[BBLISA] How do you detect bad hardware on your network?

[Alex Aminoff]

On 12/7/2015 5:49 PM, Chuck Anderson wrote:
> On Mon, Dec 07, 2015 at 03:56:01PM -0500, Alex Aminoff wrote:
>> We have an office network with about 200 desks, with Dell managed 24
>> and 48 port switches, with spanning tree and broadcast storm
>> control. At the desks and carrels, we have many small 4 and 8 port
>> non-managed switches. Sometimes one of these small switches
>> malfunctions in a way that affects large portions of our network,
>> but intermittently. What are best practices for detecting and
>> pinpointing this sort of problem?
> On the managed switches use features that immediately shut down a port
> when a BPDU is received (BPDU Guard) or when too many MAC addresses
> are learned (say more than 16 for your typical edge port).  Make sure
> you are capturing syslogs and/or SNMP traps so you know when a port
> shuts down.
I wanted to follow up on this to say that it worked! We were able to set 
up snmptrapd on our log server, and when during some scheduled down-time 
we plugged in the known bad 8-port switch, we did see both syslog and 
SNMP trap messages saying "Spanning Tree Protocol status Blocking" and 
specifying the switch and the port. The good thing about this is this 
appears to be an indicator of a problem with no false positives. So 
thank you!

  - Alex Aminoff
    NBER