[VRF]: submit vrf feature

[tylerlinp]

What I did
Submit VRF implementation according to vrf HLD #392

Why I did it
Following sonic 201908 roadmap

How I verified it
Pass vrf vs tests case and ansible test cases

Details if related

[lguohan]

this PR is too huge, can you break down into multiple small patches? there is no possibilities that this huge PR can be reviewed properly.

[tylerlinp]

this PR is too huge, can you break down into multiple small patches? there is no possibilities that this huge PR can be reviewed properly.

I had try to split but found it has heavy dependence. Huge codes have to rewite for any PR splited, furthermore it is difficult to test then. If truely rewrite to several PRs, it would be a big trouble to merge them. For example the PR 878, it is just the content of "move subnet route from intfsorch to routeorch", which is the version based origin master. But after composing other parts, it had undergone enormous changes. So it is meaningless to review PR 878 or something similar.
For esaier review, I suggest to recommit to a single PR, spliting to several commits. Maybe split based process, vrfmgrd/intfmgrd/nbrmgrd/neighsyncd/fpmsyncd/ORCHAGENT/TESTS. How do you think about?

[prsunny]

@tylerlinp , we have planned to split based on subcomponents. We won't be able to review/merge this big change. Can you check on this list:

a. One PR to change Nexthop/Neighbor Keys to include interface name. (This is generic irrespective of VRF)
b. One PR to move subnet routes to routeorch and cleanup interface orch (This is generic irrespective of VRF)
c. One PR to handle loopback interface changes
d. One PR for the rest of VRF changes.

[tylerlinp]

retest this please

[prsunny]

retest this please

[prsunny]

retest this please

[shine4chen]

@prsunny We notice that recently test_FdbAddedAfterMemberCreated testcases almost failed in all PR vs. We will help to dig it next week. We speculate that it may have something to do with vs test topology or environment.

[prsunny]

@shine4chen , thanks. That would be helpful.

[tylerlinp]

retest this please

[prsunny]

Hi Tyler, I see that the FDB test is still failing, may be it needs a sleep before checking? Also please don't change the test in this PR. It has to be another PR which can be merged and then retest this.

[tylerlinp]

@prsunny The new failure is result from a bug that is_fdb_entry_exists didn't match input key, so when fdb table is not empty then get true. I will revert this test change later and raise a new PR.

[tylerlinp]

@prsunny PR #1118 fix the fdb test failure. This PR will revert to commit 7987477 and retest.

[prsunny]

@shine4chen , thanks. Can you please resolve confict

[tylerlinp]

retest this please