up,running,lowerup> [/FONT][/COLOR]
[COLOR=#4C2F2D][FONT=Courier]veth105.0 ether 96:ce:e6:37:36:43 <broadcast,multicast,up,running,lowerup> [/FONT][/COLOR]
[COLOR=#4C2F2D][FONT=Courier]veth111.0 ether 8a:81:e2:42:5c:57 <broadcast,multicast,up,running,lowerup> [/FONT][/COLOR]
[COLOR=#4C2F2D][FONT=Courier]veth110.0 ether 76:f9:30:fb:ac:30 <broadcast,multicast,up,running,lowerup> [/FONT][/COLOR]

I googled and could not find any information regarding problems with libnl and OpenVZ. Is this problem by intention? Am I the only one affected? Because I guess somebody should have noted as the netlink library is quite common.

Does anybody have any suggestions or pointers?

Thanks
 

About

The Proxmox community has been around for many years and offers help and support for Proxmox VE, Proxmox Backup Server, and Proxmox Mail Gateway.
We think our community is one of the best thanks to people like you!

Get your subscription!

The Proxmox team works very hard to make sure you are running the best software and getting stable updates and security enhancements, as well as quick enterprise support. Tens of thousands of happy customers have a Proxmox subscription. Get yours easily in our online shop.

Buy now!