[Dynamic Buffer Calc] Enhance the logic to check maximum headroom exceeding to cover corner scenarios

[stephenxs]

What I did
Enhance the logic to check maximum headroom exceeding to cover corner scenarios

Currently, the logic to check the maximum headroom exceeding works well when a user changes any buffer configuration in the dynamic buffer model, preventing all problematic configurations from being applied to the ASIC.
However, it can fail when a problematic configuration is config_db.json and config reload is executed. To cover this scenario, the following actions need to be done:

1. Take the pending PG keys and buffer profiles into account when calculating the maximum headroom
   • Existing buffer PGs and buffer profiles can be in the pending queue since there are a large number of notifications needed to be handled during system initialization, which takes time.
2. Take the lossy PG into account when calculating the maximum headroom.
   • Non-default lossy PG can be added by the user in config_db.json
3. Pass the PG to the Lua plugin when refreshing PGs for a port

Signed-off-by: Stephen Sun stephens@nvidia.com

Why I did it

Cover corner scenarios.

How I verified it

Manual test, vs test, regression test (dynamic buffer)

Details if related

[stephenxs]

All test cases are failed due to the following error which should be fixed by sonic-net/sonic-utilities#2830

2023-05-08T08:09:14.5589099Z         # conn.request() calls http.client.*.request, not the method in
2023-05-08T08:09:14.5589452Z         # urllib3.request. It also calls makefile (recv) on the socket.
2023-05-08T08:09:14.5589655Z         try:
2023-05-08T08:09:14.5589756Z >           conn.request(
2023-05-08T08:09:14.5589871Z                 method,
2023-05-08T08:09:14.5590134Z                 url,
2023-05-08T08:09:14.5590232Z                 body=body,
2023-05-08T08:09:14.5590350Z                 headers=headers,
2023-05-08T08:09:14.5590462Z                 chunked=chunked,
2023-05-08T08:09:14.5590591Z                 preload_content=preload_content,
2023-05-08T08:09:14.5590721Z                 decode_content=decode_content,
2023-05-08T08:09:14.5590869Z                 enforce_content_length=enforce_content_length,
2023-05-08T08:09:14.5590986Z             )
2023-05-08T08:09:14.5591229Z E           TypeError: request() got an unexpected keyword argument 'chunked'

[stephenxs]

cherry-pickable to 202211 based on the lastest commit a484ab8684c45d56340bbb3a704ae6ca039e9282

[liat-grozovik]

/azp run Azure.sonic-swss

[azure-pipelines]

Azure Pipelines successfully started running 1 pipeline(s).

[stephenxs]

The same error as before

2023-05-09T04:05:18.8163888Z         # conn.request() calls http.client.*.request, not the method in
2023-05-09T04:05:18.8164229Z         # urllib3.request. It also calls makefile (recv) on the socket.
2023-05-09T04:05:18.8164485Z         try:
2023-05-09T04:05:18.8164690Z >           conn.request(
2023-05-09T04:05:18.8164887Z                 method,
2023-05-09T04:05:18.8165086Z                 url,
2023-05-09T04:05:18.8165277Z                 body=body,
2023-05-09T04:05:18.8165491Z                 headers=headers,
2023-05-09T04:05:18.8165720Z                 chunked=chunked,
2023-05-09T04:05:18.8165963Z                 preload_content=preload_content,
2023-05-09T04:05:18.8166233Z                 decode_content=decode_content,
2023-05-09T04:05:18.8166525Z                 enforce_content_length=enforce_content_length,
2023-05-09T04:05:18.8166774Z             )
2023-05-09T04:05:18.8167228Z E           TypeError: request() got an unexpected keyword argument 'chunked'
2023-05-09T04:05:18.8170916Z 
2023-05-09T04:05:18.8171542Z /usr/local/lib/python3.8/dist-packages/urllib3/connectionpool.py:496: TypeError

[keboliu]

@stephenxs maybe this PR can help to unblock the checker? #2767

[stephenxs]

@stephenxs maybe this PR can help to unblock the checker? #2767

thanks, @keboliu but I have already done it in the morning and the test failed in my case. checking.

[stephenxs]

/azpw run

[mssonicbld]

/AzurePipelines run

[azure-pipelines]

Azure Pipelines successfully started running 1 pipeline(s).

[stephenxs]

It turns out to be that the test failed on the Azure pipeline because the config_db updates were received in a different order in which they were received in my local vs testbed.
Add checkers to guarantee the order.

[stephenxs]

/azpw run

[mssonicbld]

/AzurePipelines run

[azure-pipelines]

Azure Pipelines successfully started running 1 pipeline(s).

[stephenxs]

Failed due to irrelevant test cases. Retriggered

2023-05-12T05:33:32.4509401Z test_virtual_chassis.py::TestVirtualChassis::test_chassis_add_remove_ports FAILED [ 65%]
2023-05-12T05:35:34.6375172Z test_vlan.py::TestVlan::test_VlanMemberLinkDown FAILED                   [ 80%]

[stephenxs]

@StormLiangMS could you please cherry-pick the PR to 202211? thanks.