[BGP] Implement new testcase test_traffic_shift

[bingwang-ms]

Description of PR

Summary:
Fixes # (issue)
Implement a new testcase test_traffic_shift.

• case 1 test_TSA
  Verify all routes are announced to bgp monitor, and only loopback routes are announced to neighs.

1. Start exabgp on ptf, and establise BGP session with DUT
2. Setup BGPMON over the BGP session in step 1
3. Issue TSA on DUT, and wait some time (say 30 secs)
4. Verify all routes are announced to BGPMON on ptf
5. Verify only loopback routes with certain community are announced to bgp neighbors

• case 2 test_TSB
  Verify all routes are announced to bgp monitor and all routes are announced to neighbors.

1. Start exabgp on ptf, and establise BGP session with DUT
2. Setup BGPMON over the BGP session in step 1
3. Issue TSB on DUT, and wait some time (say 30 secs)
4. Verify all routes are announced to BGPMON on ptf
5. Verify all routes are announced to bgp neighbors

Please merge after sonic-net/sonic-buildimage#5906 and sonic-net/sonic-buildimage#5968 are merged.

Type of change

• Bug fix
• Testbed and Framework(new/improvement)
• Test case(new/improvement)

Approach

What is the motivation for this PR?

This PR is to implement new testcase test_traffic_shift.

How did you do it?

1. Use exabgp to dump all received routes on BGPMON, and parsed into a set;
2. Retrieve all routes from ceos and parsed into a dict;
3. Retrieve all routes from dut and parsed into a dict;
4. Compare routes retrieved in step 3 with that retrieved in step 1 and step 2.

How did you verify/test it?

Verified on vtestbed t1

py.test --inventory veos_vtb --host-pattern vlab-01 --module-path ../ansible/library/ --testbed vms-kvm-t1 --testbed_file vtestbed.csv --junit-xml=tr.xml --log-cli-level warn -k 'not test_restart_syncd' --topology=t1,t0,any,util bgp/test_traffic_shift.py
========================================================================================= test session starts =========================================================================================
collected 2 items                                                                                                                                                                                     

bgp/test_traffic_shift.py::test_TSA ^@PASSED                                                                                                                                                      [ 50%]
bgp/test_traffic_shift.py::test_TSB ^@PASSED                                                                                                                                                      [100%]

-------------------------------------------------------------------------- generated xml file: /data/sonic-mgmt/tests/tr.xml --------------------------------------------------------------------------
===================================================================================== 2 passed in 822.05 seconds ======================================================================================

Any platform specific information?

No.

Supported testbed topology if it's a new test case?

Only run on T1 topo.

Documentation

[lgtm-com]

This pull request introduces 1 alert when merging c3f2fba into 7c5d15a - view on LGTM.com

new alerts:

• 1 for Should use a 'with' statement

[pavel-shirshov]

Please check my comments

[pavel-shirshov]

The argument name is not dump_script, but default_dump_script
As I see it here:
https://github.com/Azure/sonic-mgmt/pull/2553/files#diff-ca868a64bc6e0673c7630841edad82bd6de93e60ea7a7ba694c87467d69c8f2eR209

[pavel-shirshov]

Should we start bgpmon before TSA?

[bingwang-ms]

I think bgpmon should be started after TSA. That is because if we start bgpmon, and then set TSA, there will be a short TSB (normal) state between two operations, during which routes are possible to be announced. We might get a failure then.

[pavel-shirshov]

LGTM