New 2.6.32 Kernel for Proxmox VE 2.1 stable

I have a Chasis MFSYS25V2 with two MFS5520VI Compute Module. Each module have two Xeon CPU E5645 at 2.40GHz with 32GB of RAM (4x8GB) and with Dual Gigabit Ethernet I/O Expansion Mezzanine Card

With the kernel 2.6.32-11 I have problems with flapping network.

The problem with the network is only in the bond1 interface that have VLAN configuration. bond0 always work.

Hi!

Which controller? Anything like this?

01:00.0 Ethernet controller: Intel Corporation 82575EB Gigabit Network Connection (rev 02)
01:00.1 Ethernet controller: Intel Corporation 82575EB Gigabit Network Connection (rev 02)

it's an issue within igb drivers in version 3.4.7. Downgrading to 3.3.6 resolves the problem.
install (with dpkg -i filename) http://uploaded.to/file/673acyyl and update initrd (update-initramfs -k all -u)

regards,
patrick
 
Last edited:
I have 3x MFS5520VI Compute Module in IMS .
Replacing the driver from version 3.4.7 to 3.3.6 igb.ko solves this problem.

The problem exists only in the Linux kernel proxmox2-1 2.6.32-12-pve and IGB driver 3.4.7 and bridged vlans.
 
I tried to upgrade but no luck:

apt-get dist-upgrade
Reading package lists... Done
Building dependency tree
Reading state information... Done
Calculating upgrade... Done
0 upgraded, 0 newly installed, 0 to remove and 0 not upgraded.

Here is my version:

pveversion -v
pve-manager: 2.1-1 (pve-manager/2.1/f9b0f63a)
running kernel: 2.6.32-11-pve
proxmox-ve-2.6.32: 2.0-66
pve-kernel-2.6.32-11-pve: 2.6.32-66
lvm2: 2.02.95-1pve2
clvm: 2.02.95-1pve2
corosync-pve: 1.4.3-1
openais-pve: 1.1.4-2
libqb: 0.10.1-2
redhat-cluster-pve: 3.1.8-3
resource-agents-pve: 3.9.2-3
fence-agents-pve: 3.1.7-2
pve-cluster: 1.0-26
qemu-server: 2.0-39
pve-firmware: 1.0-15
libpve-common-perl: 1.0-27
libpve-access-control: 1.0-21
libpve-storage-perl: 2.0-18
vncterm: 1.0-2
vzctl: 3.0.30-2pve5
vzprocps: 2.0.11-2
vzquota: 3.0.12-3
pve-qemu-kvm: 1.0-9
ksm-control-daemon: 1.1-1
 
I tried to upgrade but no luck:

apt-get dist-upgrade
Reading package lists... Done
Building dependency tree
Reading state information... Done
Calculating upgrade... Done
0 upgraded, 0 newly installed, 0 to remove and 0 not upgraded.

Here is my version:

pveversion -v
pve-manager: 2.1-1 (pve-manager/2.1/f9b0f63a)
running kernel: 2.6.32-11-pve
proxmox-ve-2.6.32: 2.0-66
pve-kernel-2.6.32-11-pve: 2.6.32-66
lvm2: 2.02.95-1pve2
clvm: 2.02.95-1pve2
corosync-pve: 1.4.3-1
openais-pve: 1.1.4-2
libqb: 0.10.1-2
redhat-cluster-pve: 3.1.8-3
resource-agents-pve: 3.9.2-3
fence-agents-pve: 3.1.7-2
pve-cluster: 1.0-26
qemu-server: 2.0-39
pve-firmware: 1.0-15
libpve-common-perl: 1.0-27
libpve-access-control: 1.0-21
libpve-storage-perl: 2.0-18
vncterm: 1.0-2
vzctl: 3.0.30-2pve5
vzprocps: 2.0.11-2
vzquota: 3.0.12-3
pve-qemu-kvm: 1.0-9
ksm-control-daemon: 1.1-1


I managed to upgrade howver I am seeing two kernels (see below). How do I enable the latest Kernel?
I am sorry in advance that it seems like a simple question but the wiki pages and how to 's do not explain clearly.

pveversion -v
pve-manager: 2.1-1 (pve-manager/2.1/f9b0f63a)
running kernel: 2.6.32-12-pve
proxmox-ve-2.6.32: 2.1-68
pve-kernel-2.6.32-11-pve: 2.6.32-66
pve-kernel-2.6.32-12-pve: 2.6.32-68
lvm2: 2.02.95-1pve2
clvm: 2.02.95-1pve2
corosync-pve: 1.4.3-1
openais-pve: 1.1.4-2
libqb: 0.10.1-2
redhat-cluster-pve: 3.1.8-3
resource-agents-pve: 3.9.2-3
fence-agents-pve: 3.1.7-2
pve-cluster: 1.0-26
qemu-server: 2.0-39
pve-firmware: 1.0-16
libpve-common-perl: 1.0-27
libpve-access-control: 1.0-21
libpve-storage-perl: 2.0-18
vncterm: 1.0-2
vzctl: 3.0.30-2pve5
vzprocps: 2.0.11-2
vzquota: 3.0.12-3
pve-qemu-kvm: 1.0-9
ksm-control-daemon: 1.1-1
 
I've strange behaviour of Proxmox 2.1 kernel 2.6.32-12. I'm using this version of Proxmox since Thursday last week. I've got Kernel Panic, detailed message looks like: Kernel Panic! Kernel not support with 'relatime'. In dmesg and syslog there is no info about this event.

pveversion -v:
pve-manager: 2.1-1 (pve-manager/2.1/f9b0f63a)
running kernel: 2.6.32-12-pve
proxmox-ve-2.6.32: 2.1-68
pve-kernel-2.6.32-11-pve: 2.6.32-66
pve-kernel-2.6.32-12-pve: 2.6.32-68
lvm2: 2.02.95-1pve2
clvm: 2.02.95-1pve2
corosync-pve: 1.4.3-1
openais-pve: 1.1.4-2
libqb: 0.10.1-2
redhat-cluster-pve: 3.1.8-3
resource-agents-pve: 3.9.2-3
fence-agents-pve: 3.1.7-2
pve-cluster: 1.0-26
qemu-server: 2.0-39
pve-firmware: 1.0-16
libpve-common-perl: 1.0-27
libpve-access-control: 1.0-21
libpve-storage-perl: 2.0-18
vncterm: 1.0-2
vzctl: 3.0.30-2pve5
vzprocps: 2.0.11-2
vzquota: 3.0.12-3
pve-qemu-kvm: 1.0-9
ksm-control-daemon: 1.1-1


On the server are working few OpenVZ containers a one KVM VM with linux Debian Squeeze.
 
maybe the kernel panic was triggered by the leap second in the last days
i've read some reports that linux had problems with that
 
I am getting random freezes with pve-kernel-2.6.32-12-pve every couple of hours requiring a reboot. It's a 3 year old MSI MS-9232 1 RU server (e1000 nic Intel 82573L) which I use for testing. I reinstalled with pve-kernel-2.6.32-11-pve and it seems stable again. With the pve-kernel-2.6.32-12-pve kernel it will eventually lock up just sitting there doing nothing. When I do lsmod I don't see the igb.lo loaded.
 
Last edited:
I am also having problems with Intel NIC, in my case a Quad Port ET based on 82576.

On POST link led is on and it remains on up until GRUB passes the stick to the kernel to boot. Immediately after the kernel starts to load the link goes off and I cannot get the NIC back (rmmod igb followed by modprobe igb do not help).

No evidence of problems on dmesg. This issue has been reported on other platforms as well as with other Intel NICs here (description of the problem matches exactly what I am seeing):
https://sourceforge.net/mailarchive/message.php?msg_id=29316891

I see a lot of users reporting problems with Intel NICs here. Maybe these issues are all related.
 
A user reported that downgrading to the 3.3.6 release solved his problems. Maybe it works for you too. Have a look here, post #22
 
Just want to say that proxmox rox my sox (first post).

I had two Proxmox servers experience a kernel panic on Saturday as well. They were both older Dell 2950's and I originally thought it may be due to a needed bios update, but again only 2 out of 10 servers panicked. All of my servers are running the 2.6.32-12 kernel. These servers have been humming along perfectly for weeks on 2.1 until the leap second event...

IMAG0064.jpg
 
3.3.6 did not work for me either. I am in touch with Intel, seems to be a known (although rare) problem.

A user reported that downgrading to the 3.3.6 release solved his problems. Maybe it works for you too. Have a look here, post #22
 
Hello

Today I made ​​a fresh install of Proxmox 2.0 on a Intel Modular Server, and when I upgraded the kernel to version 2.6.32-12 the network has stopped working.

I restarted with the 2.6.32-11 kernel and the network has become operational. I think it's a bug.

Someone with an IMS could confirm this?

Same here. Networking is working in principle, but not with tagged vlans under kernel 2.6.32-12. With version 2.6.32-11 everything works just fine on the IMS.
 
Hi,

I experience a similar effect of up to 72% package loss on a direct link between two proxmox nodes with kernel 2.6.32-12-pve after upgrading today (previous kernel was 2.6.32-11-pve). The two machines are directly linked via bonded (rr) 10 Gb ethernet cards, two Intel Corporation 82599EB 10 Gigabit TN Network Connection per machine. The cards use the ixgbe driver.

After reading this thread I tried unloading the ixgbe module and force-loading the ixgbe.ko from kernel 2.6.32-11-pve on both machines and after ifup-ing the corresponding bond interfaces, there were no more package losses.

This may be related to what was observed with the ixbge in this thread https://bugzilla.redhat.com/show_bug.cgi?id=696337 (for a RedHat based system).

My question is: Will there be a new proxmox 2.6.32-xx-pve kernel package which fixes this issue by patching this driver, ie importing an upstream version or backporting it to what was used in 2.6.32-11-pve? I know that I can manually force it to load the older ixgbe.ko which is still on the system, but even symlinks to the old version of ixbge.ko from /lib/modules/2.6.32-12-pve/kernel/drivers/net/ixgbe to /lib/modules/2.6.32-11-pve/kernel/drivers/net/ixgbe won't help since the module won't be inserted unless forced by modprobe -f. Any advise on how to handle this in an automatic manner?

Thanks for your help.
 
Hi,

just for the record: I solved the problem the following way:
a) Get ixgbe-3.9.17.tar.gz from sourceforge and unpack it somewhere
b) Obtain the patch named ixgbe_fix_mac_flush.patch from an attachment in the following thread:
http://sourceforge.net/mailarchive/...FFDD51D.2080401@gt.net&forum_name=e1000-devel
c) Compile and install the kernel module (follow instructions in the ixgbe-3.9.17/README file.
Note: We need the appropriate kernel source package being installed, ie pve-headers-2.6.32-12-pve.
Then the Intel 82599EB 10 Gb card with ixgbe as driver module works again as part of of bonding device without any package losses, even after rebooting the machines.

Maybe it would be possible to introduce such a fix in the proxmox kernel package. I also tried the ixbge-3.10.16 driver, but it showed the same symptoms.
(The patch for 3.9.17 mentioned above was posted later than the 3.10.16 driver's release date.)

Thanks again.

Hi,

I experience a similar effect of up to 72% package loss on a direct link between two proxmox nodes with kernel 2.6.32-12-pve after upgrading today (previous kernel was 2.6.32-11-pve). The two machines are directly linked via bonded (rr) 10 Gb ethernet cards, two Intel Corporation 82599EB 10 Gigabit TN Network Connection per machine. The cards use the ixgbe driver.

After reading this thread I tried unloading the ixgbe module and force-loading the ixgbe.ko from kernel 2.6.32-11-pve on both machines and after ifup-ing the corresponding bond interfaces, there were no more package losses.

This may be related to what was observed with the ixbge in this thread https://bugzilla.redhat.com/show_bug.cgi?id=696337 (for a RedHat based system).

My question is: Will there be a new proxmox 2.6.32-xx-pve kernel package which fixes this issue by patching this driver, ie importing an upstream version or backporting it to what was used in 2.6.32-11-pve? I know that I can manually force it to load the older ixgbe.ko which is still on the system, but even symlinks to the old version of ixbge.ko from /lib/modules/2.6.32-12-pve/kernel/drivers/net/ixgbe to /lib/modules/2.6.32-11-pve/kernel/drivers/net/ixgbe won't help since the module won't be inserted unless forced by modprobe -f. Any advise on how to handle this in an automatic manner?

Thanks for your help.
 

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!