but i offered the OP two options.
i thought either of them would resolve the RC interface timing out issue, correct?
--- --timeout - not ideal with freebsd as rclone does not poll the mount.
--- _async=true - not ideal as need to use other rc commands to know when the prime of the vfs dir cache has completed and completed without error.
Timeout is just related to the HTTP requests going out so not related to polling as that would be dir-cache-time for a non polling remote.
Bit tricky as this hides the error. I believe if you submit to a mount with a default timeout, it would just fail in the background and you'd check to check for the status to know it failed.
~16 hours ago I started two commands: rclone mount nv:shared /mnt/shared -vv --log-file /var/log/rclone.log --debug-fuse --rc --read-only --allow-other --dir-cache-time 96h
...and after that: rclone rc vfs/refresh recursive=true --timeout 24h -vv --log-file /var/log/rclone-rc.log
rclone-rc.log:
2022/02/14 01:20:11 DEBUG : rclone: Version "v1.57.0" starting with parameters ["rclone" "rc" "vfs/refresh" "recursive=true" "--timeout" "24h" "-vv" "--log-file" "/var/log/rclone-rc.log"]
2022/02/14 12:14:06 DEBUG : 2 go routines active
2022/02/14 12:14:06 Failed to rc: connection failed: Post "http://localhost:5572/vfs/refresh": EOF
Command rclone rc ... finished a few hours ago and I started borg backup from mounted s3 bucket.
Currently I didn't hit timeout yet.
/var/log/rclone.log is currently 7G. In case I hit read timeout, what should I grep in this log?
P.S: my s3 works pretty fast. It answers slow because it has 5 millions of small files (1T total). It's not fast task to get metadata for all of them...
Backup finished without errors.
When I did rclone rc vfs/refresh recursive=true --timeout 24h, rclone mount was performing caching operation ~12 hours and comsumed ~8G RAM.
Is there way to avoid cache operation and make rclone mount do not interrupt read operations by timeout?