FFMPEG Error
Hey everyone — I recently installed Immich on a fresh Dell OptiPlex 3020 (Intel i3-4130 @ 3.4 GHz, 16GB RAM) and a TerraMaster D2-320 DAS (with two mirrored 1TB drives) running Linux Mint 22.1. I followed the official docker-compose.yml setup and used immich-go to import ~250,000 images (~200GB) from a Google Takeout export. I have it running behind Cloudflare. Everything appears to be working well overall, but I’m noticing recurring ffmpeg errors in the Docker logs, like this:
ERROR [Microservices:{"id":"0bf66864-1cee-4ede-9507-c7c12c6575d0"}] Unable to run job handler (generate-thumbnails): Error: ffmpeg exited with code 244: Conversion failed!
...
[libwebp encoder] WebPAnimEncoderAssemble() failed with error: 1
...
Error flushing encoder: Cannot allocate memory
My system has plenty of free memory (confirmed via free -h), and no memory limits are set on the containers. I’m wondering if this could be related to thumbnail generation for specific video files. I saw that only two frames were decoded in one failed case.
I’m using the ghcr.io/immich-app/immich-server:release image and the Jellyfin version of ffmpeg (7.0.2). Any guidance on:
• Why might this be failing intermittently?
• Can I safely reduce thumbnail concurrency?
• Should I retry failed jobs manually?
I'm happy to share more logs or config if it's helpful - thank you in advance!
ERROR [Microservices:{"id":"0bf66864-1cee-4ede-9507-c7c12c6575d0"}] Unable to run job handler (generate-thumbnails): Error: ffmpeg exited with code 244: Conversion failed!
...
[libwebp encoder] WebPAnimEncoderAssemble() failed with error: 1
...
Error flushing encoder: Cannot allocate memory
My system has plenty of free memory (confirmed via free -h), and no memory limits are set on the containers. I’m wondering if this could be related to thumbnail generation for specific video files. I saw that only two frames were decoded in one failed case.
I’m using the ghcr.io/immich-app/immich-server:release image and the Jellyfin version of ffmpeg (7.0.2). Any guidance on:
• Why might this be failing intermittently?
• Can I safely reduce thumbnail concurrency?
• Should I retry failed jobs manually?
I'm happy to share more logs or config if it's helpful - thank you in advance!
