[ptf_runner_reboot] cleanup ARP/FDB after test

[stepanblyschak]

Signed-off-by: Stepan Blyschak [email protected]

Description of PR

Summary:
Fixes # (issue)

Type of change

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

Approach

How did you do it?

How did you verify/test it?

Any platform specific information?

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

Documentation

[stepanblyschak]

@neethajohn The motivation is to cleanup after test.

E.g in our regression run we have the following order of test execution:
Fast-Reboot, CRM. After fast reboot there are FDB, ARP entries untill they expire. CRM test fails when some FDB expires, because it affects other resources count.

What is the problem with cleanup? What was the root cause of t1->server traffic loss with fdb flush?
The test should repopulate FDB

[liat-grozovik]

IMO each test should be standalone and make sure the next one is not affected.

[lguohan]

for warm reboot, you should not clear the mac, since it can cause data plane disruption.

[stepanblyschak]

@lguohan This is after the test, after PTF run finished we don't run test traffic so don't care about disruption

[neethajohn]

@neethajohn The motivation is to cleanup after test.
E.g in our regression run we have the following order of test execution:
Fast-Reboot, CRM. After fast reboot there are FDB, ARP entries untill they expire. CRM test fails when some FDB expires, because it affects other resources count.
What is the problem with cleanup? What was the root cause of t1->server traffic loss with fdb flush?
The test should repopulate FDB

Let me try out this patch on the platforms where I was seeing an issue earlier.

I ran the test multiple times. I am unable to get the fast reboot to run after 1 iteration of the test run (after arp and fdb clear). Received packet count is always 1 or 2 less than send for t1->server. Can you add this change conditionally only for Mellanox SKU's?