v1.10.1
What's new since v1.10.0
A maintenance release: three scheduler and diagnostics fixes found while investigating failures on a live instance.
Fixed
- Scheduled searches keep their rhythm across restarts. A job configured to search every 4 hours was re-anchored to boot time on every container start, so on a day with several restarts the real gaps stretched to 6.5 and 7.7 hours. Startup now leaves an already-correct schedule alone; only an actual interval change reschedules the job.
- The cleanup job no longer reports false timeouts. Its one-hour ceiling sat below what a large library legitimately needs, so long-but-healthy runs were recorded as failures even though the work ran to completion. The limit now reflects the measured runtime.
- Failed media probes state their cause. Probe errors were reported with an empty reason because the underlying ffprobe call suppressed its own diagnostics. The cause — a missing file, an unreadable container — now appears in the log.
Tested via
v1.10.1-rc.1 on the RC instance. The scheduling fix was additionally verified on a separate instance across a real restart: the next-run anchor survived unchanged, and it survived the production restart on promotion too.
Deployed: ghcr.io/abrechen2/sublarr:1.10.1 (also on Docker Hub as abrechen2/sublarr:1.10.1, multi-arch amd64 + arm64)
Docs: https://sublarr.de/docs/