Well, “This is currently working” kinda doesn’t really matter if you have vlan filtering off - packets are just zipping around through all the ports that are non-disabled members of the bridge (sfp 2, 3, 5, 6, 10, 11, 15, 16), preserving their existing tags (or lack of).
The config as you posted above is definitely a lot of trouble with all the emergency changes and leftover experiments.
- engage safe mode
- possibly temporarily disable the physical ports which are anyway disabled in bridge config under /interface/ethernet (sfp 1, 4, 8, 9, 12, 13, 14) so they won’t get in the way
- delete the unused bridge1 and bridge2
- clean up all the entries referring to invalid/removed items (marked with asterisks, e.g. “interface=*1B”) and all the disabled stuff, to get to the bare minimum of what’s currently active
At this point things should work the same as until now, if it is the case, commit and re-enable safe mode as a checkpoint.
I think you should also remove that layer3 vlan17 interface attached to sfp16, because you can’t have it attached to a bridge member like that.
It may be what’s disabling hardware offloading for the whole switch, although I’d sort of expect for it to just show up as invalid.
For every access port that should auto insert vlan tag on ingress / and remove on egress
- under /interface/bridge/port, add this port, set mode to “admit only untagged”, set pvid 1234
- under /interface/bridge/vlan make sure to have entry for vlan 1234, add this port to untagged list
For every trunk port that should take tagged stuff from outside, and emit tagged stuff as well (234, 456, 789)
- under /interface/bridge/port, add this port, set mode to “admit only vlan tagged”, set pvid to 1
- under /interface/bridge/vlan make sure to have entry for each of the vlans 234, 456, 789, and in those existing entries put the port in the tagged list
For every hybrid port that should do both kinds of traffic
- under /interface/bridge/port, add this port, set mode to “admit all”, set pvid to what will be used for untagged/access (native in ciscospeak)
- under /interface/bridge/vlan add to untagged list for the access vlan
- under /interface/bridge/vlan add to tagged list for all the other trunked relevant vlans
if you have stuff that is today set for untagged (not the same as vlan 1) and you want to just tunnel it through, you can do it by access port in vlan 1 or you can use any other local temporary transit vlan as you for example have 1002 on your current config above as long as it won’t conflict with the vlans you want to pass-through with already-meaningful numbers.
After all this, you should still “feel no change”, if that’s the case, again commit and re-activate safe mode, and try to enable vlan filtering on the bridge…
PS.
Can you post whether you currently see HW-Offload flags on any ports?
“H” flag in the /interface/bridge/port listing.