On Tue, Jan 31, 2017 at 10:51:36AM -0800, Ben Pfaff wrote:
> On Wed, Jan 18, 2017 at 04:56:00PM -0600, Chad Norgan wrote:
> > Thanks so much for the deeper understanding. I think from here I'm
> > going to get a mirror port setup on the switch so I can confirm that
> > the switch is indeed sending
On Wed, Jan 18, 2017 at 04:56:00PM -0600, Chad Norgan wrote:
> Thanks so much for the deeper understanding. I think from here I'm
> going to get a mirror port setup on the switch so I can confirm that
> the switch is indeed sending that pdu on the restored link. I can then
> take that to the switch
Chad,
One more thing to check with you, do you have the following
configuration enabled for the bond by any chance?
other-config:lacp-fallback-ab=true
If you do, could you please try to set it as false and test the scenario
again? The parameter is not printed out by "ovs-appctl bond/show" and
I will look at what could be done without breaking the protocol and will
report back in a couple of days.
On Jan 18, 2017 2:56 PM, "Chad Norgan" wrote:
> Shu,
>
> Thanks so much for the deeper understanding. I think from here I'm
> going to get a mirror port setup on the switch so I can confirm
Shu,
Thanks so much for the deeper understanding. I think from here I'm
going to get a mirror port setup on the switch so I can confirm that
the switch is indeed sending that pdu on the restored link. I can then
take that to the switch vendor to address.
I love the idea of patching OVS to handle
Shu, thanks for all the debugging!
If this is a correct interpretation of the standard, and the peer switch
is misbehaving, and it's common behavior across a product line, then
possibly we'll need to make OVS cope with it.
Chad, do you have thoughts on Shu's discoveries?
Thanks,
Ben.
On Wed, J
Hi Chad,
I now have a theory on what's happening in your case.
I realized that the first LACPDU packet the peer switch sent for
re-negotiation contains all zeros in Actor Informaiton TLV, line
81-84 from your gist:
20:38:17.109650 00:00:00:00:00:00 > 01:80:c2:00:00:02, ethertype Slow
Protocol
On Tue, Jan 17, 2017 at 04:54:51PM -0600, Chad Norgan wrote:
> Given that the partner port_id on the rogue packet matches the slave
> it's sent out. I lean towards #1, that the LACP implementation is
> somehow mixing up the status for the slave's pdu, rather than leaking
> eth1's pdu out the eth0 i
Short answer is no doesn't ring a bell in particular. Definitely doesn't
sound like the expected behavior.
FWIW, based on my very rusty memory of the code, it sounds like one of two
things is happening:
1) The LACP protocol implementation itself has a bug in which it's sending
the incorrect stat
Given that the partner port_id on the rogue packet matches the slave
it's sent out. I lean towards #1, that the LACP implementation is
somehow mixing up the status for the slave's pdu, rather than leaking
eth1's pdu out the eth0 interface.
-Chad
___
disc
On Thu, Jan 12, 2017 at 06:36:52PM -0600, Chad Norgan wrote:
> I've been doing some bug chasing around some unintended impacts we've
> been noticing on our bonded hypervisors. The servers have a bond with
> two slave interfaces each going to a different upstream switch which
> have been configured
I've been doing some bug chasing around some unintended impacts we've
been noticing on our bonded hypervisors. The servers have a bond with
two slave interfaces each going to a different upstream switch which
have been configured with a Virtual PortChannel (VPC). To OVS, the VPC
configuration makes
12 matches
Mail list logo
