From: Mike O (mikeeo@email.msn.com)
Date: Wed Oct 05 2005 - 23:18:39 GMT-3
Once we shut the link down , OSPF re-converged and everything was fine. We
replaced the xen pack and the link came back up without a problem. I think
the timers throttle command caused the spf calculation too fast as the link
was going up and down and caused a bad flood of LSA data.
----- Original Message -----
From: "Andrew Lissitz (alissitz)" <alissitz@cisco.com>
To: "Mike O" <mikeeo@email.msn.com>
Cc: <ccielab@groupstudy.com>
Sent: Wednesday, October 05, 2005 10:04 PM
Subject: RE: OSPF routing loop
Hey Mike and group,
Certainly sounds like you have your hands full Mike. Is everything ok
now?
During the routing loop, two (or more) routers thought they had the
correct path to a destination. These two routers were not in sync, and
they were told either dynamically or statically to route the way they
did. Were these directly connected or did you see the packets bounce a
couple of hops back and forth?
If you think the flapping and SPF calculations caused this, then you
have a big issue to work on. Your network should not go crazy with a
flapping link...
Without getting into a million options, design considerations, and
choices ... You may consider your design and how a flapping link in one
part of your network caused routing loops. If @ all possible, recreate
your network in a lab and try and simulate what occurred today. You can
buy software for your lab to simulate large scale OSPF and BGP routes.
Have you considered a route optimization software? Perhaps an eval copy
may even help. I think route science is a popular one... GS Group - is
there a good route software to recommend here?
Kindest regards Mike and good luck,
Andrew
-----Original Message-----
From: Mike O [mailto:mikeeo@email.msn.com]
Sent: Wednesday, October 05, 2005 7:39 PM
To: Andrew Lissitz (alissitz)
Cc: ccielab@groupstudy.com
Subject: Re: OSPF routing loop
Well I think the flapping link was causing OSPF to re-calaculate and I
checked the log again and it looks like it was flapping every 5 seconds.
I also should mentioned this network has about 180,000 nodes.
----- Original Message -----
From: "Andrew Lissitz (alissitz)" <alissitz@cisco.com>
To: "Mike O" <mikeeo@msn.com>; <mandingo2073@yahoo.com>
Cc: <ccielab@groupstudy.com>
Sent: Wednesday, October 05, 2005 4:30 PM
Subject: RE: OSPF routing loop
> Hello,
>
> Multicast would have problems if the Unicast table is not right. RPFs
> etc... Incoming and outgoing interface problems...
>
> The reason I asked about redist is because OSPF, by itself, is pretty
> good about keeping loops from occurring. If you saw packets expiring
> and looping, then I would check your routing sources. Loops occur
when
> routing data is not consistent / correct and up to date.
>
> The devices that bounced the packets around thought they had the
correct
> route for the destinations. Why and where did they learn their
routing
> information....
>
>
> -----Original Message-----
> From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf
Of
> Mike O
> Sent: Wednesday, October 05, 2005 3:38 PM
> To: mandingo2073@yahoo.com
> Cc: ccielab@groupstudy.com
> Subject: Re: OSPF routing loop
>
> Well we run mulitcast and it reported a unicast routing loop and when
> you did a traceroute it kept bouncing back till the TTL expired.
>
>
>>From: Carl Willias <mandingo2073@yahoo.com>
>>Reply-To: Carl Willias <mandingo2073@yahoo.com>
>>To: Mike O <mikeeo@msn.com>, alissitz@cisco.com,
ccielab@groupstudy.com
>>Subject: Re: OSPF routing loop
>>Date: Wed, 5 Oct 2005 12:29:21 -0700 (PDT)
>>
>>Why do you think it was a routing loop? I have seen issues with a
> flood of
>>LSA's causing high CPU utilization on routers causing it to drop
> packets,
>>routing and otherwise.
>>
>>----- Original Message ----
>>From: Mike O <mikeeo@msn.com>
>>To: alissitz@cisco.com; ccielab@groupstudy.com
>>Sent: Wednesday, October 05, 2005 1:59:30 PM
>>Subject: RE: OSPF routing loop
>>
>>
>>Yes we redistbute OSPF into BGP for iBGP.
>>
>>We also use the command "timers throttle spf 400 400 4000" and I think
> it
>>was trying to flood out LSA's while the link was flapping and caused a
> loop
>>in the Area
>>
>>
>> >From: "Andrew Lissitz (alissitz)" <alissitz@cisco.com>
>> >Reply-To: "Andrew Lissitz (alissitz)" <alissitz@cisco.com>
>> >To: "Mike O" <mikeeo@msn.com>, <ccielab@groupstudy.com>
>> >Subject: RE: OSPF routing loop
>> >Date: Wed, 5 Oct 2005 14:34:22 -0400
>> >
>> >Mike, are you all performing any redistribution?
>> >
>> >
>> >-----Original Message-----
>> >From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf
> Of
>> >Mike O
>> >Sent: Wednesday, October 05, 2005 12:09 PM
>> >To: ccielab@groupstudy.com
>> >Subject: OSPF routing loop
>> >
>> >Can a rapidly flapping interface cause a routing loop with OSPF?
>> >yesterday we had an interface bounce every 90 seconds and caused a
>> >routing loop within the network until we shut it down. Shouldn't the
>> >hold-down timers prevent this?
>> >
>> >
>> >Thanks,
>> >
>> >Mike
>> >
>>
>>______________________________________________________________________
_
>> >Subscription information may be found at:
>> >http://www.groupstudy.com/list/CCIELab.html
>> >
>>
>>______________________________________________________________________
_
>> >Subscription information may be found at:
>> >http://www.groupstudy.com/list/CCIELab.html
>>
>>______________________________________________________________________
_
>>Subscription information may be found at:
>>http://www.groupstudy.com/list/CCIELab.html
>>
>>______________________________________________________________________
_
>>Subscription information may be found at:
>>http://www.groupstudy.com/list/CCIELab.html
>
>
This archive was generated by hypermail 2.1.4 : Sun Nov 06 2005 - 22:00:49 GMT-3