Headline: Nexus 9000 aclqos cores - ERSPAN w/source VLAN mapped to VNI on certain ports Workarounds: You need to have physical OIR of 100mb FX xcvr to recover or you need to reload the switch. Symptoms: Link may not come up after ISSU if you have a 100mb FX xcvr. Headline: 100M link is down on N9K-C93180YC-FX side and up on peer side after port flap on post ND ISSU. Workarounds: Remove and re-enable guestshell after the upgrade.

Symptoms: Running guestshell prompts for password which is unexpected after upgrading from switch with guestshell version 2.10 to NXOS version 9.3(8) or 10.1(2). Headline: NXOS - Additional prompt showing up when running guest-shell command.

Delete the particular mac address and allow mac to be learnt again.

This is causing VXLAN unicast traffic to get dropped.

Symptoms: Mac address behind FEX AA VPC is getting deleted on primary vpc peer after reload of switch (primary reloaded followed by secondary). Headline: Vxlan unicast traffic drop due to mac learnt on FEXAA HIFVPC getting deleted on vpc primary. After doing so, 'reload' or 'reload ascii' is required to restore connectivity. This will avoid this problem and also will ensure better hashing of packets over ECMP links.

Disable 'feature nv overlay' on spine.

Other unicast communication works fine (for example - when we configure static ARPs). In fact the ARP replies are redirected to SPINE CPU instead of being forwarded. The broadcasted ARP reply is flooded correctly and reaches all hosts, however the unicast ARP reply is lost inside of the fabric. Symptoms: The L2 adjacent host are not able to resolve each other’s ARP across VxLAN fabric. Headline: 9500-R :: Feature ptp causes the spine switch to intercept unicast ARP replies in VxLAN fabric

The switch is reloaded when the issue is hit. Headline: Crash in Nexus 9000 Fatal Module Error when downgrading - service port_client hap resetĭuring downgrade from 9.3.7 to 9.3.6, vPC peer switch reloads due to 'port_client' service crash:

Nexus vortex release date