Hi,

on 5/20/08 9:53 PM Charles Steinkuehler wrote:
> bonding.txt from the kernel documentation and found the following for
> mode=1:
> 
> ~  In bonding version 2.6.2 or later, when a failover
> ~  occurs in active-backup mode, bonding will issue one
> ~  or more gratuitous ARPs on the newly active slave.
<snip>
> That could for sure explain why a newer kernel might appear to work

Thanks for pointing that. The excerpt explains the behaviour partly.

I played with the setup quite a time and found that the problematic 
behaviour relates to the particular network card or to its drivers (e100 
and eepro100).

I tried following combinations:

1) I changed primary=eth0 in bonding module options.
2) I changed primary=eth1 and loaded e100 and eepro100 driver as eth1.
3) I changed the physical interface of eth1 with driver e100 onto 
another switch.
4) The above experiments lead to experiment with the e100 and eepro100 
drivers only.

What I found is that e100 works when no vlans are set over the 
active-backup bonding interface. eepro100 driver works not at all or 
partly. This is what kernel prints right after enslaving eth0 with 
eepro100 driver.

# ifenslave bond0 eth0

# dmesg |tail
bonding: bond0: enslaving eth0 as an active interface with an up link.
bonding: bond0: link status down for active interface eth0, disabling it

# ethtool eth0
Settings for eth0:
         Supported ports: [ TP MII ]
         Supported link modes:   10baseT/Half 10baseT/Full
                                 100baseT/Half 100baseT/Full
         Supports auto-negotiation: Yes
         Advertised link modes:  10baseT/Half 10baseT/Full
                                 100baseT/Half 100baseT/Full
         Advertised auto-negotiation: Yes
         Speed: 100Mb/s
         Duplex: Full
         Port: MII
         PHYAD: 1
         Transceiver: internal
         Auto-negotiation: on
         Current message level: 0x000020c1 (8385)
         Link detected: yes


I checked the bonding.txt documentation and tried bonding option 
use_carrier=0 and miimon link monitor, but those haven't any effect on 
either eepro100 or e100 driver.

For anyone who might find above information valuable the hardware was 
"02:08.0 Ethernet controller: Intel Corporation 82801DB PRO/100 VE (LOM) 
Ethernet Controller (rev 81)" as reported by lspci.

As I mentioned earlier about my working active-backup bonding setup, 
there the e100 driver version is 3.5.23 and kernel version the latest 
stable 2.6.25.4, but the bonding setup was made somewhere around kernel 
version 2.6.2x. The hardware in this particular setup is "00:02.0 
Ethernet controller: Intel Corporation 82557/8/9 [Ethernet Pro 100] (rev 
08)" as printed by lspci on Debian Etch.

Best regards,
Erkki

-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
------------------------------------------------------------------------
leaf-user mailing list: leaf-user@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/leaf-user
Support Request -- http://leaf-project.org/

Reply via email to