Hi,
This is one that comes up in almost every network that I have been trouble shooting
lately.
In the Cisco environment set you spanning-tree portfast option on ports that have a
true one to one relationship that is, servers, clients, printers and the like.
Do NOT use portfast option on hubs, switches and routers. There is nothing worse than
spending time tracking down a segmentation type fault only to find some clown has
decided to put both ends of a cable into sockets to tidy up or cover some occupational
safety thingy.
Some other brands of switches allow you to turn spanning tree off on a port by port
basis. The same rules as above apply.
Another one that becomes critical particularly with older Catalyst 5000 type systems
is make sure you are using the same vervion of the spanning tree algorithm. A buggy
or incompatable spanning tree can start to sing. This is great for bringing down a
network and can take some time to find.
Also ensure port speeds and duplex setting are correct. I avoid Auto settings
whenever possible. They often fail and different companies do not always agree on how
they discover the remote end. Some companies allowed preferred methods and they don't
always agree with other companies preferred.
Take this for what it is. I make a nice living walking around talking about this
stuff.
Teunis,
Hobart, Tasmania
Australia
On Saturday, December 16, 2000 at 03:10:25 PM, Craig Columbus wrote:
> Yep. Seen this one. I had a fairly large switched environment (Bay, not
> Cisco) where people were failing to obtain DHCP addresses. Spanning tree
> was removed and the problem went away.
>
> Craig
>
> At 02:45 PM 12/16/2000 -0500, you wrote:
> >The potential problem is spanning tree will shut down the link for the first
> >thirty seconds (default) before passing traffic. If your computer is far
> >enough along in the boot proccess and makes a dhcp request before the switch
> >activates the port, the dhcp request will fail. Use portfast to prevent this
> >problem. Portfast is fine so long as you are sure no switch will be
> >connected to that port. This problem can be seen more frequently with PCMCIA
> >NICS. The computer may not assert the link on the card until late in the
> >boot process and then almost immediatly issue the DHCP request.
> >""A. Ward"" <[EMAIL PROTECTED]> wrote in message
> >91dluo$tgr$[EMAIL PROTECTED]">news:91dluo$tgr$[EMAIL PROTECTED]...
> > > Has anyone had issues where DHCP limited the use of the Spanning tree =
> > > protocol?
> > >
> > > Can you send me a lead to a white paper or give me any information on =
> > > this issue?
> > >
> > > Thanks.
> > >
> > > A. Ward
> > >
> > > _________________________________
> > > FAQ, list archives, and subscription info:
> >http://www.groupstudy.com/list/cisco.html
> > > Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
> > >
> >
> >
> >_________________________________
> >FAQ, list archives, and subscription info:
> >http://www.groupstudy.com/list/cisco.html
> >Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
>
> _________________________________
> FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html
> Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
>
>
--
www.tasmail.com
_________________________________
FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html
Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
