Today, our network ground to a total halt during the afternoon.  First the
Internet access was lost, then stations slowed right down, froze and would
not reboot.

I checked the servers which showed over 90% system idle and then saw that
the main network switch (HP procurve 5308 with 6 modules - 4 modules with
1000T connections and 2 modules with a total of 6 active fibre modules) had
all its LEDs on continuously, as were secondary switches.
Rebooting the switches did not help so I removed all but the server
connections to the main switch, normal flashing, so I added the remaining 13
copper patch leads one at a time, all LEDs flashing merrily.  Then, as I
reconnected the fibre leads, the last one caused the LEDs to flash more and
more and then lock on, as soon as I removed it normal activity resumed.

I went to the science lab to which this fibre run led and found that a Cat5
patch cable had been connected from one outlet port to another.  Once
disconnected, I found I could reconnect the fibre lead at the main switch
and all was fine.

Is this what is called a network storm, am I right in thinking that enabling
Spanning Tree protocol would prevent this problem from happening again if
our little dears find out about this rather simple way of causing chaos.
(Actually I think this was done by a helpful science teacher who saw the end
of a lead used to connect laptops to the network dangling across the bench
and assumed it had 'fallen out')

David Thomas