Try reconfiguring the backend (back up the config file first) and see if that helps.
I still get the exact same error. Copying over with 1.58.1 files again and it starts working again with no issues.
I redid the config, updated the token using legacy authentication, got my sms and everything seems to work except when mounting it complains with the error i posted.
I also did a new test mount and it also displayed exact same error and added suffix to name aswell.
I got no clue.
Hi Nick, do you have an ETA for bringing back multithread downloads when cache full is used? If you don't that's fine, obviously.
Please follow the issue:
Tried my best to make one
I am already. Nick stated in April ‘21 that the feature was gonna come back soon and that’s why I was asking if he had a time frame in mind. Thank you.
I had a sponsorship deal to implement multithread downloads into the vfs cache, but unfortunately it fell though at the last minute I don't currently have a timescale for implementing it, but I do have quite a good plan which is a start!
Thanks for the reply. Not sure what you mean by "sponsorship deal" but I get that there have been setbacks. As always, you don't owe us anything and I was asking only due to Drive's problems with single threaded downloads... which are not your fault to begin with.
I''l wait. Thanks again!
what is the difference between Union and Combine remotes?
The docs are not exactly clear in what will be used when this is enabled, only size?
Also perhaps if you are really like me and just want maximum performance and have files that NEVER change, wouldn't be better to just disable all fingerprinting?
Then if I really update a file in the cache, i should be able to use a RC command to force fingerpriting check to refresh them or something like that
I did and it's still not clear? Union intertwines all files and combine only directories?
It depends on the backend. For local this will just use modtime and size when it's on and it will include hash when off.
For drive it will make no difference.
For S3 it will use hash and size when on and include modtime when off.
Setting this to on will speed things up slightly if you are using s3 but not at all if using drive.
With fast fingerprint, removing the check wouldn't speed things up as rclone doesn't have to do any network operations to check it.
Union intermingles files but combine attaches remotes to top level directories without intermingling files.
Here is the example from the docs which hopefully makes a bit more sense now.
The combine backend joins remotes together into a single directory tree.
For example you might have a remote for images on one provider:
$ rclone tree s3:imagesbucket
And a remote for files on another:
$ rclone tree drive:important/files
The combine backend can join these together into a synthetic directory structure like this:
$ rclone tree combined:
│ ├── file1.txt
│ └── file2.txt
That assumes your backend is local, but not actually a network mount mounted, so you'd still have to make external connections to retrieve size of the items.
Is the use-case of combine for mounting? You can mount a single remote with a single process as opposed to one independent rclone process for each?
Yes, that's perfect use case. Also mounting lots of shared drives individually.
This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.