Skip to content

[Mellanox] Update test_exceeding_headroom due to adding checker when enabling SHP #14479

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
liat-grozovik merged 1 commit into from
Dec 3, 2024
Merged

[Mellanox] Update test_exceeding_headroom due to adding checker when enabling SHP #14479

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
91 changes: 46 additions & 45 deletions tests/qos/test_buffer.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2384,24 +2384,24 @@ def _get_max_speed_from_list(speed_list_str):


@pytest.mark.disable_loganalyzer
def test_exceeding_headroom(duthosts, rand_one_dut_hostname, conn_graph_facts, port_to_test): # noqa F811
"""The test case for maximum headroom

If the accumulative headroom of a port exceeds the maximum value,
the new configuation causing the violation should not be applied to prevent orchagent from exiting
@pytest.mark.parametrize("disable_shp", [True, False])
def test_exceeding_headroom(duthosts, rand_one_dut_hostname, conn_graph_facts, port_to_test, disable_shp): # noqa F811
"""The test case is to verify If the accumulative headroom(shared headroom) of a port exceeds the maximum threshold,
the relevant configuration should not be applied successfully, and there will are the corresponding error logs.

Args:
port_to_test: Port to run the test

The flow of the test case:
1. Find the longest possible cable length the port can support.
It will also verify whether a super long cable will be applied
The test will be skipped if such limit isn't found after the cable length has been increased to 2km.
The test will be skipped if such limit isn't found after the cable length has been increased to 10km.
2. Add extra PGs to a port, which causes the accumulative headroom exceed the limit
3. Configure a headroom-override on a port and then enlarge the size of the profile.
Verify whether the large size is applied.
4. Configure a long cable length with shared headroom pool enabled.
Verify the size in the profile is updated when shared headroom pool is disabled.
3. Configure a headroom-override on a port and then enlarge the headroom of the profile(when SHP is disabled,
the headroom is size. When SHP is enabled, the headroom is xoff).
Verify the config cannot be applied to the profile
4. Configure a violating cable length which causing the headroom exceed the limit threshold.
Verify the relevant pg table for the violating cable length doesn't exist in app db

In each step, it also checks whether the expected error message is found.
"""
Expand All @@ -2428,13 +2428,15 @@ def test_exceeding_headroom(duthosts, rand_one_dut_hostname, conn_graph_facts, p
'redis-cli hget BUFFER_POOL_TABLE:ingress_lossless_pool xoff')['stdout']

try:
# Test case runs with shared headroom pool disabled
# because the headroom size is very small with shared headroom pool enabled
if original_over_subscribe_ratio and original_over_subscribe_ratio != '0':
duthost.shell(
'config buffer shared-headroom-pool over-subscribe-ratio 0')
if original_configured_shp_size and original_configured_shp_size != '0':
duthost.shell('config buffer shared-headroom-pool size 0')
if disable_shp:
logging.info("shp is disabled")
if original_over_subscribe_ratio and original_over_subscribe_ratio != '0':
duthost.shell(
'config buffer shared-headroom-pool over-subscribe-ratio 0')
if original_configured_shp_size and original_configured_shp_size != '0':
duthost.shell('config buffer shared-headroom-pool size 0')
else:
duthost.shell('config buffer shared-headroom-pool over-subscribe-ratio 2')

# 1. Find the longest possible cable length the port can support.
loganalyzer, marker = init_log_analyzer(
Expand Down Expand Up @@ -2585,16 +2587,23 @@ def test_exceeding_headroom(duthosts, rand_one_dut_hostname, conn_graph_facts, p
['BUFFER_PROFILE .* cannot be updated because .* referencing it violates the resource limitation',
'Unable to update profile for port .*. Accumulative headroom size exceeds limit'])

logging.info('[Update headroom override to a larger size]')
duthost.shell(
'config buffer profile set test-headroom --size {}'.format(int(maximum_profile['size']) * 2))
def _update_headroom_exceed_Larger_size(param_name):
logging.info(
'[Update headroom exceed the headroom threshold with the 2*maximum_profile[param_name]]')
duthost.shell(
f'config buffer profile set test-headroom --{param_name} {int(maximum_profile[param_name]) * 2}')

# This should make it exceed the limit, so the profile should not applied to the APPL_DB
time.sleep(20)
size_in_appldb = duthost.shell(
'redis-cli hget "BUFFER_PROFILE_TABLE:test-headroom" size')['stdout']
pytest_assert(size_in_appldb == maximum_profile['size'],
'The profile with a large size was applied to APPL_DB, which can make headroom exceeding')
# This should make it exceed the limit, so the profile should not applied to the APPL_DB
time.sleep(20)
size_in_appldb = duthost.shell(
f'redis-cli hget "BUFFER_PROFILE_TABLE:test-headroom" {param_name}')['stdout']
pytest_assert(size_in_appldb == maximum_profile[param_name],
f'The profile with a large size was applied to APPL_DB, which can make headroom exceeding. '
f'size_in_appldb:{size_in_appldb}, '
f'maximum_profile_{param_name}: {maximum_profile[param_name]}')

param_name = "size" if disable_shp else "xoff"
_update_headroom_exceed_Larger_size(param_name)

# Check log
check_log_analyzer(loganalyzer, marker)
Expand All @@ -2604,36 +2613,28 @@ def test_exceeding_headroom(duthosts, rand_one_dut_hostname, conn_graph_facts, p
'config interface buffer priority-group lossless set {} {}'.format(port_to_test, '3-4'))
duthost.shell('config buffer profile remove test-headroom')

# 4. Configure a long cable length with shared headroom pool enabled.
loganalyzer, marker = init_log_analyzer(
duthost,
'Toggle shared headroom pool',
['BUFFER_PROFILE .* cannot be updated because .* referencing it violates the resource limitation',
'Unable to update profile for port .*. Accumulative headroom size exceeds limit',
'refreshSharedHeadroomPool: Failed to update buffer profile .* when toggle shared headroom pool'])
['.*Unable to update profile for port .*. Accumulative headroom size exceeds limit',
'.*ERR swss#buffermgrd: :- doTask: Failed to process table update.*',
'.*ERR swss#buffermgrd: :- refreshPgsForPort: Update speed .* and cable length .* for port.* failed,'
' accumulative headroom size exceeds the limit.*'])

# Enable shared headroom pool
duthost.shell(
'config buffer shared-headroom-pool over-subscribe-ratio 2')
time.sleep(20)
# And then configure the cable length which causes the accumulative headroom exceed the limit
duthost.shell(
'config interface cable-length {} {}m'.format(port_to_test, violating_cable_length))
expected_profile = make_expected_profile_name(
original_speed, '{}m'.format(violating_cable_length))
check_pg_profile(
duthost, 'BUFFER_PG_TABLE:{}:3-4'.format(port_to_test), expected_profile)

# Disable shared headroom pool
duthost.shell(
'config buffer shared-headroom-pool over-subscribe-ratio 0')
time.sleep(20)
# Make sure the size isn't updated
profile_appldb = _compose_dict_from_cli(duthost.shell(
'redis-cli hgetall BUFFER_PROFILE_TABLE:{}'.format(expected_profile))['stdout'].split('\n'))
assert profile_appldb['xon'] == profile_appldb['size']

# Check log
# Make sure the profile isn't updated
# This pg table for the violating cable length doesn't exist in app db
excepted_pg_table = 'BUFFER_PG_TABLE:{}:3-4'.format(port_to_test)
pg_table_in_app_db = check_pg_profile(
duthost, excepted_pg_table, expected_profile, fail_test=False)
assert not pg_table_in_app_db, f"{expected_profile} should not exist in {excepted_pg_table} in app db"
# Check syslog includes relevant error log
check_log_analyzer(loganalyzer, marker)
finally:
logging.info('[Clean up]')
Expand Down