Fixes to netconf on SSH tunnels

[Sripadvallabh]

Summary

Fixes NETCONF per-connection logging when the NETCONF connection is established through an SSH tunnel.

Previously, logs emitted during SSH tunnel setup and ncclient session negotiation were not captured in the NETCONF per-connection log file. This change forwards the relevant tunnel and ncclient log records to the NETCONF file handler.

Also fixes a logging error seen when the pyATS tasklog stream is unavailable:

AttributeError: 'NoneType' object has no attribute 'write'

The NETCONF logger now skips attaching the pyATS tasklog adapter when the tasklog stream is not available.

Manual Test

Using a testbed with a NETCONF connection configured with sshtunnel, ran:

device.connect(via='netconf')

Verified that the generated NETCONF per-connection log file contains SSH tunnel setup logs, ncclient session logs, and the NETCONF connected message.

Example sanitized log content:

%NETCONF-INFO: Connecting proxy host <proxy>
%NETCONF-INFO: Adding local tunnel 127.0.0.1:<local-port> for <device-host>:830
%NETCONF-INFO: Device '<device>' connection 'netconf' via new SSH tunnel 127.0.0.1:<local-port>
%NETCONF-INFO: [host 127.0.0.1 ...] Sending:
%NETCONF-INFO: [host 127.0.0.1 ...] Received message from host
%NETCONF-INFO: [host 127.0.0.1 ...] initialized: session-id=<session-id>
%NETCONF-INFO: NETCONF CONNECTED

Automated Tests

python -m unittest discover src/yang/connector/tests -v
# Ran 26 tests: OK, skipped=3

python -m unittest discover . -v
# from ncdiff/src/yang/ncdiff/tests
# Ran 78 tests: OK

Total:

104 tests passed
3 tests skipped