Clear fabric counters queue/port

[jfeng-arista]

What I did

Raise a PR ( the same as sonic-net/sonic-utilities#2892 ) in the https://github.com/Azure/sonic-utilities.msft/tree/202205

How I did it

How to verify it

tested under https://github.com/Azure/sonic-buildimage-msft/tree/202205

Previous command output (if the output of a command-line utility has changed)

New command output (if the output of a command-line utility has changed)

[jfeng-arista]

@gechiang This the clear fabric counters pr for 202205.

[gechiang]

@jfeng-arista Thanks! BUt this is rather interesting. From this PR it looks like the changes are exactly the same as that we previously cherry-picked to here... Did I missed anything? If there is no change, then how were you able to ensure it is fixed and not causing the build issue that we saw previously?

[jfeng-arista]

@jfeng-arista Thanks! BUt this is rather interesting. From this PR it looks like the changes are exactly the same as that we previously cherry-picked to here... Did I missed anything? If there is no change, then how were you able to ensure it is fixed and not causing the build issue that we saw previously?

The major change is the cli. I guess the cli parse infra changes are not here yet,

command = ["fabricstat", "-C", "-q"] is the one in sonic-net/sonic-utilities#2892
here it still use the old way
command = "fabricstat -C -q"

I tried to copy all the files in https://github.com/Azure/sonic-buildimage-msft/tree/202205 locally and tried to run the sonic-utilities tests and saw the test failures, and I revised code based on what I see and they passed in my local testing, so I raise the PR based on that test in this repo.

Though I don't know exactly what the errors you saw before ( there is no error log / information for the previous failure you mentioned ). All my changes are based on my local testing.

[gechiang]

@jfeng-arista Thanks! BUt this is rather interesting. From this PR it looks like the changes are exactly the same as that we previously cherry-picked to here... Did I missed anything? If there is no change, then how were you able to ensure it is fixed and not causing the build issue that we saw previously?

The major change is the cli. I guess the cli parse infra changes are not here yet,

command = ["fabricstat", "-C", "-q"] is the one in sonic-net/sonic-utilities#2892 here it still use the old way command = "fabricstat -C -q"

I tried to copy all the files in https://github.com/Azure/sonic-buildimage-msft/tree/202205 locally and tried to run the sonic-utilities tests and saw the test failures, and I revised code based on what I see and they passed in my local testing, so I raise the PR based on that test in this repo.

Though I don't know exactly what the errors you saw before ( there is no error log / information for the previous failure you mentioned ). All my changes are based on my local testing.

Thanks. I will have to hold off on picking up this for now until I get the build team to add the necessary PR tests into the pipeline for the MSFT repo so that we can better identify if a PR is really safe without build issues... last time when I picked it up we did not realize there was issue with it until community member and myself started to build with it... I will try to experiment picking up this change in my local branch and attempt a build there to see if the problem is somehow fixed...
Will update you... Thanks!

[jfeng-arista]

@jfeng-arista Thanks! BUt this is rather interesting. From this PR it looks like the changes are exactly the same as that we previously cherry-picked to here... Did I missed anything? If there is no change, then how were you able to ensure it is fixed and not causing the build issue that we saw previously?

The major change is the cli. I guess the cli parse infra changes are not here yet,
command = ["fabricstat", "-C", "-q"] is the one in sonic-net/sonic-utilities#2892 here it still use the old way command = "fabricstat -C -q"
I tried to copy all the files in https://github.com/Azure/sonic-buildimage-msft/tree/202205 locally and tried to run the sonic-utilities tests and saw the test failures, and I revised code based on what I see and they passed in my local testing, so I raise the PR based on that test in this repo.
Though I don't know exactly what the errors you saw before ( there is no error log / information for the previous failure you mentioned ). All my changes are based on my local testing.

Thanks. I will have to hold off on picking up this for now until I get the build team to add the necessary PR tests into the pipeline for the MSFT repo so that we can better identify if a PR is really safe without build issues... last time when I picked it up we did not realize there was issue with it until community member and myself started to build with it... I will try to experiment picking up this change in my local branch and attempt a build there to see if the problem is somehow fixed... Will update you... Thanks!

thank you

[jfeng-arista]

@jfeng-arista Thanks! BUt this is rather interesting. From this PR it looks like the changes are exactly the same as that we previously cherry-picked to here... Did I missed anything? If there is no change, then how were you able to ensure it is fixed and not causing the build issue that we saw previously?

The major change is the cli. I guess the cli parse infra changes are not here yet,
command = ["fabricstat", "-C", "-q"] is the one in sonic-net/sonic-utilities#2892 here it still use the old way command = "fabricstat -C -q"
I tried to copy all the files in https://github.com/Azure/sonic-buildimage-msft/tree/202205 locally and tried to run the sonic-utilities tests and saw the test failures, and I revised code based on what I see and they passed in my local testing, so I raise the PR based on that test in this repo.
Though I don't know exactly what the errors you saw before ( there is no error log / information for the previous failure you mentioned ). All my changes are based on my local testing.

Thanks. I will have to hold off on picking up this for now until I get the build team to add the necessary PR tests into the pipeline for the MSFT repo so that we can better identify if a PR is really safe without build issues... last time when I picked it up we did not realize there was issue with it until community member and myself started to build with it... I will try to experiment picking up this change in my local branch and attempt a build there to see if the problem is somehow fixed... Will update you... Thanks!

thank you

I was helped with others here during the weekend to try to pull the full source locally, and the build of broadcom ( which is the one close to upstream ) passed. This makes me more confident about this change within the CURRRENT code base. thank you