One cluster network port does not return home after upgrading one C3132Q switch
Applies to
- 40G cluster network ports on 10/40 Gigabit Ethernet Controller XL710 QSFP+ card
- Two cluster LIFs per node and port.
- Cisco N3K-C3132Q-V cluster network switches (CNS)
- NX-OS 7.0(3)I7(9) with Nexus_3132QV_RCF_v1.6-Cluster-HA-Breakout
- NX-OS 7.0(3)I7(10) with Nexus-3132QV-RCF-v1.8-Cluster-Breakout
Issue
- First Cluster Network Switch Nexus N3K-C3132Q-V model updated, from NX-OS 7.0(3)I7(9) to 7.0(3)I7(10) and new RCF 1.8 applied
- When just one cluster network port in one node is online, multiple rings in that node are offline (vifmgr, vldb and mgwd) with
CLUSTER NETWORK DEGRADED
errors. - No more errors when that port is manually offlined and the cluster LIF changes the home port.
- Cable re-seat/replace/swap doesn't fix the issue.
- Controller port is properly detected in the switch side.
- All cluster network pings fail between LIFs on both switches but work for LIFs connected to the same switch.