Fix qos/test_qos_dscp_mapping.py#13119
Conversation
This test is mainly failing due to:
- Sending 10K packets at once on a ptf interface is causing nnpy
socket to be non-functional for a moment sometimes and connection
is getting timedout.
- Sometimes egress_queue_count on queue 7 is way greater than no of
packets being sent and the queue_count is not within the TOLERANCE,
this could be due to protocol packets.
- Test is not enhanced to be run on dualtor topologies (packets are
going to unexpected ToR).
- Fixtures "upstream_links" and "downstream_links" are always
referring to upper ToR, thus return values of these fixtures is not
really considering the ToR type and causing failures.
|
The pre-commit check detected issues in the files touched by this pull request. Detailed pre-commit check results: To run the pre-commit checks locally, you can follow below steps:
|
|
The pre-commit check is mandatory. Please check. |
pre-commit check is being failed with "toggle_all_simulator_ports_to_rand_selected_tor" imported but unused error. Added "# noqa" to suppress this failure as this import is being used as fixture argument (similar to other tests).
Thanks took care of this. |
|
@XuChen-MSFT FYI |
|
@vkjammala-arista Do you need to backport to 202311 branch? |
|
hi @XuChen-MSFT @developfast could you help to review this PR also? |
developfast
left a comment
There was a problem hiding this comment.
changes lgtm - thanks for the fixture change to provide support on dualtor
|
@yxieca can you pls help merge? |
Hi @wsycqyz, we need this fix in 202311 as well (as mentioned in description). Thanks! |
* Fix qos/test_qos_dscp_mapping.py
This test is mainly failing due to:
- Sending 10K packets at once on a ptf interface is causing nnpy
socket to be non-functional for a moment sometimes and connection
is getting timedout.
- Sometimes egress_queue_count on queue 7 is way greater than no of
packets being sent and the queue_count is not within the TOLERANCE,
this could be due to protocol packets.
- Test is not enhanced to be run on dualtor topologies (packets are
going to unexpected ToR).
- Fixtures "upstream_links" and "downstream_links" are always
referring to upper ToR, thus return values of these fixtures is not
really considering the ToR type and causing failures.
* Fix pre-commit check failures
pre-commit check is being failed with
"toggle_all_simulator_ports_to_rand_selected_tor" imported but unused
error. Added "# noqa" to suppress this failure as this import is being
used as fixture argument (similar to other tests).
|
Cherry-pick PR to 202405: #13247 |
* Fix qos/test_qos_dscp_mapping.py
This test is mainly failing due to:
- Sending 10K packets at once on a ptf interface is causing nnpy
socket to be non-functional for a moment sometimes and connection
is getting timedout.
- Sometimes egress_queue_count on queue 7 is way greater than no of
packets being sent and the queue_count is not within the TOLERANCE,
this could be due to protocol packets.
- Test is not enhanced to be run on dualtor topologies (packets are
going to unexpected ToR).
- Fixtures "upstream_links" and "downstream_links" are always
referring to upper ToR, thus return values of these fixtures is not
really considering the ToR type and causing failures.
* Fix pre-commit check failures
pre-commit check is being failed with
"toggle_all_simulator_ports_to_rand_selected_tor" imported but unused
error. Added "# noqa" to suppress this failure as this import is being
used as fixture argument (similar to other tests).
|
@vkjammala-arista PR conflicts with 202305 branch |
|
Cherry-pick PR to 202311: #13260 |
|
hi @vkjammala-arista, could you file separate PR for 202305, it is a conflict. |
Hi @StormLiangMS, this change is NA for 202305 (test doesn't exist there). Thanks! |
* Fix qos/test_qos_dscp_mapping.py
This test is mainly failing due to:
- Sending 10K packets at once on a ptf interface is causing nnpy
socket to be non-functional for a moment sometimes and connection
is getting timedout.
- Sometimes egress_queue_count on queue 7 is way greater than no of
packets being sent and the queue_count is not within the TOLERANCE,
this could be due to protocol packets.
- Test is not enhanced to be run on dualtor topologies (packets are
going to unexpected ToR).
- Fixtures "upstream_links" and "downstream_links" are always
referring to upper ToR, thus return values of these fixtures is not
really considering the ToR type and causing failures.
* Fix pre-commit check failures
pre-commit check is being failed with
"toggle_all_simulator_ports_to_rand_selected_tor" imported but unused
error. Added "# noqa" to suppress this failure as this import is being
used as fixture argument (similar to other tests).
* Fix qos/test_qos_dscp_mapping.py
This test is mainly failing due to:
- Sending 10K packets at once on a ptf interface is causing nnpy
socket to be non-functional for a moment sometimes and connection
is getting timedout.
- Sometimes egress_queue_count on queue 7 is way greater than no of
packets being sent and the queue_count is not within the TOLERANCE,
this could be due to protocol packets.
- Test is not enhanced to be run on dualtor topologies (packets are
going to unexpected ToR).
- Fixtures "upstream_links" and "downstream_links" are always
referring to upper ToR, thus return values of these fixtures is not
really considering the ToR type and causing failures.
* Fix pre-commit check failures
pre-commit check is being failed with
"toggle_all_simulator_ports_to_rand_selected_tor" imported but unused
error. Added "# noqa" to suppress this failure as this import is being
used as fixture argument (similar to other tests).
Description of PR
Summary: Fix qos/test_qos_dscp_mapping.py failures
Fixes # aristanetworks/sonic-qual.msft#132
Type of change
Back port request
Approach
What is the motivation for this PR?
This test is mainly failing due to:
How did you do it?
Fixes involve adding right fixtures to the test methods (to consider dualtor topologies) and minor changes to address above issues.
How did you verify/test it?
With the fix test is passing well (earlier it was failing most of the times with one or other reason). Verified the fix on Arista-7260CX3-C64 202311 branch.
Any platform specific information?
Supported testbed topology if it's a new test case?
Documentation