Reservation allocated to place without resources

[brherger]

Hello! Thank you for a great tool.

Environment

• Linux
• Python 3.10
• labgrid v23.0.6

Background We are making use of the Place Scheduling functionality using tags to reserve and acquire.

Problem: When a device (with an exporter) goes offline, then the place exists in the coordinator, but its resource capabilities are rightfully lost. It does not seem like the reservation logic takes this into account as it can allocate a place without any resources, which ultimately will fail to be acquired.

$ labgrid-client -p some-device show
Place 'some-device':
  tags: board=foo
  matches:
    */some-device/*/*
  acquired: None
  acquired resources:
  created: 2024-08-09 14:20:53.035338
  changed: 2024-08-13 15:24:40.631180
$ labgrid-client resources
$ labgrid-client reserve board=foo
Reservation 'M29NQZJS42':
  owner: host/user
  token: M29NQZJS42
  state: allocated
  filters:
    main: board=foo
  allocations:
    main: some-device
  created: 2024-08-15 15:21:41.387948
  timeout: 2024-08-15 15:22:41.387987
$ labgrid-client -p +M29NQZJS42 acquire
labgrid-client: error: Match */some-device/*/* has no matching remote resource

I am wondering if there is a possible workaround to this situation aside from passing name=<place-name> to the reserve call, as this defeats the purpose of reserving any of the possible tag matches. Can I pass the resource name or something similar?