Skip to content
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.

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

[recover] improve adaptive recover methods #1652

Merged
merged 1 commit into from
May 12, 2020
Merged

Conversation

yxieca
Copy link
Collaborator

@yxieca yxieca commented May 7, 2020

Summary:
Fixes # (issue)

Type of change

  • Bug fix
  • Testbed and Framework(new/improvement)
  • Test case(new/improvement)

Approach

How did you do it?

  • When process is missing, reload_config.
  • When port channle or vlan link is down, reload_config.
  • redirect output of config reload and load minigraph to /dev/null.

Signed-off-by: Ying Xie ying.xie@microsoft.com

How did you verify/test it?

Run all pytest cases, currently some test case will leave the system in a bad state that would require config reload to recover. Making sure that the test would continue.

This is a step towards better recovery. However, specific scenarios like config_db.json got contaminated is not addressed by this change.

- When process is missing, reload_config.
- When port channle or vlan link is down, reload_config.
- redirect output of config reload and load minigraph to /dev/null.

Signed-off-by: Ying Xie <ying.xie@microsoft.com>
@daall
Copy link
Contributor

daall commented May 7, 2020

However, specific scenarios like config_db.json got contaminated is not addressed by this change.

What do you mean by this? Just that we can't detect config_db.json being contaminated?

@yxieca
Copy link
Collaborator Author

yxieca commented May 7, 2020

However, specific scenarios like config_db.json got contaminated is not addressed by this change.

What do you mean by this? Just that we can't detect config_db.json being contaminated?

There is no good way of doing so. The only signal I see is having /etc/sonic/config_db.json.bak but I cannot be sure that config_db.json.bak is good. Maybe instead of using config reload, we should always go back to load minigraph. But then how do we know that minigraph is good? I feel that these are bigger problems that we need to think through before acting. Also some community members doesn't have minigraph to begin with. config reload is the common ground for all.

@yxieca yxieca merged commit 6e6c81d into sonic-net:master May 12, 2020
@yxieca yxieca deleted the recover branch May 12, 2020 20:11
noaOrMlnx added a commit to noaOrMlnx/sonic-mgmt that referenced this pull request May 19, 2020
* Convert dip_sip ansible test to pytest (sonic-net#1488)

* Convert dip_sip ansible test to pytest

* Fix Review Comments:
- Add ansible wrapper to dip_sip.yml
- Delete unneccessary files

Signed-off-by: Noa Or <noaor@mellanox.com>

* Fix review comments

Check if topology is lag group by minigraph facts.

* Update test_dip_sip.py

* [pmon daemon check] refactoring pmon daemon state check code (sonic-net#1537)

* [daemon utils] move platform/check_daemon_status.py to common/platform/daemon_utils.py

Signed-off-by: Ying Xie <ying.xie@microsoft.com>

* [pmon daemon] refactoring pmon daemon state checking

- Refactor get_pmon_daemon_list to get_pmon_daemon_states:
  returning list of daemon states.
  not returning daemons that are known short lived.
  not returning daemons disabled by configuration.
- Refactor daemon state check to check the returned states.

Signed-off-by: Ying Xie <ying.xie@microsoft.com>

* Fix test_techsupport.py so it could run with T1 topology (sonic-net#1538)

* Fix test_techsupport.py so it could run with T1 topology

* [lgtm] Configure LGTM to analyze test code (sonic-net#1544)

Signed-off-by: Danny Allen <daall@microsoft.com>

* [tests] Fix LGTM errors in tests folder (sonic-net#1545)

Signed-off-by: Danny Allen <daall@microsoft.com>

* Convert vxlan-decap testing to pytest (sonic-net#1518)

Signed-off-by: Xin Wang <xinw@mellanox.com>

Co-authored-by: Xin Wang <xinw@mellanox.com>

* Add the APC PSU support for SNMP PSU controller (sonic-net#1461)

Add APC SNMP MIB for PDU control

PORT_NAME_BASE_OID = ".1.3.6.1.4.1.318.1.1.4.4.2.1.4"
PORT_STATUS_BASE_OID = ".1.3.6.1.4.1.318.1.1.12.3.5.1.1.4"
PORT_CONTROL_BASE_OID = ".1.3.6.1.4.1.318.1.1.12.3.3.1.1.4"

How did you verify/test it?
{username}@{sonic-mgmt-docker}:/var/sonic-mgmt-int/tests/common/plugins/psu_controller$ python
Python 2.7.12 (default, Oct 8 2019, 14:14:10)
[GCC 5.4.0 20160609] on linux2
Type "help", "copyright", "credits" or "license" for more information.

from snmp_psu_controllers import *
psucntl = get_psu_controller('<controller_ip>',"<device_name>")
psucntl.get_psu_status()
[{'psu_id': 0, 'psu_on': True}]
psucntl.turn_off_psu('0')
True
psucntl.get_psu_status()
[{'psu_id': 0, 'psu_on': False}]
psucntl.turn_on_psu('0')
True
psucntl.get_psu_status()
[{'psu_id': 0, 'psu_on': True}]
print psucntl.psuType
APC

* [Mellanox] Fix issues for thermal control test cases (sonic-net#1524)

* [thermal fix] 1. should not mock PSU fan speed less than 100; 2. wait 5 seconds after turning on PSU
* Add some check for test case test_thermal_control_psu_absence

* [pytest/hash] add hash key ingress-port test (sonic-net#1509)

1. Iterate once for every port when hash key is ingress port.
2. In loose mode, do not need to enter the check_balancing function.

* [advanced reboot] Add Paramiko module for device connection (sonic-net#1542)

* [advanced reboot] Add Paramiko module for device connection

Parmiko module provides fallback mechanism to using username/password
This is required if we are rebooting into new image using advanced
reboot test fixture.

signed-off-by: Tamer Ahmed <tamer.ahmed@microsoft.com>

* [pytest] Port advanced reboot based test cases (sad path) (sonic-net#1543)

* [pytest] Port advanced reboot based test cases (sad path)

This patch ports remaining sad path advanced reboot test cases
to pytest infra.

signed-off-by: Tamer Ahmed <tamer.ahmed@microsoft.com>

* Update the importing of conn_graph_facts after platform_fixtures.py is removed (sonic-net#1547)

PR sonic-net#1492 copied the implementation of conn_graph_facts in tests/platform/platform_fixtures.py
to tests/common/fixtures/conn_graph_facts.py.

Then PR sonic-net#1503 removed file tests/platform/platform_fixtures.py. But not all related importings
were updated.

This change is to update all the platform_fixtures related importings.

Signed-off-by: Xin Wang <xinw@mellanox.com>

Co-authored-by: Xin Wang <xinw@mellanox.com>

* Create README.testbed.cEOS.md (sonic-net#1549)

* Configure and enable core uploader (sonic-net#1522)

* If core-storage secret key is available, add to /etc/sonic/core_analyzer.rc.json and enable & start core_uploader service
If https_proxy is provided, update /etc/sonic/core_analyzer.rc.json.

* Check the entire dict path before de-referencing.

* Improved regex per comments.

* Fixed syntax error.

* Add a sample file for newly introduced ansible facts.wq

* Removed a redundant empty line.

Co-authored-by: Ubuntu <remanava@remanava-kube-1.hblknyhzkmnujibhxvn3dmavjb.xx.internal.cloudapp.net>

* [tests/common/devices] Add FanoutHost (sonic-net#1532)

* [tests/common/devices] Add FanoutHost

- Add FanoutHost which aggregate different fanout switch host
- Add fanouthosts fixture shortcut to get fanouthost easily
- Support select host instance by os type
- Support get fanout host by both mgmtip and hostname

Co-authored-by: Jing Kan<jika@microsoft.com>

* [tacacs]: replace shell command with service module (sonic-net#1550)

Signed-off-by: Guohan Lu <gulv@microsoft.com>

* [tests/lag]: Convert lag_fallback.yml to pytest (sonic-net#1526)

* [tests/lag]: Convert lag_fallback.yml to pytest

- Convert lag_fallback.yml to pytest
- Reuse int shutdown mehod to shut a member of lag
- Use EOSHost class method shutdown/no_shutdown
- Add post check for EOSHost shutdown/no_shutdown
- Fix typos

Co-authored-by: Jing Kan<jika@microsoft.com>

* [pytest] Fix get_asic_type method and logging (sonic-net#1557)

* [pytest] Fix get_asic_type method and logging

The get_asic_type method dereferences dut which is not a member of the the class.
the dut is an instance of SonicHost class and so should reference self.

signed-off-by: Tamer Ahmed <tamer.ahmed@microsoft.com>

* Added iptables rules to make sure BGP don't ack back to BGP peer syn messages (sonic-net#1561)

* Added iptables rules to make sure BGP don't
ack to peer messages. Otherwise this TX packets
can make egress buffer accounting  off

Signed-off-by: Abhishek Dosi <abdosi@microsoft.com>

* Review Comments Address

* [pytest/hash] Generate available send packet ports instead of hard coding src_ports in hash_test.py. (sonic-net#1511)

Generate available send packet ports instead of hard coding src_ports in hash_test.py.

* [fanout switch] build fanout switch list from device_conn map (sonic-net#1566)

* [fanout switch] build fanout switch list from device_conn map

- Build fanout switch from connection map.
- Build fanout switch DUT port map for tests.
- Build credential dict according to the DUt's inventory.
- Set fanout switch password with right variable name.
- Return command outpus.

Signed-off-by: Ying Xie <ying.xie@microsoft.com>

* Fix an indentation

* add comment for the os type

* Stop arp_update before crm test to prevent DUT learning mac address unexpectly, which caused the fdb test fail. (sonic-net#1559)

* Stop arp_update before crm test to prevent DUT learning mac address unexpectly, which caused the fdb test fail.

* [pytest] Add replace fast-reboot script test option (sonic-net#1563)

Adding an option to replace fast reboot script on dut as part of
testbed preparation.

signed-off-by: Tamer Ahmed <tamer.ahmed@microsoft.com>

* [pytest/creds]: load groups vars into creds for the dut (sonic-net#1575)

Signed-off-by: Guohan Lu <gulv@microsoft.com>

* [pytest/common/devices] Remove immediately checking of shut/no_shut (sonic-net#1576)

- The check can be put in a wait_unitl if user want

Co-authored-by: Blueve <jika@microsoft.com>

* SPyTest documentation initial version (sonic-net#1565)

Co-authored-by: Rama Sasthri, Kristipati <rama.kristipati@broadcom.com>

* [tests/conftest] Read os_type from ansible inventory vars to construct FanoutHost (sonic-net#1577)

- Read os var from ansible inventory manager

Co-authored-by: Jing Kan<jika@microsoft.com>

* [link flap] add link flap pytest (sonic-net#1573)

[link flap] add link flap pytest

* Use show interface status to obtain operational status of interfaces.
* Filter out operational down interfaces from test list.
* Address show_interface.py issue with single interface output.

Signed-off-by: Ying Xie <ying.xie@microsoft.com>

* Fix sensor data for 3700 (sonic-net#1568)

* [Mellanox] Fix issue: remove non-exist sensor sysfs files (sonic-net#1567)

* [pytest/snmp_lldp]: exclude mgmt interface from active interface list (sonic-net#1564)

align the criteria for checking lldpRemManAddrTable, mgmt interface 'eth0' is ruled out from the lldp neighbor list.

* [pytest/features] Add test for show features command (sonic-net#1546)

* [pytest assersion] introduce assertion handling code (sonic-net#1592)

Signed-off-by: Ying Xie <ying.xie@microsoft.com>

* [advanced-reboot] Refactor prepare ssh keys into ssh_utils (sonic-net#1571)

The ssh keys preparation is also required by wr_arp test suite,
and so putting it into common utils file for sharing

signed-off-by: Tamer Ahmed <tamer.ahmed@microsoft.com>

* [tests/pc] Use test_po_update.py in config.yml directly (sonic-net#1589)

* [tests/pc] Use test_po_update.py in config.yml directly

- Replace origin logic of config.yml with test_po_update.py
- Refine the test_po_udpate test with recommand fixture usage
- Remove useless import
- Remove useless var

Co-authored-by: Blueve <jika@microsoft.com>

* [pytest] Convert control plane assisted warm-reboot test (wr_arp) (sonic-net#1572)

* [pytest] Convert control plane assisted warm-reboot test (wr_arp)

This PR converts Ansible control assisted warm-reboot (wr_arp) test
case to pytest.

signed-off-by: Tamer Ahmed <tamer.ahmed@microsoft.com>

* [recovery] introduce adaptive recover method (sonic-net#1583)

* [recovery] introduce adaptive recover method

- For interfaces down failure, try to bring up the interfaces.
- For service down failure, reload config to bring them back up.
- For all other failures, reboot to recover.
- Make adaptive the default recover method.
- reboot dut to recover if database service is down
- Only allow cover service to override None action

Signed-off-by: Ying Xie <ying.xie@microsoft.com>

* [tests/lag_2] Refactor test_lag_2.py (sonic-net#1582)

* [tests/lag_2] Refactor test_lag_2.py

- Refine the code structure by using class to share a static context
- Remove most useless code and defines
- Rename some vars
- Use fanouthosts fixture for single_lag test
- Fix existing LGMT alarm
- Use wait_until to check if a int has been shut
- Fix pytest warnings regarding Test case detector

Co-authored-by: Jing Kan <jika@microsoft.com>

* [ansible] Fix symlink to ferret.conf.j2 (sonic-net#1595)

Template file was moved to arp/files/ferret.conf.j2 and symlink update was missed

signed-off-by: Tamer Ahmed <tamer.ahmed@microsoft.com>

* [test/procdockerstatsd] Verify daemon is active/running and uploads data successfully (sonic-net#1548)

* [pfc_storm_icos.j2]: jinja2 template to start PFC storm on ICOS. (sonic-net#1590)

[pfc_storm_stop_icos.j2]: jinja2 template to stop PFC storm on ICOS.

Note: Jinja2 considers physical to logical mapping of interfaces as per ICOS user manual.
fpti1_0_1 0/1
fpti1_0_2 0/2
fpti1_0_3 0/3

Co-authored-by: Praveen Chaudhary <pchaudha@pchaudha-mn1.linkedin.biz>

* [pytest/hash] Add hash keys: 'src-mac', 'dst-mac', 'ip-proto', 'vlan-id'. (sonic-net#1512)

* Revise according to the following review comments:
1. Not enable src-mac, dst-mac and vlan-id hash keys by default.
2. Get all untag vlan ports move to a separate function.
3. setup vlans for hash_key vlan-id test when vlan-id in hash_keys.

* [pytest/default_route_check]: check various aspect related to default route learnt from bgp (sonic-net#1598)

- check if default route has correct set src address
- check if ipv6 default route use global nexthop or not

Signed-off-by: Guohan Lu <gulv@microsoft.com>

* Fix below issue: (sonic-net#1596)

add topo is giving error as part of VEOS task
TASK [eos : Get VM front panel interface number]
with error FAILED! => {"msg": "Incorrect sudo password"}

Instead of using ansible_sudo_password make it ansible_become_password
based on below thread.
ansible/ansible#62042

* [tests/lag_2] Use test_lag_2.py in lag_2.yml directly (sonic-net#1602)

- Remove lag_fallback.yml
- Remove single_lag_lacp_rate_test.yml
- Remove single_lag_test.yml
- Use pytest runner to run pytest lag_2 test in lag_2.yml

Co-authored-by: Jing Kan<jika@microsoft.com>

* [tests/sensors] Convert sensors_check.yml to pytest (sonic-net#1601)

* [tests/sensors] Convert sensors_check.yml to pytest

- Convert sensors_check.yml to test_sensors.py
- Skip test if platform not supported
- Replace origin playbook with pytest runner
- Use pytest_assert helper instead of native assert to reduce noise
- Refactor _platform_info method in SonicHost and expose it function as a getter

Co-authored-by: JIng Kan<jika@microsoft.com>

* [tests/telemetry] Verify default config parameters (sonic-net#1530)

* [pytest/test_default_route]: support 4.9 kernel 201911/201811 release (sonic-net#1600)


Signed-off-by: Guohan Lu <gulv@microsoft.com>

* [TestbedProcessing] Minor fixes to suite latest changes (sonic-net#1604)

* Remove witespases

* Add ptf_pb_ip parsing

* Fix testbed.csv generating

* Use ansible_become_pass

* [pytest]: reorg tests info subfolder (sonic-net#1613)

Signed-off-by: Guohan Lu <gulv@microsoft.com>

* [sanity_checks]: add critical process check in sanity checks (sonic-net#1617)

- first read the critical process list (/etc/supervisord/critical_process)
from the container, and then check if any of the process crashed.

- add snmp container to the critical service list

- add auto_recover support

Signed-off-by: Guohan Lu <gulv@microsoft.com>

* add comment

Signed-off-by: Guohan Lu <gulv@microsoft.com>

* add auto_recover support if process check failed

* [tests/sensors] Move test_sensors.py to domain specific folder (sonic-net#1616)

Co-authored-by: Jing Kan <jika@microsoft.com>

* Fix sensors_check link (sonic-net#1611)

* [pytest]: add get_ip_route_info in SonicHost (sonic-net#1618)

get_ip_route_info returns ip route info in the kernel
for a given destination ip

refactor test_default_route to use the new method

Signed-off-by: Guohan Lu <gulv@microsoft.com>

* [Mellanox/platform]Compare CPU's temperature against critical threshold (sonic-net#1585)

* [loganalyzer] Fail flag support in context manager and callback execution (sonic-net#1619)

* Fail flag support in context manager and callback execution

Signed-off-by: Neetha John <nejo@microsoft.com>

* [Mellanox] add test cases for dynamic minimum fan speed and psu fan speed policy (sonic-net#1552)

* [pytest/nbrhost]: add neighbor config info in nbrhost class (sonic-net#1621)

* [vxlan-decap]: Flush ptf rx buffer before ptf send packet. (sonic-net#1608)

* add links to virtual switch testbed setup

* [telemetry certs] deploy certs for telemetry in deploy-mg (sonic-net#1614)

* adding server and dsmsroot certs for telemetry
* adding support for ptfhost copy certs

* [pytest] Add support to populate DUT FDB entries (sonic-net#1593)

* [pytest] Add support to populate DUT FDB entries

This new PTF plugin and test case that populates DUT FDb entries.
It generates n Packets and sends them to DUT vlan IP. The number
of distinct MAC to distinct ip addresses is configurable.

signed-off-by: Tamer Ahmed <tamer.ahmed@microsoft.com>

* Revert "[telemetry certs] deploy certs for telemetry in deploy-mg (sonic-net#1614)" (sonic-net#1628)

This reverts commit 19e92b1.

* [Mellanox]Fix sensor data for 3700/3700c/3800 (sonic-net#1627)

* Ignore test in test_turn_on_off_psu_and_check_psustatus (sonic-net#1606)

* Fixed fixture scope conflict for DUT monitor (sonic-net#1630)

Signed-off-by: Yuriy Volynets <yuriyv@mellanox.com>

* [test plans] Moving test plans from sonic-wiki repo to sonic-mgmt repo (sonic-net#1631)

Signed-off-by: Ying Xie <ying.xie@microsoft.com>

* [pytest/bgp_gr_test]: add bgp graceful restart helper test (sonic-net#1623)

Signed-off-by: Guohan Lu <gulv@microsoft.com>

* [Mellanox] Adjust test cases for fan led support (sonic-net#1580)

* Fixes in test_drop_counters.py (sonic-net#1610)

- Added more attempts to read and verify if expected drop counter was changed to the specific - value. It was observed delays in counter refreshing, so this fix was added.
- Fixed 'test_not_expected_vlan_tag_drop', removed vlan 1, because port channels are members of vlan 1.
- Skipped 'test_loopback_filter', added description.

Signed-off-by: Yuriy Volynets <yuriyv@mellanox.com>

* [swap_syncd] Fix failures in swap syncd (sonic-net#1632)

Signed-off-by: Danny Allen <daall@microsoft.com>

* [pytest]: fix typo in parse default route output (sonic-net#1634)

* Update README.md

* Update README.md

* [pytest/ansible] Add support for multi-duts (sonic-net#1432)

* [pytest/ansible] Add support for multi-duts

* review comment: use duts instead of dut

* Fixing various test cases to use duthost instead of testbed_devices['dut']

* trim spaces if any when present in the duts list

signed-off-by: Tamer Ahmed <tamer.ahmed@microsoft.com>

* Pytest organization proposal (sonic-net#1605)

* Pytest organization proposal

Signed-off-by: Neetha John <nejo@microsoft.com>

* [test/dir_bcast] Convert dir_bcast.yml to pytest (sonic-net#1609)

* [test/dir_bcast] Convert dir_bcast.yml to pytest
* Force refresh router before testing and move to ipfwd folder
* Fix LGTM alerts

Co-authored-by: Jing Kan<jika@microsoft.com>

* [copp] Convert COPP test to pytest (sonic-net#1633)

Signed-off-by: Danny Allen <daall@microsoft.com>

* [doc]: Update document to specify non-empty password file (sonic-net#1638)

ansible 2.8 expects a non-empty file

* Revert "[copp] Convert COPP test to pytest (sonic-net#1633)" (sonic-net#1643)

This reverts commit 9a8fc1e.

* [pytest] Fix ptf_runner issue introduced by COPP test conversion (sonic-net#1644)

* [copp] Convert COPP test to pytest (sonic-net#1633)
* Fix missing space in ptf_runner

Signed-off-by: Danny Allen <daall@microsoft.com>

* [fast-reboot]: Fix IP range overlapping. (sonic-net#1637)

Signed-off-by: Nazarii Hnydyn <nazariig@mellanox.com>

* Add support of specifying multiple inventory files (sonic-net#1645)

The pytest-ansible plugin always assumes that the value
of cli argument '--inventory' is a single inventory file.
With this enhancement, we can pass in multiple inventory
files using the cli argument '--inventory'. The multiple
inventory files can be separated by comma ','.

For example:
        pytest --inventory "inventory1, inventory2" <other arguments>
        pytest --inventory inventory1,inventory2 <other arguments>

Signed-off-by: Xin Wang <xiwang5@microsoft.com>

* [pytest] Fix module import issue when running whole test suite (sonic-net#1642)

* [pytest] Fix module import issue when running whole test suite

When having multiple conftest, dir leading to this conftest has
to be Python package (presence of __init__.py.) Also, adding
pytest basedir to conftest.py

signed-off-by: Tamer Ahmed <tamer.ahmed@microsoft.com>

* [advanced-reboot] Fix IP range overalapping (sonic-net#1653)

This code ports PR 1637 to advanced reboot.

ported-pr: sonic-net#1637
signed-off-by: Tamer Ahmed <tamer.ahmed@microsoft.com>

* Deprecate ansible_host.py (sonic-net#1658)

* Deprecate ansible_host.py

The functionalities in ansible_host.py , including the
AnsibleHost class and some exceptions, have been
implemented in the common library. This file can
be deprecated in case people new to sonic-mgmt
waste time on old libraries.

Changes:
1. Remove ansible_host.py
2. Replace all the AnsibleHost related calls with
    more appropriate fixtures or functions.
3. Remove duplicated localhost fixture definition

Signed-off-by: Xin Wang <xiwang5@microsoft.com>

* [Mellanox] Fix sensor data for SN4700 (sonic-net#1660)

* Fix for SN4700 sensors

Signed-off-by: Shlomi Bitton <shlomibi@mellanox.com>

* Edit sensors file with correct labels

* [drop counters] Improve support for combined L2/L3 drop counters (sonic-net#1649)

- Get the correct interface for combined L2/L3 counters
- Add Arista and Dell SKUs to the combined counter list

Signed-off-by: Danny Allen <daall@microsoft.com>

* [pytest/ntp] Use local time when behind proxy (sonic-net#1640)

When testbed is behind a proxy then NTP inside the ptf container cant synchronize with public NTP servers and test hung on ntpd -gq command

* [recover] improve adaptive recover methods (sonic-net#1652)

- When process is missing, reload_config.
- When port channle or vlan link is down, reload_config.
- redirect output of config reload and load minigraph to /dev/null.

Signed-off-by: Ying Xie <ying.xie@microsoft.com>

* [sensors] Added support of Montara nad Maverics platforms to sku-sensors-data.yml (sonic-net#1646)

Signed-off-by: Vitaliy Senchyshyn <vsenchyshyn@barefootnetworks.com>

* [pytest] PFCWD config test (sonic-net#1620)

* Pytest PFCWD config test

Signed-off-by: Neetha John <nejo@microsoft.com>

* Address LGTM

Signed-off-by: Neetha John <nejo@microsoft.com>

* Rename marker to syslog_marker

Signed-off-by: Neetha John <nejo@microsoft.com>

* Separate out verbose comments

Signed-off-by: Neetha John <nejo@microsoft.com>

* [advanced-reboot] Fix testing hang when doing BGP shutdown on Arista VM (sonic-net#1579)

Fix warm-reboot-sad-bgp testing hang after vEOS upgraded to 4.20.15M.
It pops a confirm message when doing BGP shutdown.

* [file organization] Rename folder names to allow markers to work again (sonic-net#1671)

* Revert "[pytest] Fix module import issue when running whole test suite (sonic-net#1642)"

This reverts commit f88cff2.

* [file organization] Rename folder names to allow markers to work again

pytest has an issue dealing with sub-folders with the same name and being
included from. In this case, the problem was with several components has
'args' as sub-folder to hold argument parsing utilities.

This issue was initially caught when we were tryign to run the whole test
suite. Attemp was made to address this issue. However there are more in
pytest that cannot tolerate same name sub-folders.

Change 'args' to '<feature>_args' and change conftest.py accordingly.

Signed-off-by: Ying Xie <ying.xie@microsoft.com>

* Rename 'platform' directory to 'platform_tests' to prevent conflicts (sonic-net#1662)

* Deprecate fixture testbed_devices (sonic-net#1665)

The testbed_devices fixture is not a good design. Multiple testbed device
objects are initialized in this fixture. It is not common that a test
script needs all the devices. This fixture may cause unnecessary overhead.
It would be better for test scripts to use the fixtures for different
devices on needed basis.

Changes:
1. Remove the definition of fixture testbed_devices.
2. Replace the call to testbed_devices fixture with other fixtures that
   fit better.

Signed-off-by: Xin Wang <xiwang5@microsoft.com>

* [Mellanox] Fix sensor data for 2010 (sonic-net#1673)

* Update Th buffer params for Alpha change (sonic-net#1676)

Signed-off-by: Neetha John <nejo@microsoft.com>

* [QoS] Fix qos issues (sonic-net#1664)

Fix some issues found during QoS test.

- Support ARP populate for ptf topo
- Add debug info in WRRtest for the purpose of providing detail result in case of failure.

Co-authored-by: Ying Xie <yxieca@users.noreply.github.com>
Co-authored-by: Danny Allen <daall@microsoft.com>
Co-authored-by: Xin Wang <wangxin.wang@gmail.com>
Co-authored-by: Xin Wang <xinw@mellanox.com>
Co-authored-by: Sujin Kang <sujkang@microsoft.com>
Co-authored-by: Junchao-Mellanox <57339448+Junchao-Mellanox@users.noreply.github.com>
Co-authored-by: William-zx <47626856+William-zx@users.noreply.github.com>
Co-authored-by: Tamer Ahmed <tamer.ahmed@microsoft.com>
Co-authored-by: lguohan <lguohan@gmail.com>
Co-authored-by: Renuka Manavalan <47282725+renukamanavalan@users.noreply.github.com>
Co-authored-by: Ubuntu <remanava@remanava-kube-1.hblknyhzkmnujibhxvn3dmavjb.xx.internal.cloudapp.net>
Co-authored-by: Blueve <672454911@qq.com>
Co-authored-by: abdosi <58047199+abdosi@users.noreply.github.com>
Co-authored-by: Iris Hsu <iris_hsu@edge-core.com>
Co-authored-by: Blueve <jika@microsoft.com>
Co-authored-by: Rama Sasthri, Kristipati <rama.kristipati@gmail.com>
Co-authored-by: Rama Sasthri, Kristipati <rama.kristipati@broadcom.com>
Co-authored-by: Kebo Liu <kebol@mellanox.com>
Co-authored-by: pra-moh <49077256+pra-moh@users.noreply.github.com>
Co-authored-by: Praveen Chaudhary <pchaudhary@linkedin.com>
Co-authored-by: Praveen Chaudhary <pchaudha@pchaudha-mn1.linkedin.biz>
Co-authored-by: Myron Sosyak <49795530+msosyak@users.noreply.github.com>
Co-authored-by: Stephen Sun <5379172+stephenxs@users.noreply.github.com>
Co-authored-by: Neetha John <nejo@microsoft.com>
Co-authored-by: yvolynets-mlnx <50697593+yvolynets-mlnx@users.noreply.github.com>
Co-authored-by: Xin Liu <xinxliu@microsoft.com>
Co-authored-by: Mahesh Maddikayala <10645050+smaheshm@users.noreply.github.com>
Co-authored-by: Nazarii Hnydyn <nazariig@mellanox.com>
Co-authored-by: Xin Wang <xiwang5@microsoft.com>
Co-authored-by: shlomibitton <60430976+shlomibitton@users.noreply.github.com>
Co-authored-by: Vitaliy Senchyshyn <43479243+vsenchyshyn@users.noreply.github.com>
Co-authored-by: Vincent Chiang <47546216+vincentchiang-ec@users.noreply.github.com>
Co-authored-by: Joe LeVeque <jleveque@users.noreply.github.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants