Unable to run commands

[healy93]

Hi, I’m having trouble with Nutify sending commands to a UPS through a remote NUT server.

Setup:

• Nutify running in Docker compose (dartsteven/nutify:amd64-latest)
• Remote NUT server is reachable
• upsd.users allows a non-root account with actions = SET, actions = FSD, instcmds = ALL

From CLI everything works:

• upsc returns UPS status correctly
• upscmd -u <user> -p <pass> apc@192.168.xx.xx beeper.enable works (returns OK)

From Nutify WebUI:

• Any command (e.g. beeper.enable, test.battery.start.quick) fails with:

Output command: Username (root):
Errors: Error reading from stdin!

It seems Nutify tries to run upscmd as root inside the container instead of using the configured NUT user.
The socket connection is fine and metrics display correctly, only instant commands fail.

Any ideas?

Thanks, Healy

[DartSteven]

Closing this issue due to the Nutify v0.2.0 release.

Nutify 0.2.0 includes breaking changes and is not backward compatible with previous versions. A full clean setup is required: recreate the database from scratch and start from a new empty folder to avoid dirty or incompatible data.

If you still experience this problem on v0.2.0, please open a new ticket and provide:

Environment: Docker + image tag/version
Profile: Single/Multi
Topology:
Symptom:
Expected:
When started: <date/time>
Recent changes:
TopBar active target:
Logs (first error):
Tried already: <checks/fixes performed>

[healy93]

I've done some further testing on this and confirmed that the issue is still present in v0.3.0.

The problem appears to be in the UPS command handling. The Commands page was still using the primary ups_config.host / ups_config.name rather than resolving the currently selected Multi-NUT target.

I made a local patch to:

• core/upscmd/upscmd.py
• core/upscmd/api_upscmd.py

The patch:

• Passes the selected target_id from /api/upscmd/list to get_ups_commands().
• Resolves the selected Multi-NUT target before constructing UPS_NAME@UPS_HOST.
• Uses the same target resolution in execute_command().
• Retains the existing primary UPS configuration as the fallback.

After applying the patch and restarting the container, I tested this with both of my UPS devices.

Command discovery now follows the selected UPS, and I have successfully tested:

• Beeper enable/disable independently on both UPS devices.
• Battery-test commands independently on both UPS devices.
• Command discovery independently on both UPS devices.

Commands are now being sent to the UPS currently selected in the Multi-UPS interface, rather than always going to the primary UPS.

The relevant changes are essentially:

def get_ups_commands(target_id=None):
    ...
    UPS_NAME, UPS_HOST = _get_command_target(target_id)

and:

def execute_command(command, target_id=None):
    ...
    UPS_NAME, UPS_HOST = _get_command_target(target_id)

The API list endpoint now also resolves and passes the selected target:

scoped_target_id = _resolve_target_scope_id()
commands = get_ups_commands(target_id=scoped_target_id)

I also added a _get_command_target(target_id=None) helper which resolves the selected Multi-NUT target using the existing Multi-NUT storage/target handling, while falling back to the existing primary UPS configuration when no Multi-NUT target is selected.

After making these changes:

• py_compile passes successfully.
• The Nutify container restarts normally.
• The web interface operates normally.
• Commands have been tested successfully against both physical UPS devices.
• Switching UPS in the Multi-UPS interface now also switches the UPS used by the Commands page.

Hopefully this helps narrow down where the Multi-UPS command routing needs updating upstream.

I can provide the complete local diff/patch if useful.

[DartSteven]

Pls make a pull requests