[Xorp-users] OLSR test results on Linux.

[Ben Greear]

I've been doing a few preliminary tests of OLSR + pim-sm testing on Linux.

With a patch to fea (posted previously on the xorp-hackers list), it works fine,
including with virtual tables and multiple instances of Xorp.

I had to add a policy to export connected and static routes into olsr to get
it to act as I wanted, but that also worked fine as soon as I configured
xorp properly.

Normal UDP (unicast) traffic works fine.  I can't get mcast working, but
that may be some other problem on my end (it works fine with OSPF in the
same configuration)

Next, I tried dynamic changes (using xorpsh to add/remove interfaces).  I had 3 virtual
nodes, linked in a triangle.  I was running traffic between N2 and N3.  I dropped the link
between N2 and N3 and traffic started flowing through N1 as expected.

I re-added the link, and traffic quickly switched back to flowing directly between N2 and N3.

So far, the only weird thing is that it seems to take even longer than ever to execute
commands to add the interfaces.  I'll need to do more testing to figure out if OLSR
is to blame for this, or if it's something else:

XORPsh-helper:  Completed command: 8 successfully in 5296.577ms against router-id: 2,  cmds: nice -17 xorpsh  -c configure -c set interfaces interface 2.3.2 vif 
2.3.2 address 10.2.3.2 prefix-length 24 -c commit -c set protocols olsr4 interface 2.3.2 vif 2.3.2 address 10.2.3.2 interface-cost 1 -c set protocols igmp 
interface 2.3.2 vif 2.3.2 disable false -c set protocols pimsm4 interface 2.3.2 vif 2.3.2 disable false -c set plumbing mfea4 interface 2.3.2 vif 2.3.2 disable 
false -c commit -c exit -c exit

(This is on an 8-core nehalem box)

Thanks,
Ben

-- 
Ben Greear <greearb at candelatech.com>
Candela Technologies Inc  http://www.candelatech.com