Thread (8 messages) 8 messages, 3 authors, 2011-08-02

Re: Bond, round robin mis-config​uration.

From: Eduard Sinelnikov <hidden>
Date: 2011-08-01 05:50:14

I have put the bond down: 'ifconfig bond0 down.'
Then changed the mode: 'echo 1 > /sys/class/net/bond0/bonding/mode'

Without puting the bond down, you couldn't change the mode.

2011/7/31 Peter Senna Tschudin [off-list ref]:
On Sun, Jul 31, 2011 at 12:15 PM, Eduard Sinelnikov
[off-list ref] wrote:
quoted
Hi,

The scenario is:
* Create a bond with 3 interfaces (connect them to switch).
* Change bond's mode to active/backup.
* Physicly remove two cables form interfaces ( not the active interface ).
* Put the cables back
* Change the mode to round robin.
How did you change the mode?
quoted
* Try to ping some other computer.

Now only one interface is pinging to remote computer.
Without removing the cables all three interface will ping to remote
computer periodicly.


I did some debuging,in the code, and I see that in round robin all the
interface is in active (and all of them transmiting periodically).
After removing and puting back the cables(in active/backup mode). the
interfaces change their status to backup.
After this only one interface is transmiting ( the one which was the active).

Thanks in advance,
? ? ? ? ? Eduard

2011/7/31 Peter Senna Tschudin [off-list ref]:
quoted
Hi Eduard,

On Sun, Jul 31, 2011 at 4:59 AM, Eduard Sinelnikov
[off-list ref] wrote:
quoted
Hi,

In the kernel 2.6.39.3 ( /drivers/net/bond/bond_main.c).
I followed the code you mentioned. The file is actually at:

./drivers/net/bonding/bond_main.c
quoted
In the function ??bond_xmit_roundrobin?
The code check if the bond is active via
?bond_is_active_slave(slave)? Function call.
Which actually checks if the slave is backup or active
What is the meaning of slave being ?backup in round robin mode?
Correct me if I wrong but in round robin every slave should send a
packet, regardless of being active or backup.
I'm not sure about this but my best guess is that even using all
slaves to send packages, the slaves must be used one at a time, to
send packages sequentially. And one slave can be deactivated when a
problem is detected. I think that this two scenarios that justify the
check.
quoted
Thank you,
? ? ? ? ? ?Eduard

_______________________________________________
Kernelnewbies mailing list
Kernelnewbies at kernelnewbies.org
http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
Peter

--
Peter Senna Tschudin
peter.senna at gmail.com
gpg id: 48274C36


--
Peter Senna Tschudin
peter.senna at gmail.com
gpg id: 48274C36
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help