Command line switch took longer than the scheduled update interval *recently*

Configuration
1

I have had a few command line switches working perfectly fine for about 2 years. I usually apply updates as soon as they become available (w/ backups), however I cannot pin point the exact update that caused this issue.

My switch looks like:

- platform: command_line
  scan_interval: 90
  switches:
    test_switch:
      command_on: "ssh -o UserKnownHostsFile=/dev/null -o StrictHostKeyChecking=no -i /config/.ssh/id_rsa [email protected] 'some linux command'"
      command_off: "ssh -o UserKnownHostsFile=/dev/null -o StrictHostKeyChecking=no -i /config/.ssh/id_rsa [email protected] 'some linux command'"
      command_state: "ssh -o UserKnownHostsFile=/dev/null -o StrictHostKeyChecking=no -i /config/.ssh/id_rsa [email protected] 'some linux command'"

I confirmed that these commands work w/o password request and perform the desired actions when I execute them manually by dropping into docker comatainer that HA lives in by issuing the following command.

docker exec -it homeassistant /bin/bash

The ssh line executes and everything w/o issues. However, when HA attempts to do it I see the following in my logs

Updating command_line switch took longer than the scheduled update interval 0:01:30

Originally I had 30 for update interval as the commands only takes like 2 seconds to execute but increased to 90 but no success.

If I restart HA it seems like to switches update but then stop working right after and the above entries are in my logs.

Any ideas?

2

Did you ever resolve this? It just started happening to me.

3

I noticed the same warning shortly after upgrading to 2023.06, but the command only took a second to complete when I tried it from the shell.

I am also starting to get this:
Command failed (with return code 255): ssh dockerservice@(redacted)

When tested from the shell (ssh), return code is 0.

4

I am also seeing this on all of my command line switches which have been working fine for at least a year. They are still functioning, but report this message every time they execute the status call. Again, like the other posters, the commands when execute from the shell take a fraction of a second. Has anyone got to the bottom of this yet?

5

This has been fixed and will be released with .7 (Hopefully!) see: