[AG-TECH] Testing multicast connectivity...

[Kelly, Vaughn [NCSUS Non J&J]]

I've seen similar problems at multiple sites for various reasons, one that
hasn't been mentioned though, (aside from you checking YOUR lists, which
you've already done) is to verify that no connected to you (i.e. your
provider) isn't filtering any of the routes prior to delivery.  The problem
may also lie not in the route advertisements themselves, but MSDP source
filtering from remote networks.  Be wary when any access list between a
necessary source and your receivers gets modified!

-----Original Message-----
From: Richard Coffey [mailto:richardc at cs.utah.edu] 
Sent: Tuesday, May 20, 2003 11:00 PM
To: Bill Nickless
Cc: ag-tech at mcs.anl.gov; Mitch Kutzko
Subject: Re: [AG-TECH] Testing multicast connectivity...

I was referring to both problems. The Cisco bug might be a problem as well,
but I think most of our core routers have been upgrade beyond this problem.
I'll check with our NOCC on the IOS revs.

Thanks Bill and Mitch!

-Richard


On 5/20/03 8:24 PM, "Bill Nickless" <nickless at mcs.anl.gov> wrote:

> Ah, this is a different problem.  I thought you were talking about the
> beacon traffic not being heard by other beacons.  Now I understand you're
> talking about your beacon crashing from time to time (or one of the three
> threads).
> 
> Mitch Kutzko <mitch at ncsa.uiuc.edu> has a slightly modified version of the
> beacon code, with larger pre-compiled constant buffers.  I'm not sure if
> Mitch thinks it's ready for prime time yet.  I've copied him on this
> note--Mitch, is it ready for distribution?
> 
> At 05:12 PM 5/20/2003 -0600, Richard Coffey wrote:
>> Hi Bill,
>> 
>> Avtwo.cs.utah.edu is the beacon is question. Every once and a while, Java
>> throws and exception and kills the connection. Thoughts on that?
>> 
>> -Richard
>> 
>> On 5/20/03 4:15 PM, "Bill Nickless" <nickless at mcs.anl.gov> wrote:
>> 
>>> This is consistent with an old bug in Cisco IOS prior to 12.0(10).  If
the
>>> router directly connected to the source is running an earlier version,
then
>>> the behavior you mention below is possible.
>>> 
>>> What specific beacon sources are you referring to?
>>> 
>>> At 03:57 PM 5/20/2003 -0600, Richard Coffey wrote:
>>>> Hi all,
>>>> 
>>>> As of late, I've noticed problems with individual multicast addresses
>>>> working. The beaconserver's address seems to indicate that we cannot
send,
>>>> yet I've gone into the test room and successfully collaborated. Also,
>> people
>>>> cannot see me in a particular room, but in other rooms they can, and
vice
>>>> versa.
>>>> 
>>>> We're not running tight ACL's on our routers, so it's not that.
Thoughts?
>>>> 
>>>> -Richard
>>>> 
>>>> --
>>>> Richard C., SCI Evangelist
>>>> SCI Institute
>>>> University of Utah
>>>> http://www.sci.utah.edu/
>>>> 801.918.5050
>>>> 801.585.6513 fax
>>> 
>>> ===
>>> Bill Nickless    http://www.mcs.anl.gov/people/nickless      +1 630 252
>> 7390
>>> PGP:0E 0F 16 80 C5 B1 69 52 E1 44 1A A5 0E 1B 74
>> F7     nickless at mcs.anl.gov
>>> 
>>> 
> 
> ===
> Bill Nickless    http://www.mcs.anl.gov/people/nickless      +1 630 252
7390
> PGP:0E 0F 16 80 C5 B1 69 52 E1 44 1A A5 0E 1B 74 F7
nickless at mcs.anl.gov
> 
> 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/ag-tech/attachments/20030521/45646614/attachment.htm>