Symptom
When configured with "switchport dot1q ethertype 0x88a8", packets are encapsulated in 0x8100 after chassis reload
Ethernet II, Src: 70:6d:15:41:6b:43 (70:6d:15:41:6b:43), Dst: e4:c7:22:dc:f8:01 (e4:c7:22:dc:f8:01)
Destination: e4:c7:22:dc:f8:01 (e4:c7:22:dc:f8:01)
Address: e4:c7:22:dc:f8:01 (e4:c7:22:dc:f8:01)
.... ...0 .... .... .... .... = IG bit: Individual address (unicast)
.... ..0. .... .... .... .... = LG bit: Globally unique address (factory default)
Source: 70:6d:15:41:6b:43 (70:6d:15:41:6b:43)
Address: 70:6d:15:41:6b:43 (70:6d:15:41:6b:43)
.... ...0 .... .... .... .... = IG bit: Individual address (unicast)
.... ..0. .... .... .... .... = LG bit: Globally unique address (factory default)
Type: 802.1Q Virtual LAN (0x8100) <â Remains 0x8100
802.1Q Virtual LAN, PRI: 0, CFI: 0, ID: 82
000. .... .... .... = Priority: 0
...0 .... .... .... = CFI: 0
.... 0000 0101 0010 = ID: 82
Type: 802.1Q Virtual LAN (0x8100)
802.1Q Virtual LAN, PRI: 0, CFI: 0, ID: 10
000. .... .... .... = Priority: 0
...0 .... .... .... = CFI: 0
.... 0000 0000 1010 = ID: 10
Type: IP (0x0800)
Conditions
Trunk port on BCM based n9k/n3k with non-default ethertype setting (0x88a8 or 0x9100).
Upon reload, ethertype setting on that port reverts to default =0x8100.
Workaround
1. default the ethertype after port flap
no switchport dot1q ethertype
2. Re-apply the required ethertype
switchport dot1q ethertype
Further Problem Description
See also CSCvu24785, where the issue with reverting back to default=0x8100 after link flap addressed starting from 9.3(5) NXOS version.
