Enhance Functional Testing: Daemon Startup Validation, Log Improvements, and Test Restart Refactoring

[moisesPompilio]

Description and Notes

A validation step was added after node startup to ensure the daemon actually started successfully. After executing the startup command, the workflow waits for 1 second to verify if the node is running. If the daemon fails to start(such as due to misconfigured flags or command errors) the process is stopped, a relevant error message is logged (including the output of the command and the affected node), and an error is returned. This provides much clearer feedback for developers writing or running tests and makes it easier to detect incorrect flags or startup commands.

The port detection logic was also improved. Previously, the test only read the last line of the log within a specified timeout, which could cause failures if the output timing was off, resulting in missed port detection and unstable CI runs. Now, all log lines are read and checked, significantly reducing the chance of missing the port assignment. Since port information is always printed at the beginning of the logs, this change does not impact performance. Note: This is not a definitive solution to the architectural issue behind port detection; a more comprehensive fix is planned for PR #733.

Additionally, the restart test was refactored for better clarity and relevance. Instead of starting two nodes, stopping them, and checking their data directories (which did not accurately reflect a restart scenario), the test now starts a node, stops it, and then starts it again, verifying that restart works correctly and that stopping the node does not corrupt files.

How to verify the changes you have done?

• Start a node with both valid and intentionally invalid flags or commands. Confirm that, if the daemon fails to start, a clear log error is shown with the command output and node information.
• Run the functional test suite and observe improved reliability in port detection; verify that logs are parsed correctly and the port is assigned even with slight output timing differences.
• Review the restart test logic: ensure it starts a node, stops it, and starts the same node again, checking for correct restart behavior and file integrity.

[Davidson-Souza]

Could you explain a little on each commit what's the logic behind those changes? I like the PR description, but remember that people may read each commit independently.

[moisesPompilio]

Could you explain a little on each commit what's the logic behind those changes? I like the PR description, but remember that people may read each commit independently.

I improved the commit descriptions.

[jaoleal]

I went commit by commit and i got the feeling that they are backwards ?

I feel that the one changing the daemon startup should be the first one and the one simplifying the restart.py by last

[Davidson-Souza]

ACK e87a4d2

It seems this attempts to fix the same problem as #731. Do you guys have any preference about the order for merging? IIUC #731's approach is the final one?

[moisesPompilio]

It seems this attempts to fix the same problem as #731. Do you guys have any preference about the order for merging? IIUC #731's approach is the final one?

I think this PR can be merged first without any problem. #731 tends to be the final approach because it will stop reading things from the log, and #731 already excludes the detect_ports part, so it would remove this small fix from this PR without any issues I hope. This PR just makes a small adjustment to how ports are read, because in the current code it doesn't tolerate any timeout during node initialization. Since this PR inserts a small 1 second timeout during node initialization to check if any error occurred, this timeout is already enough to make most tests break because the port line would have already passed.

[jaoleal]

ACK e87a4d2

[joaozinhom]

ACK e87a4d2