We ran into a bug on the 3.3.2 code on our FCIP links. We did an upgrade to the SAN core to resolve the issue. My case was with EMC so here is the information from them.
Problem: Host is unable to communicate with disks using IVR because the Host's PLOGI does not reach the disk and IVR sends a PLOGI reject back to the host.
Root Cause: Cisco bug tracking number CSCsu38135. Access to this bug requires access to Cisco's Bug toolkit.
Fix: Workarounds:
- Have the host logout or login to the fabric (e.g. by flapping the port) multiple times. If this is done every 2 minutes, then 1 in 3 tries should catch the appropriate window, and the host should be able to communicate.
- Perform a FCNS refresh on the affected VSAN.
- Perform an IVR refresh.
Permanent fix: Fixed in SAN OS 3.3(3) and NX OS 4.1(1b).