[f-nsp] Port State Blocking, FastIron SX 800

[Nick Morrison]

On Wed, Sep 30, 2009 at 10:50 PM, Chuck Ufarley <markentime at gmail.com> wrote:
> Crisis averted. A few more reboots and this thing started to behave.  Thanks
> for your time, everyone.

s/averted/postponed/ :-)

I've had all sorts of bizarre issues with XG optics in FI-SX's - the
2XG management modules, linking to various other bits of Foundry kit.
We've had link but no traffic; link but some traffic; no link lights
but UP reported in show int..

Also recently had a problem where the management card CPU load was
being driven up by ... wait for it ... a high number of spurious
interrupts from the PSUs!  Took 'em a long time to find that one.
They're quality switches, to be sure!  Brocade have our PSUs in for
analysis at the moment.

Thankfully we limit our use of FI-SX switches around the network.

I'm interested to know how many other people are closely monitoring
their Foundry/Brocade switches, and possibly seeing occasional various
oddities.  Do you monitor and graph your CPU load, memory usage, error
rates, and so on?  Do you watch and analyse your syslogs and snmp
traps?

(I should point out that the network I'm working on comprises mainly
RX-4 to -16 switches, with a few SXs doing some of the edge work.)

Nick

-- 
Nick Morrison <nick at nick.on.net>