The config_reload in the fixture dut_disable_ipv6 waits until all critical processes are up after issuing config reload command and it timeouts in case of dualtor because mux container doesn't come up. Mux container is disabled by another fixture stopServices in the same file. These two fixtures have no dependency on each other hence the execution can happen in any order, so if the teardown of dut_disable_ipv6 happens before stopServices then this issue is seen.
This change ensures that the teardown of stopServices happens before dut_disable_ipv6 so that mux is no longer disabled at the time of config_reload.
How did you verify/test it?
Ran qos/test_qos_sai.py on Arista-7260CX3 platform with dualtor topology with 202305 and 202311 images.
Any platform specific information?
Supported testbed topology if it's a new test case?
Description of PR
Summary: Fix qos/tes_qos_sai.py teardown failure for dualtor. Fixes #130
Type of change
Back port request
Approach
What is the motivation for this PR?
qos/test_qos_sai.py fails at teardown
Regression introduced by https://github.com/sonic-net/sonic-mgmt/pull/10651 for dualtor.
How did you do it?
The config_reload in the fixture
dut_disable_ipv6
waits until all critical processes are up after issuing config reload command and it timeouts in case of dualtor because mux container doesn't come up. Mux container is disabled by another fixturestopServices
in the same file. These two fixtures have no dependency on each other hence the execution can happen in any order, so if the teardown ofdut_disable_ipv6
happens beforestopServices
then this issue is seen.This change ensures that the teardown of
stopServices
happens beforedut_disable_ipv6
so that mux is no longer disabled at the time of config_reload.How did you verify/test it?
Ran qos/test_qos_sai.py on Arista-7260CX3 platform with dualtor topology with 202305 and 202311 images.
Any platform specific information?
Supported testbed topology if it's a new test case?
Documentation