[ofa-general] Re: IB Bonding errors with recent kernel

[Pradeep Satyanarayana]

Or Gerlitz wrote:
> Pradeep Satyanarayana wrote:
>> I was attempting to execute a few bonding tests with the 2.6.28-rc7
>> kernel and see the following error when I try to add the first slave:
>> ib0: multicast join failed for
>> 0001:0000:0000:0000:0000:0000:0000:0000, status -22
> Hi PRadeep,
> 
> This (junk MGID) issue is well know for somehow long time and irrelevant
> to bonding, we see it with non bonded IPoIB devices as well. I added
> Yossi Etigin from Voltaire to the distribution list as he might have
> more details on the issue.
> 
Or,

If I am not mistaken the issue you mention is a little different from the one I pointed out.

Without bonding I see the following:

kernel: ib0: multicast join failed for ff12:401b:ffff:0000:0000:0000:ffff:ffff, status -11

However, with bonding what I see is :

ib0: multicast join failed for 0001:0000:0000:0000:0000:0000:0000:0000, status -22

The latter is seen only when IPoIB bonding is activated (i.e. when a slave is added)

The latter junk MGID is because a Ethernet mapped address is being added to the dev->mc_list.

Subsequently an ib-bond status does not reveal any slave as active as shown below:

ib-bond --status
bond0: 80:00:04:04:fe:80:00:00:00:00:00:00:00:05:ad:00:00:03:05:b9
slave0: ib0
slave1: ib1

Pradeep