SCIENTIFIC-LINUX-USERS Archives

April 2011

SCIENTIFIC-LINUX-USERS@LISTSERV.FNAL.GOV

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
"Raimondo Giammanco [VKICC]" <[log in to unmask]>
Reply To:
Raimondo Giammanco [VKICC]
Date:
Wed, 6 Apr 2011 17:41:19 +0200
Content-Type:
text/plain
Parts/Attachments:
text/plain (138 lines)
On 04/04/2011 12:48, Alan Bartlett wrote:
> On 4 April 2011 09:59, Raimondo Giammanco [VKICC]<[log in to unmask]>  wrote:
>
>>   I've just tested once again the card with ubuntu, now I'm 100% positive it
>> works (i.e. the e1g44etblk was the only wired port during iperf, eth0 was
>> disconnected).
>>
>>   It works with ubuntu 10.04 server:
>> kernel 2.6.35-24-server
>> igb version 2.1.0-k2
>>
>> It remains to be seen where the problem comes in SL (tested kernel:
>> 2.6.32-71.18.2.el6.x86_64)
>>
>> I'll see if I can compile kernel version 2.6.35 on SL or if the problem is
>> linked to my boot parameters (noapic acpi=off).
> Hi Raimondo,
>
> Thank you for keeping this m/l updated with the results of your
> experiments. I would suggest the following steps:
>
> (1) Check the Red Hat bug tracker for any related issues.
> (2) Review the need for your current boot parameters.
> (3) Build your own testing kernel from the latest stable long-term
> support tarball (linux-2.6.35.12.tar.bz2).
>
> Regards,
> Alan.
Hello Alan,

  I had no time in the last few days to go back to my servers.

  Today I compiled my own kernel, version 2.6.38.2 (I compiled before 
reading your comment about long term kernel support)...

  I booted on one server and the very first test seems encouraging: I 
can boot without the previous flags "acpi=off noapic" and finally the 
pci-e network card work as supposed:

iperf -s (on a bonded interface)
########
------------------------------------------------------------
Server listening on TCP port 5001
TCP window size: 85.3 KByte (default)
------------------------------------------------------------
[  4] local 10.130.0.102 port 5001 connected with 10.130.0.103 port 56568
[ ID] Interval       Transfer     Bandwidth
[  4]  0.0-10.0 sec  2.10 GBytes  1.80 Gbits/sec
[  5] local 10.130.0.102 port 5001 connected with 10.130.0.103 port 56569
[  5]  0.0-10.0 sec  2.11 GBytes  1.81 Gbits/sec
[  4] local 10.130.0.102 port 5001 connected with 10.130.0.103 port 56570
[  4]  0.0-10.0 sec  2.12 GBytes  1.82 Gbits/sec
########

and ifconfig gives

#######
bond3     Link encap:Ethernet  HWaddr 00:00:00:00:02:00
           inet addr:10.130.0.102  Bcast:10.130.0.255  Mask:255.255.255.0
           UP BROADCAST RUNNING MASTER MULTICAST  MTU:1500  Metric:1
           RX packets:9403147 errors:0 dropped:0 overruns:0 frame:0
           TX packets:1716406 errors:0 dropped:0 overruns:0 carrier:0
           collisions:0 txqueuelen:0
           RX bytes:14236223600 (13.2 GiB)  TX bytes:155593304 (148.3 MiB)

eth0      Link encap:Ethernet  HWaddr 00:25:90:1A:62:DE
           inet addr:10.1.0.133  Bcast:10.1.7.255  Mask:255.255.248.0
           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
           RX packets:5637 errors:0 dropped:299 overruns:0 frame:0
           TX packets:397 errors:0 dropped:0 overruns:0 carrier:0
           collisions:0 txqueuelen:1000
           RX bytes:557787 (544.7 KiB)  TX bytes:39951 (39.0 KiB)
           Memory:fafe0000-fb000000

eth8      Link encap:Ethernet  HWaddr 00:00:00:00:02:00
           UP BROADCAST RUNNING SLAVE MULTICAST  MTU:1500  Metric:1
           RX packets:4698471 errors:0 dropped:0 overruns:0 frame:0
           TX packets:858208 errors:0 dropped:0 overruns:0 carrier:0
           collisions:0 txqueuelen:1000
           RX bytes:7113403080 (6.6 GiB)  TX bytes:77796780 (74.1 MiB)
           Memory:f87e0000-f8800000

eth9      Link encap:Ethernet  HWaddr 00:00:00:00:02:00
           UP BROADCAST RUNNING SLAVE MULTICAST  MTU:1500  Metric:1
           RX packets:4704676 errors:0 dropped:0 overruns:0 frame:0
           TX packets:858198 errors:0 dropped:0 overruns:0 carrier:0
           collisions:0 txqueuelen:1000
           RX bytes:7122820520 (6.6 GiB)  TX bytes:77796524 (74.1 MiB)
           Memory:f7fe0000-f8000000

lo        Link encap:Local Loopback
           inet addr:127.0.0.1  Mask:255.0.0.0
           UP LOOPBACK RUNNING  MTU:16436  Metric:1
           RX packets:0 errors:0 dropped:0 overruns:0 frame:0
           TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
           collisions:0 txqueuelen:0
           RX bytes:0 (0.0 b)  TX bytes:0 (0.0 b)
#########

where cat /proc/net/bounding/bond3
#######
Ethernet Channel Bonding Driver: v3.7.0 (June 2, 2010)

Bonding Mode: load balancing (round-robin)
MII Status: up
MII Polling Interval (ms): 100
Up Delay (ms): 0
Down Delay (ms): 0

Slave Interface: eth8
MII Status: up
Speed: 1000 Mbps
Duplex: full
Link Failure Count: 0
Permanent HW addr: 00:1b:21:93:b4:08
Slave queue ID: 0

Slave Interface: eth9
MII Status: up
Speed: 1000 Mbps
Duplex: full
Link Failure Count: 0
Permanent HW addr: 00:1b:21:93:b4:09
Slave queue ID: 0
############

Where eth8 and eth9 are in the second pci-e e1g44 card.

If tomorrow's tests will confirm further, I'll need to run to run a 
custom kernel for the time being, until the el6 kernel are ok with my 
hardware.

I'll see maybe for a kernel with drbd included, that would be really nice.

Regards and thanks for the help.

Raimondo

ATOM RSS1 RSS2