I never used cache. My mount point was always a gdrive > crypt
remote. My drive will soon reach 200TB everything included. So there are a few more files to be scanned
You are right, rclone
will only download the needed bytes, that mediainfo reads. But the call that rclone
sends to the remote will request the whole file, starting at byte 0 until the end. This ârequested bytesâ number is counted towards the daily quota, not the âdownlaoded bytesâ (at least for Google Drive).
My config is quite simple:
[gsuite]
type = drive
client_id = my.client_id.apps.googleusercontent.com
client_secret = my_client_secret
token = {...}
[gsuite-crypt]
type = crypt
remote = gsuite:data
filename_encryption = standard
password = password
password2 = password2
When running a Plex scan using this config against letâs say 1TB new files, this would result in a 24h ban.
Plex will open every file, read some data and seek to multiple positions in the file to collect the metadata.
As described above, every open and seek will count to the ârequested bytesâ limit (which seems to be 10TB for Google Drive). When Plex seekâs 10 times, 1TB of new files will exceed the 10TB limit.
Here is a example strace of mediainfo where this behavior can bee seen.
I annotated the calls that trigger the rclone request with the byte range requested. Running mediainfo on this 77 GB file caused 3 opens and 13 seeks, in total 16 requests send by rclone. Summing up the bytes requested, this would add 1080.09 GB to the ârequested bytesâ limit.
Running the same command with --vfs-read-chunk-size 128M
would only add 2 GB to this limit.
I hope I could explain the difference between âdownlaoded bytesâ and ârequested bytesâ.