test for align watermark flow with port configuration

[dbarashinvd]

Description of PR

Summary:
Fixes # (issue)

Type of change

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

Back port request

• 201911
• 202012
• 202205

Approach

What is the motivation for this PR?

add a verification test for feature align watermark flow with port configuration

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

[roy-sror]

Hi Doron,

1. To cover the different scenarios, I suggest using pytest parametrization. It has some advantages comparing to the existing implementation.
2. There's no usage of allure step, it would be much more readable to have allure step when running the test with allure reporter.
3. Did you consider adding tests to cover the functional correctness of the counters?

[dbarashinvd]

Hi Doron,

1. To cover the different scenarios, I suggest using pytest parametrization. It has some advantages comparing to the existing implementation.
2. There's no usage of allure step, it would be much more readable to have allure step when running the test with allure reporter.
3. Did you consider adding tests to cover the functional correctness of the counters?

since we agreed to use randomization, didn't use pytest parametrization
added alure steps
random choice of reload/reboot and queue/watermark/pg-drop minimized the test duration to ~ 500 seconds.

[JibinBao]

It seems there is missing a test: "Adding a port when watermark is enabled" which is defined in HLD.
Can we need to add it?

[dbarashinvd]

It seems there is missing a test: "Adding a port when watermark is enabled" which is defined in HLD. Can we need to add it?

Hi Jibin,
this is a test defined in the feature implementation HLD, not the test plan HLD.
it's covered in a unit test and is part of the public PR for the implementation in sonic-swss 2525

Also note that it's not just adding a port now since there was an improvement adding counterpoll flex counters only for ports with queue or pg buffers: sonic-swss 2473

so now it's creating a buffer queue and buffer pg for a certain port and checks the flex counters created for it

if needed I can add such a test also for this test here but keep in mind it'll make the test longer than it's now, about 500 seconds after randomization requested by Roy.

[JibinBao]

It seems there is missing a test: "Adding a port when watermark is enabled" which is defined in HLD. Can we need to add it?

Hi Jibin, this is a test defined in the feature implementation HLD, not the test plan HLD. it's covered in a unit test and is part of the public PR for the implementation: sonic-swss 2525

Also note that it's not just adding a port now since there was an improvement adding counterpoll flex counters only for ports with queue or pg buffers: sonic-swss 2473 so now it's creating a buffer queue and buffer pg for a certain port and checks the flex counters created for it

if needed I can add such a test also for this test here but keep in mind it'll make the test longer than it's now, about 500 seconds after randomization requested by Roy.

Hi dbarashinvd,
I am not sure if the unit test can cover this scenario or not. If yes, I think it is ok. Otherwise, we will lose coverage.