03-20-2012 05:27 AM
Hi, I've been trying to understand the interaction between DCFM and Fabric Watch. Am I correct in the following?
- Fabric OS will just report events via commands like errdump and other logs on the switch
- Fabrc Watch enables SNMP traps, thresholds and performance montoring
- DCFM brings togther the Fabric Watch messages/traps to a central reporting tool
If this is the case, do I need to worry about the Fabric Watch configuration? Does DCFM use the Fabric Watch configuration or override it with it's own config?
Part of the reason I'm getting confused is that if I go to DCFM > Fabric Watch > Configuration, for example I take a look at the alarm configuration for Port, nothing is selected. However, if I disable a port a SNMP trap is generated. Why?
If anyone can explain how it hangs together, I'd be most grateful.
03-27-2012 03:49 AM
For those interested, after further research and testing, I have concluded the following:
- out of the box Fabric OS can be configured to send both snmp traps and syslogs to other servers
- if you configure the DCFM management server to be one of the recipients, it will process both the syslog and snmp traps. In our case, the administrator account is configured to get emails (a team email account) if either the syslog entry or snmp trap is severity level error or above.
- Fabric Watch is not required to run DCFM except for performance monitoring, e.g. in our case, we monitor ISL utilization and Fabric Watch is required for this.
So in the example that was confusing me, the port disable creating a SNMP trap and alert in openview, this happens for the following reason. On the switch, snmpmibcapshow shows that swFCPortScn in SW-TRAP is enabled so a SNMP trap is generated. Fabric Watch ignores this trap because port monitoring is disabled. DCFM receives the trap (this can be seen in the master log) but does nothing with it as the severity level is warning. However, Openview did generate an alert as it alerts on warning and above.