[upgrade_image] Fix python interpreter mismatch after upgrade

[lolyu]

Description of PR

Summary:
Fixes # (issue)

Type of change

• Bug fix
• Testbed and Framework(new/improvement)
• New Test case
  • Skipped for non-supported platforms
• Test case improvement

Back port request

• 202205
• 202305
• 202311
• 202405
• 202411
• 202505
• 202511

Approach

What is the motivation for this PR?

202511 nightly upgrade has the following image upgrade issue:

2026-01-19 04:47:52,053 ansible_hosts.py#502 DEBUG - /var/src/sonic-mgmt_testbed-bjw3-can-dual-t0-7260-1/ansible/devutil/devices/sonic.py::post_upgrade_actions#237: "localhost" -> AnsibleModule::pause | Results =>{"hostname": "localhost", "reachable": true, "failed": false, "changed": false, "rc": 0, "stderr": "", "stdout": "Paused for 60.0 seconds", "start": "2026-01-19 04:46:52.047977", "stop": "2026-01-19 04:47:52.049217", "delta": 60, "echo": true, "user_input": "", "_ansible_no_log": false}
2026-01-19 04:47:52,054 ansible_hosts.py#426 DEBUG - ===== ['bjw3-can-7260-13', 'bjw3-can-7260-14'] -> shell ================================================================
2026-01-19 04:47:52,054 ansible_hosts.py#444 DEBUG - /var/src/sonic-mgmt_testbed-bjw3-can-dual-t0-7260-1/ansible/devutil/devices/sonic.py::post_upgrade_actions#242: ["bjw3-can-7260-13", "bjw3-can-7260-14"] -> AnsibleModule::shell, {"module_name": "shell", "args": ["sed -i \"s/^ClientAliveInterval [0-9].*/ClientAliveInterval 900/g\" /etc/ssh/sshd_config && systemctl restart sshd"], "kwargs": {}, "module_attrs": {"become": true}}
2026-01-19 04:47:53,211 ansible_hosts.py#502 DEBUG - /var/src/sonic-mgmt_testbed-bjw3-can-dual-t0-7260-1/ansible/devutil/devices/sonic.py::post_upgrade_actions#242: ["bjw3-can-7260-13", "bjw3-can-7260-14"] -> AnsibleModule::shell | Results =>{"bjw3-can-7260-13": {"hostname": "bjw3-can-7260-13", "reachable": true, "failed": true, "module_stdout": "", "module_stderr": "Warning: Permanently added '10.150.238.70' (RSA) to the list of known hosts.\r\nDebian GNU/Linux 13 \\n \\l\n\n/bin/sh: 1: /usr/bin/python3.11: not found\n", "msg": "The module failed to execute correctly, you probably need to set the interpreter.\nSee stdout/stderr for the exact error", "rc": 127, "_ansible_no_log": false, "changed": false}, "bjw3-can-7260-14": {"hostname": "bjw3-can-7260-14", "reachable": true, "failed": true, "module_stdout": "", "module_stderr": "Warning: Permanently added '10.150.238.72' (RSA) to the list of known hosts.\r\nDebian GNU/Linux 13 \\n \\l\n\n/bin/sh: 1: /usr/bin/python3.11: not found\n", "msg": "The module failed to execute correctly, you probably need to set the interpreter.\nSee stdout/stderr for the exact error", "rc": 127, "_ansible_no_log": false, "changed": false}}
2026-01-19 04:47:53,212 sonic.py#265 ERROR - Post upgrade actions failed, devices: ['bjw3-can-7260-13', 'bjw3-can-7260-14'], error: RunAnsibleModuleFailed('/var/src/sonic-mgmt_testbed-bjw3-can-dual-t0-7260-1/ansible/devutil/devices/sonic.py::post_upgrade_actions#242: ["bjw3-can-7260-13", "bjw3-can-7260-14"] -> AnsibleModule::"shell" failed, Results => {"bjw3-can-7260-13": {"hostname": "bjw3-can-7260-13", "reachable": true, "failed": true, "module_stdout": "", "module_stderr": "Warning: Permanently added \'10.150.238.70\' (RSA) to the list of known hosts.\\r\\nDebian GNU/Linux 13 \\\\n \\\\l\\n\\n/bin/sh: 1: /usr/bin/python3.11: not found\\n", "msg": "The module failed to execute correctly, you probably need to set the interpreter.\\nSee stdout/stderr for the exact error", "rc": 127, "_ansible_no_log": false, "changed": false}, "bjw3-can-7260-14": {"hostname": "bjw3-can-7260-14", "reachable": true, "failed": true, "module_stdout": "", "module_stderr": "Warning: Permanently added \'10.150.238.72\' (RSA) to the list of known hosts.\\r\\nDebian GNU/Linux 13 \\\\n \\\\l\\n\\n/bin/sh: 1: /usr/bin/python3.11: not found\\n", "msg": "The module failed to execute correctly, you probably need to set the interpreter.\\nSee stdout/stderr for the exact error", "rc": 127, "_ansible_no_log": false, "changed": false}}')

The issue is due to the PREV image and the upgrade-to image have different python interpreter versions. When Ansible runs first time on a target device, it will cache the python interpreter path in the memory; the error arises when the device boots up with the upgrade-to image and Ansible fails to find the python interpreter using the path that is from the PREV image.

This is observed on nightly that tries to upgrade to 20251110.03:

admin@bjw3-can-7260-13:~$ show version | head -n 5

SONiC Software Version: SONiC.20251110.02
SONiC OS Version: 12
Distribution: Debian 12.12
Kernel: 6.1.0-29-2-amd64
admin@bjw3-can-7260-13:~$ python --version
Python 3.11.2

admin@bjw3-can-7260-13:~$ show version | head -n 5

SONiC Software Version: SONiC.20251110.03
SONiC OS Version: 13
Distribution: Debian 13.2
Kernel: 6.12.41+deb13-sonic-amd64
admin@bjw3-can-7260-13:~$ python --version
Python 3.13.5

Signed-off-by: Longxiang lolv@microsoft.com

How did you do it?

Let's reset the facts cache in the postupgrade before running any Ansible modules on the new image.

How did you verify/test it?

2026-01-19 06:29:24,312 upgrade_image.py#187 INFO - SONiC host bjw3-can-7260-13 current version 20251110.03
2026-01-19 06:29:24,312 upgrade_image.py#187 INFO - SONiC host bjw3-can-7260-14 current version 20251110.03
2026-01-19 06:29:24,312 upgrade_image.py#202 INFO - Skip enabling FIPS
2026-01-19 06:29:24,313 upgrade_image.py#220 INFO - Use default docker folder size
2026-01-19 06:29:24,313 upgrade_image.py#232 INFO - ===== UPGRADE IMAGE DONE =====

Any platform specific information?

Supported testbed topology if it's a new test case?

Documentation

[mssonicbld]

/azp run

[azure-pipelines]

Azure Pipelines successfully started running 1 pipeline(s).

[mssonicbld]

Cherry-pick PR to 202511: #21969