I’ve gone back to trying to get rclone cache working as I just can’t get unionfs to work for me (sonarr/radarr won’t import new files). My situation is I’ve created a fresh cache-tmp-upload-path and I’m transferring old pending files from the old cache-tmp-upload-path (different folder) via a standalone rclone move job i.e. the new cache-tmp-upload-path is only getting new files.
At the moment radarr is adding a 17GB file. Not only is it taking forever, but it’s hammering my RAM and CPU:
I have a good spec system - E5-2683V3 and 64GB of ram so this is very surprising. Are there any settings on rclone cache that will stop it being a resource hog or speed it up? e.g. could the pending files go to disk rather than ram if that’s what’s happening? I only use the cache to upload files to gdrive and to keep radarr/sonarr in sync as I play files via a separate vfs mount, so I dont’ need things to happen in ram for ‘speed’
I must be doing something wrong as this can’t be normal. Thanks in advance for any help
# Mount rclone vfs mount
rclone mount --allow-other --dir-cache-time 96h --cache-dir=/tmp/rclone/vfs --vfs-read-chunk-size 64M --vfs-read-chunk-size-limit 1G --buffer-size 256M --syslog --umask 002 --rc --log-level INFO --stats 1m gdrive_media_vfs: /mnt/disks/rclone_vfs &
# Mount rclone cache
rclone mount --allow-other --dir-cache-time=360h --cache-chunk-path=/tmp/rclone/cache --cache-db-path=/tmp/rclone/cache --cache-chunk-size=512M --cache-total-chunk-size=8G --cache-info-age=368h --cache-db-purge --cache-tmp-upload-path=/mnt/user/rclone_upload --cache-tmp-wait-time=1h --log-level INFO --stats 1m gdrive_media: /mnt/disks/rclone_cache &
# Mount old rclone cache so can decrypt files and upload via rclone move job
rclone mount --allow-other --dir-cache-time 5m --cache-dir=/tmp/rclone/old_cache --vfs-cache-max-age 1h --vfs-read-chunk-size 128M --vfs-read-chunk-size-limit 512M --buffer-size 128M --syslog --umask 002 --rc --log-level INFO --stats 1m rclone_cache_old: /mnt/disks/rclone_cache_old &