Friends,
We have just brought up a cluster with two F740 filers, running ONTAP 5.3.6R2. They are connected by Gigabit Ethernet to a Cisco Catalyst 5500 switch. This is on a 9 port Gigabit Etherchannel module. We didn't get a chance to test the failover behavior, and have questions for anyone who might be running a similar configuration.
Is the failover instantaneous, as suggested by our NetApp sales rep and CE, or is there a delay while things sort themselves out in the switch? Is there special configuration on the switch port that should be configured for these (e.g. "set spantree portfast") that should be applied?
Thanks for your insights...
Regards,
Daniel Oxenhandler Senior Network Engineer Sprint Laboratories 650-375-4281
Is the failover instantaneous, as suggested by our NetApp sales rep and CE, or is there a delay while things sort themselves out in the switch?
Its not instantaneous, there's 15 seconds or so until the partner kicks in, IIRC. As for the switch, as long as you're not doing things like port security or static spantree databases, you should be fine. 'set spantree portfast' is only useful when a port transistions from down to up. In the case of the cluster failover, the interface on the surviving partner just gets an additional mac addr..
..kg..