RE: ASR9k - ospf stuck in exchange

I had a few weird things with ospf between my asr9k and asr901....i think
this was seen with 9k on IOS XR 4.1.2 and ASR901 on 15.1(2)SNG (btw, don't
bother pulling your hair out trying to get NTP to synch with this verion of
ASR901 code, go ahead and upgrade to 15.1(2)SNH and forego the ntp
weirdness)

One of the weird things was IF I used a /30 between 9k and 901 I had to use
network type p-to-p (I think this was on xr 4.1.2.....and strangely I think
I saw that I did *not* have to do this with xr 4.0.1)

Another thing I had to deal with was mtu mismatch on neighboring...i think I
was stuck in exchange here til I resolved it. Notice that at face value it
seems that the mtu's are not same, but read my notes...

My notes during my deployment....

---------------------------------------------------------------------
--------begin change mtu on gige link between 9k and 901-------------

...this was done since during testing i was unable to send tagged traffic
over the mpls pw from end to end, strange one-way traffic flow was seen.
seemed that it was due to a 1500 default mtu setting on an asr gige link to
a 901. then once i changed it on the 9k to 9216, the ospf neighbor session
drops and stuck in exstart....this was seen on debugs with mtu
mismatch...whichever side sees larger says it sees larger on opposite
side....also, 14 bytes less was needed on setting of 901 *OR* ospf
mtu-ignore needed to be set to allow for mismatch. i opted for setting the
same mtu on both sides.
---------------------------------------------------------------------

--- 901

conf t
interface Vlan1
mtu 9202
exit
exit
wr

--- 9k

conf
interface gig0/1/0/1
mtu 9216
commit

------------ notes....

* there appears to be an MTU mismatch between the ip interfaces that the 901
and 9k neighbor on..901 vlan 1 9202 and 9k g0/1/0/1 9216..however, see 9k
output..it appears that ospf is advertising an mtu of 9202 to 901. and this
seems to be calculated on the set/config'd mtu of g0/1/0/1 minus 14 (eth
hdr)

RP/0/RSP0/CPU0:9k#sh int GigabitEthernet0/1/0/1 br
Tue Apr 3 08:31:21.758 CST

               Intf Intf LineP Encap MTU
BW
               Name State State Type (byte)
(Kbps)
----------------------------------------------------------------------------

----
          Gi0/1/0/1          up          up               ARPA  9216
1000000
RP/0/RSP0/CPU0:crml-9k#sh ip os in GigabitEthernet0/1/0/1 | in MTU
Tue Apr  3 08:31:29.597 CST
  Transmit Delay is 1 sec, State POINT_TO_POINT, MTU 9202, MaxPktSz 9000
---------------------------------------------------------------------
--------end change mtu on gige link between 9k and 901---------------
---------------------------------------------------------------------
-----Original Message-----
From: nobody_at_groupstudy.com [mailto:nobody_at_groupstudy.com] On Behalf Of
AKHILESH THAKUR
Sent: Monday, April 23, 2012 11:56 PM
To: chryan_at_cisco.com; ccielab_at_groupstudy.com
Subject: RE: ASR9k - ospf stuck in exchange
Dear Experts,
The IP MTU is set to 9202 at ASR end and 9202 at 7600 end. The problem
occurs when the physical link only supports 1500 not 9202. WHile the same
problem is not observed connecting 7600s, configure mtu-ignore enable on
interface and ospf will come up no matter what mtu is at physical link.
Since this is production network, still will try to get some debug ip ospf
adj and check.
Regards
Akhilesh
> Date: Mon, 23 Apr 2012 23:14:16 -0500
> From: chryan_at_cisco.com
> To: ccielab_at_groupstudy.com
> Subject: Re: ASR9k - ospf stuck in exchange
>
> Not sure if you are able to do this, because it's a production 
> network, but "debug ip ospf adj" should give you some clues as to what 
> the problem is.
>
> Chuck
>
> On 4/23/12 12:43 PM, AKHILESH THAKUR wrote:
> > Dear Experts,
> > I have migrated my network from 7600 to ASR9k. Before migration if 
> > the
link
> > between two router supports max mtu of 1536 and interface is 
> > configured with mtu9216, the ospf will come up.
> >
> > Now at one end is ASR9k with same configuration it doesnot work. I 
> > know
the
> > ASR9k put additional mtu of 14 bytes on ethernet ports and has been
adjusted.
> >
> > so at ASR9k end mtu is 9216 and 7600 end mtu is 9202. The ospf just 
> > stuck
in
> > exchange.
> > I have opened the tac case and since the link is max support 1536,
accordingly
> > needs to change the mtu at asr9k. I tired with mtu-ignore but failed.
> >
> > Wanted to know whats the difference between 7600 and asr9k ospf DBD
packet
> > handling ?
> >
> > Any of you guys faced similar issue and the way to deal with it.
> >
> > Regards
> > Akhilesh
> >
> >
> > Blogs and organic groups at http://www.ccie.net
> >
> > ____________________________________________________________________
> > ___ Subscription information may be found at:
> > http://www.groupstudy.com/list/CCIELab.html
> >
> >
> >
> >
> >
> >
> >
> >
>
>
> --
> ------------------------------------
> Chuck Ryan
> Customer Support Engineer
> High Touch Technical Support (HTTS)
> Cisco Systems
> (469) 255-5672
> Email: chryan_at_cisco.com
> Work Hours: 6pm - 2am CST, Sun-Thurs
> ------------------------------------
> **** If you need assistance outside of these hours, please call the 
> HTTS
Toll Free number (800-495-9121) and ask to speak to the next available
engineer. ****
>
> This email may contain confidential and privileged material for the 
> sole use
of the intended recipient. Any review, use, distribution or disclosure by
others is strictly prohibited. If you are not the intended recipient (or
authorized to receive for the recipient), please contact the sender by reply
email and delete all copies of this message.
>
> For corporate legal information go to:
> http://www.cisco.com/web/about/doing_business/legal/cri/index.html
>
>
> Blogs and organic groups at http://www.ccie.net
>
> ______________________________________________________________________
> _ Subscription information may be found at:
> http://www.groupstudy.com/list/CCIELab.html
Blogs and organic groups at http://www.ccie.net