My VFS SweetSpot - Updated 30-Aug-2018


Thanks, I’ve just created my own API with that little tutorial. Hopefully I can see some API hits tomorrow. Will update later.

@BinsonBuzz: I’ve moved from a 200/20 to a 500/500 so I expected at least some more speed, so it’s weird to me. I will try your advise for the changed buffer!

For the rclone rc command, do you run it as a standalone snippet right after the mount or is it integrated in a longer string?

Currently I’m not testing with Plex but with Emby, since previously I had better results with that. My library scan just finished on Emby so I will test some more tomorrow. Then I will also compare it to Plex with your throttle buffer setting.


As a standalone snippet right after the mount - that’s the full string


I think I didn’t do it right yesterday with my own API, didn’t get any hits. So I created the API and deleted my rclone mount and tried to reconfigured. So I go through the rclone wizard, with client ID filled it from my dashboard. Then I get the 0auth screen and paste in the code. Then I get the following error:

2018/11/06 09:59:38 Failed to configure token: failed to get token: oauth2: cannot fetch token: 401 Unauthorized
Response: {
“error”: “invalid_client”,
“error_description”: “Unauthorized”

So could it be that my 0auth hasn’t been processed by Google yet, or am I doing something wrong? I only got asked for the Client ID during rclone config, not for the client password.

EDIT: just went through the Rclone config for the 5th time and this time it did ask for my client secret from the start. So now I’ve got this configured. Hopefully I will get api hits now.


I’ve put this code behind my mount but it gives me the following error:
Failed to rc: connection failed: Post http://localhost:5572/vfs/refresh: dial tcp connect: connection refused

My code looks like this now:

rclone mount --allow-other --dir-cache-time 72h --vfs-read-chunk-size 128M --vfs-read-chunk-size-limit off --buffer-size 100M --log-level INFO gdrive_crypt: /mnt/user/mount_rclone/Gdrive --stats 1m &

rclone rc --timeout=1h vfs/refresh recursive=true


Sorry for dirtying up the thread. I’ve got the API working and the performance now is amazing! Startup times are within 5 seconds for my short tests with Emby Will also try Plex, see how it goes there. This feels like my previous performance again. Really amazing how big of a difference this makes!

Now if I could get that RC working then I’m pretty much set again I hope :).


hmm just checked my logs and I have the same - it was working before. Will investigate tonight.

Glad it’s working - basically it’s about the same as spinning up a local disk :slight_smile:

I’ve got 80% of my media on gdrive now and I only really keep some of my 4K stuff local just to remove the bandwidth requirements, not because rclone can’t cope


What did you use in the end for your mount BTW? drive-chunk-size 32M or 128M?


Currently testing with 128m. But will try 32m later today. What made you switch back from 128 to 32?


Sorry, I meant vfs-read-chunk-size 64M Vs your vfs-read-chunk-size 128M

I went with 64M as I was trying to get faster start times, but I might test again with 128M as I got upgraded last month from 200/200 to 300/300, so it might be a bit faster.


Happy to hear it’s working well for you. Using your own API key will definitely do better.


I saw you somewhere post your Plex settings to stop analyzing. I browsed through your posts, but can’t find it anymore. Would you mind sharing it here again? I’m noticing Emby performs quite a lot better than Plex, which shouldn’t happen.


Well, it does have to analyze all the files in both Plex and Emby.

I keep all the thumbnails/previews/etc off:

Everything else really doesn’t matter too much with the exception of deep analyze. The box is extensive media analysis.

Emby does a ffprobe when it adds a file to the library for every item to get the information which is similiar to what Plex does as well.

I use a script to check the analyzed files in Plex:

06.11.2018 09:00:41 PLEX LIBRARY STATS
Media items in Libraries
Library = Movies
  Items = 1977

Library = TV Shows
  Items = 20295

Library = xMMA
  Items = 53

Library = zExercise
  Items = 279

Time to watch
Library = Movies
Minutes = 211705
  Hours = 3528
   Days = 147

Library = TV Shows
Minutes = 816201
  Hours = 13603
   Days = 566

Library = xMMA
Minutes = 9240
  Hours = 154
   Days = 6

Library = zExercise
Minutes = 11278
  Hours = 187
   Days = 7

22731 files in library
0 files missing analyzation info
0 media_parts marked as deleted
0 metadata_items marked as deleted
0 directories marked as deleted
22606 files missing deep analyzation info.

Once that files missing analyze is 0, Plex scans are just about instant (10-20 seconds).


Thanks! That made me change some settings (not a lot). So I will test a bit with the read chunk size to see if that speeds up playing time. But I think it doesn’t make much a problem. At this point the mediaplayer device is probably the only limitation.


If I am not misreading it, you probably need to add the --rc argument to the rclone mount for the rc commands to be available.


yep - fixed it:

rclone mount --rc --rc-addr= --allow-other --buffer-size 1G --dir-cache-time 72h --drive-chunk-size 32M --fast-list --log-level INFO --vfs-read-chunk-size 128M --vfs-read-chunk-size-limit off gdrive_media_vfs: /mnt/user/mount_rclone/google_vfs


How did you come up with rc-addr= is that the error you got back? I got a different IP, that’s why Im asking.


The IP address he is used is more related to his environment.

If you plan to run the rc commands from the same box, you can just use --rc and it’ll bind to the local loopback, which means only that box can access it.

I just use the --rc and you can run it like:

felix@gemini:~$ rclone rc vfs/refresh recursive=true


What am I missing here then, this my error:

2018/11/06 15:46:56 NOTICE: Serving remote control on
2018/11/06 15:46:56 ERROR : rc: “vfs/refresh”: error: couldn’t find method “vfs/refresh”
2018/11/06 15:46:56 Failed to rc: Failed to read rc response: 405 Method Not Allowed: {
“error”: “couldn’t find method “vfs/refresh””,
“input”: {
“recursive”: “true”

My rclone mount:
rclone mount --rc --allow-other --buffer-size 1G --dir-cache-time 72h --drive-chunk-size 32M --fast-list --log-level INFO --vfs-read-chunk-size 128M --vfs-read-chunk-size-limit off gdrive_crypt: /mnt/user/mount_rclone/Gdrive --stats 1m &

And the seperate code after the mount:
rclone rc vfs/refresh recursive=true


What does rclone version show?


rclone v1.43-150-g249a523d-beta

  • os/arch: linux/amd64
  • go version: go1.11.1


Update to 1.44 which is the latest version.