From: Mas Kato (loomis_towcar@xxxxxxxxxxxxxx)
Date: Wed Oct 10 2001 - 21:03:40 GMT-3
[demime could not interpret encoding binary - treating as plain text]
Jeremy,
Other culprits could include interface resets (any way to track them?) and perh
aps network diameter-related timer settings either being mismatched or too low,
or actual diameter too big for legal topology, etc.
Do any of the C'trons keep track of where the last 'Topology Change Notificatio
n' BPDU came from? Or perhaps your sniffer can filter on them? Might lead to a
smoking gun...
Regards,
Mas Kato
https://ecardfile.com/id/mkato
> "Wright, Jeremy" <JA_WRIGHT@admworld.com>Cc: ccielab@groupstudy.com
> RE: BPDU/Topology ChangesDate: Wed, 10 Oct 2001 12:24:25 -0500
>Reply-To: "Wright, Jeremy" <JA_WRIGHT@admworld.com>
>
>Thanks for the info....no I don't see any topology changes during off hours.
>We see them sparatically through the day but more towards the beginning of
>the day and the end which makes me think that it has to do with people
>booting up and shutting down there computer and something is generating
>topology changes.
>
> -----Original Message-----
> From: Craig King [mailto:cking007@hotmail.com]
> Sent: Wednesday, October 10, 2001 12:15 PM
> To: Wright, Jeremy
> Cc: ccielab@groupstudy.com
> Subject: Re: BPDU/Topology Changes
>
> Have you taken a look at the utilization/traffic patterns of
>your network?
> You don't mention anything about your topology except the
>1000-user base.
> If your network is suffering from congestion problems, the
>switches could be
> adjusting the topology because BPDUs are delayed and the
>switches think that
> certain links are down. Do you see any topology changes in
>offhours? If
> not, then it could be related to network traffic patterns
>throughout the
> day.
>
> Take a look at other stats that might indicate a
>oversaturated or problem
> link (collisions, errors, delay exceeded, etc...). This
>might help isolate
> the problem to a specific portion of your network.
>
> HTH,
>
> CK
>
> ----- Original Message -----
> From: "Wright, Jeremy" <JA_WRIGHT@admworld.com>
> To: <jezerski@broadcom.com>; "'Chris Allen'"
><chris.allen@callisma.com>;
> "Wright, Jeremy" <JA_WRIGHT@admworld.com>
> Cc: <ccielab@groupstudy.com>
> Sent: Wednesday, October 10, 2001 11:03 AM
> Subject: RE: BPDU/Topology Changes
>
>
> > Thanks for the advice....I have defined the root and
>secondary bridge
> > before, so I don't think that is an issue. My uplink ports
>are also not in
> > port fast. We haven't had to power cycle or add any new
>switches
> > recently.I'm just having trouble tracking down the
>device(s) that are
> > causing the topology change. If we had Cisco devices I
>could find it...but
> > since my boss loves his Cabletron and the prices...we are
>now paying for
> it.
> > I have been suggesting Cisco since I got here and what's
>funny is, my boss
> > asked "do you know of any good Cisco switches".....I would
>hate for him to
> > actually take my advice for once. Thanks a lot for the
>input...anymore
> would
> > be appreciated...
> >
> > -----Original Message-----
> > From: Joseph Ezerski [mailto:jezerski@broadcom.com]
> > Sent: Wednesday, October 10, 2001 10:49 AM
> > To: 'Chris Allen'; 'Wright, Jeremy'
> > Cc: ccielab@groupstudy.com
> > Subject: RE: BPDU/Topology Changes
> >
> > I might also add that you do not want a portfast type of
> > technology on your
> > trunk links between switches. At least, in the Cisco
>world,
> > turning on
> > Portfast bypasses MAXAGE and the Forward Delay timers and
> > goes right into
> > Forwarding, thus effectively ignoring Spanning Tree. I
> > would venture to say
> > that you might still have STP loops but are not noticing
> > them, since
> > enjoying the blissful peace of non-convergence.
> >
> > In a corrrectly functioning network, you should have NO
> > topolgy changes
> > unless you added more switches, changed the root bridge,
>or
> > changed the path
> > costs.
> >
> > Do what Chris suggests. Set your root and secondary
> > bridges. Don't leave
> > those to chance. It could be as simple as that. It is a
> > good start. Then
> > work your way down from there.
> >
> > I had an example where a Cat 3548XL, in a lab no less,
>that
> > had a better MAC
> > and BID than my core 6509s and was assuming the duties of
> > the root bridge
> > (this was before I set my root priorities). Everytime
>that
> > switch was
> > rebooted or unplugged we had a convergence. Needless to
> > say, it sucked,
> > until I set my root bridge priority to 0 and my secondary
>to
> > 1. Now, the
> > only way someone can assume root is if they have a 0
> > priority and a better
> > MAC, which is highly unlikely. Add to that a newer
>feature
> > from Cisco
> > called RootGuard that will eliminate even that slight
> > chance.
> >
> > I hope this helps.
> >
> > -Joe
> >
> > -----Original Message-----
> > From: nobody@groupstudy.com
>[mailto:nobody@groupstudy.com]On
> > Behalf Of
> > Chris Allen
> > Sent: Wednesday, October 10, 2001 8:27 AM
> > To: Wright, Jeremy
> > Cc: ccielab@groupstudy.com
> > Subject: RE: BPDU/Topology Changes
> >
> >
> > Are you forcing Root and Secondary Bridge Elections??? If
> > not I would....
> > Also you can make the network even more predictable and
> > stable by manually
> > configuring port costs to force a certain path....
> >
> > -----Original Message-----
> > From: nobody@groupstudy.com
>[mailto:nobody@groupstudy.com]On
> > Behalf Of
> > Wright, Jeremy
> > Sent: Wednesday, October 10, 2001 10:18 AM
> > To: 'cisco@groupstudy.com'
> > Cc: 'ccielab@groupstudy.com'
> > Subject: BPDU/Topology Changes
> >
> >
> > My network recently was having a lot of topology changes.
>I
> > then used
> > fastforwarding (Cabletron's version of port fast, and no I
> > don't make the
> > decision on what hardware to purchase, I just configure)
>and
> > it reduced the
> > number of changes tremendously. We are still having
>several
> > per day, we have
> > about a 1000 users here. I'm running Sniffer with a filter
> > on BDPU's and was
> > wondering if there were any suggestions on a better way to
> > track the
> > topology changes down. Also, I would like to know what is
> > the recommended
> > min/max number of topology changes. I realize it is
>specific
> > to your network
> > but I just want to get a general feel. Thanks team
> >
> >
This archive was generated by hypermail 2.1.4 : Thu Jun 20 2002 - 22:33:16 GMT-3