we have full mesh network with fabric path between N7700 (spine) and Nexus 5600 (leaf). There are Nexus 2200 FEX-es dual-homed to Nexus 5600.
We have problem with traffic loss after Nexus 5600 reloads. Fabric path is up, FEX-es are still down and spine switch forwards traffic (because FP is up) to restarted leaf switch and it drops traffic (it should send it to vPC peer over peer-link). Traffic starts after FEX come online on restarted Nexus 5600.
We opened TAC case and TAC said that this is BUG and for resolving it they need to develop new feature (there is no any date for this).
Workaround are EEM and python scripts.
EEM script should shutdown fabric path uplink ports after switch boots up (match PFMA-5-MOD_STATUS: Module 1 current-status is MOD_STATUS_ONLINE) and start python script.
Python script should wait some time (timer should be addopted to number of FEX-es, more FEX-es means higher timer value) and after timer expires (all FEX-es come online) it should enable fabric path uplink ports.
Original Blog published @ https://blogs.cisco.com/datacenter/spinning-up-an-nvme-over-fibre-channel-strategy
Every so often there comes a time when we witness a major shift in the networking industry that fundamentally changes the landscap...
Let me start by telling you that there is already a very nice document in the Cisco Community Forums which describes the procedure to upgrade the APIC CIMC. In case you came across this document before looking into the mentioned one, here is...