Re: Link aggregation between NetBSD agr and Linux bond interfaces

[BERTRAND Joël <joel.bertrand%systella.fr@localhost>]

Gert Doering a écrit :
> Hi,
>
> On Wed, Dec 07, 2016 at 11:27:29PM +0100, BERTRAND Joël wrote:
> >       Even with interfaces up on both sides, tunnel is unusable.
>
> What does "unusable" mean, exactly?  Is the agr interface down, are
> packets not being sent to tap0/tap1, are they not passed to openvpn,
> are they not being received, etc.?

	No connection. Both OpenVPN links run as expected. But even if agr0 is 
up (and configured of course with an IP address), no data is received 
from aggregated interface. I suppose that NetBSD try to send data, but 
Linux host doesn't receive any packets.

> You'll need to do some tcpdumping on all interfaces involved to see
> how far the packets get...

	Of course, I have tried to check link with tcpdump. WHen I try to ping 
Linux host from NetBSD, I obtain on agr0 :
10:34:46.034686 ARP, Request who-has 192.168.100.1 tell 192.168.100.2, 
length 28
10:34:47.034689 ARP, Request who-has 192.168.100.1 tell 192.168.100.2, 
length 28
10:34:48.034682 ARP, Request who-has 192.168.100.1 tell 192.168.100.2, 
length 28
10:34:49.034679 ARP, Request who-has 192.168.100.1 tell 192.168.100.2, 
length 28
10:34:50.034666 ARP, Request who-has 192.168.100.1 tell 192.168.100.2, 
length 28
10:34:51.034655 ARP, Request who-has 192.168.100.1 tell 192.168.100.2, 
length 28
10:34:52.034655 ARP, Request who-has 192.168.100.1 tell 192.168.100.2, 
length 28
10:34:53.034647 ARP, Request who-has 192.168.100.1 tell 192.168.100.2, 
length 28
10:34:54.034642 ARP, Request who-has 192.168.100.1 tell 192.168.100.2, 
length 28
10:34:55.034639 ARP, Request who-has 192.168.100.1 tell 192.168.100.2, 
length 28
10:34:56.034624 ARP, Request who-has 192.168.100.1 tell 192.168.100.2, 
length 28
10:34:57.034635 ARP, Request who-has 192.168.100.1 tell 192.168.100.2, 
length 28
10:34:58.034615 ARP, Request who-has 192.168.100.1 tell 192.168.100.2, 
length 28
10:34:59.034613 ARP, Request who-has 192.168.100.1 tell 192.168.100.2, 
length 28
10:35:00.034612 ARP, Request who-has 192.168.100.1 tell 192.168.100.2, 
length 28

	If I try to ping NetBSD from Linux, I obtain the same message.

> What I hear from people in the Linux world, this *should* work (aka
> "people have done it successfully on Linux"), though I've never done
> it myself.  But it should work...
>
> Looking at the man page, agr seems to default to use LACP, which might
> or might not be the problem - so I'd start by turning it off ("link1")
> to see if a static tunnel works.  If that works, check that the linux
> side is also speaking LACP and whether that part comes up.

	I have turned of link1. And on linux side, I suppose bond0 is 
configured to use LACP :
rayleigh:[~] > cat /proc/net/bonding/bond0
Ethernet Channel Bonding Driver: v3.7.1 (April 27, 2011)

Bonding Mode: IEEE 802.3ad Dynamic link aggregation
Transmit Hash Policy: layer2 (0)
MII Status: up
MII Polling Interval (ms): 100
Up Delay (ms): 200
Down Delay (ms): 200

802.3ad info
LACP rate: slow
Min links: 0
Aggregator selection policy (ad_select): stable
...


> Maybe OpenVPN is getting confused by the LACP frames and not forwarding
> them (it *should* be totally transparent in TAP mode, but I'm not sure
> anyone has tested LACP-over-TAP yet)

	Best regards,

	JKB