Fixed
- Jobs stuck "running" with a spinner are now self-healed. A worker task that ended without writing a terminal status left a stale
runningrow — a permanent spinner, and the visible running count drifted above the parallel limit. The worker now reaps orphanedrunningrows (local jobs with no live task, past a short grace window) back topendingevery 60s so they re-run cleanly.
Docker images
Pinned to this release:
docker pull ghcr.io/i-ial9000/shrinkerr:v0.9.108 # portable (amd64 + arm64, CPU)
docker pull ghcr.io/i-ial9000/shrinkerr:v0.9.108-nvenc # x86_64 + NVIDIA GPU (NVENC)
docker pull ghcr.io/i-ial9000/shrinkerr:v0.9.108-edge # as above but ffmpeg master
docker pull ghcr.io/i-ial9000/shrinkerr:v0.9.108-edge-nvencFloating tags (:latest, :edge, :nvenc, :edge-nvenc) track the most recent release of each lineage.