New Proxmox VE kernel - 2.6.32 (pvetest)

[martin]

We just released updates for our default 2.6.32 kernel to the pvetest repository. Everybody is encouraged to test and give feedback before we move it to the stable repository.

Release notes:

- pve-kernel-2.6.32 (2.6.32-34)

• update to debian squeeze kernel 2.6.32-35
• update ARECA RAID driver (including support for 1880 series)
• enable HP Smart Array SCSI driver SCSI_HPSA=m
• enable PMC-Sierra SPC 8001 SAS/SATA SCSI_PM8001=m
• enable Brocade 1010/1020 10Gb Ethernet BNA=m
• add Brocade firmware
• update bnx2 drivers
• update e1000e driver to 1.3.17
• update igb to 3.0.22
• update ixgbe to 3.3.9
• fix applydebpatches.pl: include ${rev}squeeze1 series patches

About our kernel branches:
Proxmox VE Kernels
__________________
Best regards,
Martin Maurer

 

[screenie]

hohooo.... i was waiting for the areca drivers since 8 months

big thanks!

 

[chipitsine]

hello, we are experiencing problem with Intel NIC on that kernel, where should I report a bug ?

 

[dietmar]

hello, we are experiencing problem with Intel NIC on that kernel, where should I report a bug ?

Please use this forum to report bugs.

 

[chipitsine]

Please use this forum to report bugs.

ok.

NIC is (lspci -v):

07:00.0 Ethernet controller: Intel Corporation 82571EB Gigabit Ethernet Controller (rev 06)
Subsystem: Intel Corporation Device 346e
Flags: bus master, fast devsel, latency 0, IRQ 53
Memory at 98960000 (32-bit, non-prefetchable) [size=128K]
Memory at 98940000 (32-bit, non-prefetchable) [size=128K]
I/O ports at 3020
Expansion ROM at 99200000 [disabled] [size=128K]
Capabilities: [c8] Power Management version 2
Capabilities: [d0] Message Signalled Interrupts: Mask- 64bit+ Queue=0/0 Enable+
Capabilities: [e0] Express Endpoint, MSI 00
Capabilities: [100] Advanced Error Reporting <?>
Capabilities: [140] Device Serial Number b4-cf-e5-ff-ff-23-04-00
Kernel driver in use: e1000e
Kernel modules: e1000e



after upgrading to 2.6.32-5-pve network is not functional (works with 2.6.32-4-pve well).
after link up/link down on switch network works again.

we are using 4 nics in failover bonding (active nic is I described).

 

[chipitsine]

looks like that bug: http://lists.debian.org/debian-kernel/2011/04/msg01016.html

 

[chipitsine]

wouldn't be better to deploy trac.proxmox.com ?
I have experience in deploying and supporting Trac, if you are interested.

 

[tom]

wouldn't be better to deploy trac.proxmox.com ?
I have experience in deploying and supporting Trac, if you are interested.

thanks for the offer, but we already use Bugzilla for bug tracking. currently only active developers (internal and external) have access, but we plan to open it more for the 2.0 series, soon (see also http://pve.proxmox.com/wiki/Roadmap#Roadmap_for_2.x)

 

[tom]

I remember an similar issue with on board Intel Nics (network sometimes down after reboot) - after flashing a new mainboard bios (Intel S3420GP) the issue seems resolved. maybe it helps also here.? we did not use bonding, just a plain 1 gbit link.

 

[chipitsine]

I remember an similar issue with on board Intel Nics (network sometimes down after reboot) - after flashing a new mainboard bios (Intel S3420GP) the issue seems resolved. maybe it helps also here.? we did not use bonding, just a plain 1 gbit link.

no, BIOS update didn't help.

 

[chipitsine]

I noticed about Roadmap, but I was not sure whether it is new or existing bug tracking system.

 

[dietmar]

Just upload a new kernel wit updated e1000e driver. Please can you test:

ftp://download.proxmox.com/debian/d...4/pve-kernel-2.6.32-5-pve_2.6.32-36_amd64.deb

 

[chipitsine]

Just upload a new kernel wit updated e1000e driver. Please can you test:

ftp://download.proxmox.com/debian/d...4/pve-kernel-2.6.32-5-pve_2.6.32-36_amd64.deb

still doesn't work

 

[dietmar]

still doesn't work

Is there a way to reproduce the bug?

 

[chipitsine]

Is there a way to reproduce the bug?

it reproduces on certain type of motherboards.
I'm afraid, we cannot put such server in Internet, it is located in our LAN.

we can help in any kind. I'll take a look at patches and will try to recompile kernel myself.

 

[dietmar]

I'll take a look at patches and will try to recompile kernel myself.

I just uploaded the actual kernel sources to:

ftp://download.proxmox.com/sources/pve-kernel-2.6.32_2011-08-10.tar.gz

We use the driver from http://sourceforge.net/projects/e1000/ because many users reported unstable behaviour with the debian e1000e driver.

It would be great if you help to debug that.

 

[chipitsine]

on the next week 3 servers of that kind become spare and I'll play with compiling kernel.

 

[chipitsine]

so

I just uploaded the actual kernel sources to:

ftp://download.proxmox.com/sources/pve-kernel-2.6.32_2011-08-10.tar.gz

We use the driver from http://sourceforge.net/projects/e1000/ because many users reported unstable behaviour with the debian e1000e driver.

It would be great if you help to debug that.

there two kind of NICs:

82571EB (good)
80003ES2LAN (bad)

after all experiments...

1) there're bonding problem with mixing Intel 82751 with other Intel NICs
2) that problem exists in all kernels, except new uploaded 2.6.32 with sf's e1000e drivers (well, I updated 1.4.4 with 1.5.1)
3) 2.6.18-6 kernel is affected with that problem, someone, please describe 2.6.18 build environment ?


yes, new uploaded kernel with different e1000e drivers is good

 

[dietmar]

3) 2.6.18-6 kernel is affected with that problem, someone, please describe 2.6.18 build environment ?

see Makefile

yes, new uploaded kernel with different e1000e drivers is good

so version 1.4.4 works?

 

[chipitsine]

see Makefile



so version 1.4.4 works?

I replaced 1.4.4 with 1.5.1 and I can say 1.5.1 works well.

Makefile in 2.6.18 didn't say anything on OS version, I tried to build 2.6.18 under debian 6.0 and ubuntu 10.10, no luck, many strange compilation errors, which say me, I've chosen wrong build environment.