[BUG] --cloud-only returns false data for some hosts, schedule logic issue

[sadsfae]

Summary

• We have noticed what appears to be a CLI bug and the way the wiki cloud membership differs from --cloud-only output.
• It appears for at least (2) recent e19 R660 hosts their first schedule is not recognized as active.

These are the hosts to look at, currently released in cloud51

• e19-h21-000-r660.example.com
• e19-h22-000-r660.example.com

Details

This environment here cloud51 has 10x hosts evident from both system schedules as well as the wiki:

d41-h11-000-r660.example.com
e19-h07-000-r660.example.com
e19-h21-000-r660.example.com
e19-h22-000-r660.example.com
e19-h23-000-r660.example.com
e19-h25-000-r660.example.com
e19-h26-000-r660.example.com
e20-h02-000-r660.example.com
f25-h16-000-r660.example.com
f25-h18-000-r660.example.com

However --quads--cloud-only --cloud cloud51 only lists 8 of the 10 hosts, even though they have correct schedules.

quads --cloud-only --cloud cloud51 | wc -l
8

Here we can check the hosts we should have listed and count:

d41-h11-000-r660.example.com  = cloud51
e19-h07-000-r660.example.com  = cloud51
e19-h21-000-r660.example.com  = cloud01
e19-h22-000-r660.example.com  = cloud01
e19-h23-000-r660.example.com  = cloud51
e19-h25-000-r660.example.com  = cloud51
e19-h26-000-r660.example.com  = cloud51
e20-h02-000-r660.example.com  = cloud51
f25-h16-000-r660.example.com  = cloud51
f25-h18-000-r660.example.com  = cloud51

Note that 2/10 hosts show cloud01 via --ls-host-cloud however let's look at their current schedules:

for host in e19-h21-000-r660.example.com e19-h22-000-r660.example.com; do quads --ls-schedule --host $host; done
Default cloud: cloud01
Current cloud: cloud01
8060| start=2026-01-18T22:00, end=2026-02-08T22:00, cloud=cloud51
Default cloud: cloud01
Current cloud: cloud01
8061| start=2026-01-18T22:00, end=2026-02-08T22:00, cloud=cloud51

Experimenting further, I tried to cancel a schedule on one of these so they move back to cloud01

quads --shrink --host e19-h21-000-r660.rdu2.scalelab.redhat.com --now

The selected host does not have any active schedules
Would you like to shrink a future allocation of e19-h21-000-r660.example.com? (y/N): y
Traceback (most recent call last):
  File "/usr/bin/quads", line 33, in <module>
    sys.exit(load_entry_point('quads==2.0', 'console_scripts', 'quads')())
             ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^
  File "/usr/lib/python3.13/site-packages/quads/cli/main.py", line 64, in main
    _exit_code = qcli.run(
        action=cli_args.get("action"),
        cli_args=cli_args,
    )
  File "/usr/lib/python3.13/site-packages/quads/cli/cli.py", line 67, in run
    exit_code = action_meth()
  File "/usr/lib/python3.13/site-packages/quads/cli/cli.py", line 840, in action_shrink
    if end_date < schedule.start or end_date > schedule.end or (not now and end_date < threshold):
       ^^^^^^^^^^^^^^^^^^^^^^^^^
TypeError: '<' not supported between instances of 'datetime.datetime' and 'str'

I believe I can "trick" QUADS by modifying the current cloud by setting it''s current assignment_id to active and/or cloud_idto that ofcloud51`but I think this needs to be examined closely in its current error state first.

[sadsfae]

Apparently these were marked broken, not a bug.