02-05-2012 03:41 PM
We had a cable suddenly go bad from the core switch (DCX-4S) to a disk array. It didn't die so was causing the servers using that connection to hang. I was able to spot this and disabled the offending port and got the engineers to check out the cable.
My problem is the time it took to pick this up - in errshow we see multiple MS-1009 "Exceeded bit error rate threshold" errors. We have all snmp traps sent to a server and I was hoping to see a bunch of traps correlating to these messages but was disappointed. We have Fabric Watch enabled but I suspect that we have something misconfigured. Should I expect to see an snmp alert for each MS-1009 message?
02-06-2012 09:44 AM
for my understanding FabricWatch events and events in the errdump are two differernt things.
First of all you can can forawrd each event of the errorlog to a syslog server. But this is not a trap which you have asked for.
You can configure FabriWatch to monitor errorcounter on ports. These can be configured with custom values and can be send as traps, syslog or e-mails in case of crossing a threshold. If a threshold has been crossed you will get a trap or syslog ort e-mail if configured.
I am not sure if you can setup the SNMP stuff to generate to each error message a dedicated trap.
I am working since years with the syslog implementation successfully and get all importrant messages.
I hope this helps,