[Date Prev][Date Next] [Thread Prev][Thread Next]
[Thread Index]
[Date Index]
[Author Index]
RE: [K12OSN] PXE boot problems...........resolved
- From: R P Herrold <herrold owlriver com>
- To: james <james mail nbsd org>
- Cc: "'Bert Rolston'" <BertR wanganui-city school nz>,"'K12OSN'" <k12osn redhat com>
- Subject: RE: [K12OSN] PXE boot problems...........resolved
- Date: Tue Mar 5 22:29:18 2002
On Tue, 5 Mar 2002, james wrote:
> OT: I don't see why an extra hub (You mean, a regular 10 or 100 or
> 10/100 Megabit hub? Not a switch or router?) would interfere.
>
> Hubs are dumb, they just boost signals and repeat everything told to
> them. They shouldn't interfere with any network operations.
Of course they interfere -- they add latency, and cannot
detect, and can hide collision traffic -- the so-called 5-4-3
rule of thumb -- end to end, 5 segments max, with 4 hubs max,
and no more than 3 segments occupied -- was probably in play
-- more of any of these limits can mask errors which logic
says cannot be happening, down at the physical layer.
Adding a bridge, or a switch for layer 2 regeneration, or a
router for layer 3 can help, but MAC layer traffic cannot be
routed (such as the PXE broadcasts) at layer 3.
Me ? -- I strive to 4-3-3, to be on the safe side. I had a
hubbed speed mismatch negotiaion biting me just last week with
a Cisco <==> Linux 2.4 kernel/(3com or Intel) dual speed NIC
transition -- a utility dual speed switch was dropped in, and
the issue went away.
-- Russ Herrold
[Date Prev][Date Next] [Thread Prev][Thread Next]
[Thread Index]
[Date Index]
[Author Index]