start_move_batch returns 500 (EntryNotFound) on reclaim moves with no current schedule

[sadsfae]

Environment

• quads-stage-3.0.0-20260930.noarch on quads2-stage.rdu2.scalelab.redhat.com

Problem

quads --move-hosts (cron) logs on every reclaim:

Failed to start move batch: Check the flask server logs

Traceback (quads-server):

File ".../blueprints/moves.py", line 125, in start_move_batch
    result = ScheduleDao.start_move_batch(hostnames)
File ".../server/dao/schedule.py", line 430, in start_move_batch
    raise EntryNotFound
quads.server.dao.baseDao.EntryNotFound

Reproduced 2026-09-30 18:25 UTC shrinking cloud05 (e28-h26-000-r650, f30-h17-000-dl360 moved cloud05 -> cloud01 after schedule 465 ended at 18:20) and earlier at 16:01 (cloud03 -> cloud01). 8 EntryNotFound occurrences in journal since deployment.

Root cause

• GET /api/v3/moves (ScheduleDao.get_moves) intentionally reports hosts with NO current schedule as a move to host.default_cloud (reclaim after assignment end).
• POST /api/v3/moves/progress/batch (ScheduleDao.start_move_batch, added in "feat: polling/status API" [RFE] Tracking and Polling API #661) requires get_current_schedule(host) to return a row and raises EntryNotFound otherwise. Reclaim moves never have a current schedule by definition, so every reclaim 500s.
• Impact: move still executes (switch config + move_and_rebuild complete), but move progress is not tracked and the CLI warns; any consumer of schedule_ids is silently skipped.

Expected

start_move_batch should not fail for valid reclaim moves. Hosts without a current schedule should be skipped (no schedule row to mark pending); host-not-found should still raise.

Proposed fix

In ScheduleDao.start_move_batch, continue (skip) when get_current_schedule returns empty instead of raising EntryNotFound.

[sadsfae]

Fix uploaded: https://review.gerrithub.io/c/quadsproject/quads/+/1252676 (branch movebatchfix, off development).

• ScheduleDao.start_move_batch now skips hosts without a current schedule (valid reclaim moves) instead of raising EntryNotFound; host-not-found still raises.
• Regression test added: start_move_batch with mixed hosts returns 201 and omits the unscheduled host.
• Verified locally against the podman postgres test DB: tests/api 310 passed; the only failure (test_valid_now_keyword, 3600s TZ offset) reproduces on base code and is unrelated.