Charles Chan

I'd suggest adding that in the troubleshooting section at the end of the exercise. You are welcome to submit a patch. I can review and merge it. Thanks.

Seems similar to https://github.com/p4lang/tutorials/issues/204

The error message seems suggesting that we should omit the entire match field instead of using wildcard mask. Have you tried removing the entire match field on `hdr.ethernet.dst`?

I was suggesting skipping the entire ethernet dst match field. Something like this: ``` final PiCriterion unmatchedTrafficCriterion = PiCriterion.builder().build(); ``` This should theoretically create a "catch all" match

Hmm... alright, looks like it doesn't like it either. Will need to take a closer look to figure out why the original PiCriterion get rejected.

Could you elaborate the purpose this patch? It seems that this new class is not invoked anywhere else.

We agreed to put this on hold during the stand up on 1/24. Will revisit this once we need to support dynamic update / move slicing service to ONOS core....

I also hit the same issue recently. Got around by retesting it.

Just providing numbers for record - in current scale, we have ~ 120 groups in a leaf switch. The group limit is 1024, which means we will hit the limit...

> Just providing numbers for record - in current scale, we have ~ 120 groups in a leaf switch. The group limit is 1024, which means we will hit the...