Rclone mount not updating

Help and Support
1

I'm currently having an issue with rclone and dropbox, where my mount doesn't seem to be updating. I have two servers, one of the servers downloads files to disk and then uploads to dropbox and the mount is updated immediately.

On my 2nd server, it is a mount only, but is not syncing what's been deleted/uploaded as when doing a 1:1 comparison between both servers, the main server shows files that have been deleted correctly, but on the 2nd server they are still visible. My rclone command is;

rclone mount --allow-other --drive-impersonate admin@jumbo.stream --vfs-cache-mode full --cache-dir /dev/shm/cache/ --vfs-cache-poll-interval 1m --tpslimit 12 --tpslimit-burst 12 --vfs-cache-max-age 24h --vfs-cache-max-size 8G --drive-pacer-min-sleep 5ms --user-agent=JumboStream --multi-thread-streams 1 --vfs-read-ahead 150Mi --buffer-size 300Mi --vfs-read-chunk-size 150Mi --attr-timeout 8700h --dir-cache-time 8760h --poll-interval 1m --no-traverse dropenc: /home/plex/mounts/dropbox/

Is there something that I am missing?

2

if you use a debug log, should see
poll-interval is not supported by this remote

as per rclone docs,
changes made directly on the cloud storage by the web interface or a different copy of rclone will only be picked up once the directory cache expires if the backend configured does not support polling for changes.

so by default, it will take five minutes for the second mount to see changes on the first mount
--dir-cache-time duration Time to cache directory entries for (default 5m0s)

3

So removing

--vfs-cache-poll-interval 1m

Should resolve my issue?

4

no, i shared the relevant flag in the previous post.

5

I see that now. Thank you.

6

yeah, that makes sense, as i was re-edited my post, during the time you posted.

7

if you want, can force a refresh on the mount

  1. add --rc --rc-no-auth to the mount command.
  2. and then run this command
    rclone rc vfs/refresh recursive=true -vv
8

That won't help me to refresh the current mount though? I don't have the add --rc --rc-no-auth to the mount command. on my command.

I just checked my command and I currently have --dir-cache-time 8760h

Is this bad? I question why my mount has been updating fine up until now, though if this is the culprit.

9

edit:
i made a mistake, dropbox does support polling.

i was working on another topic and the OP was using box.
so i got confused

and you did not answer the questions in the help and support template???

10

You have a baffling mount command.

Are you using dropbox or Google Drive as you have a ton of Google Drive commands in there.

Dropbox supports polling so changes should show up.

The super best thing is to actually use the help and support template as that collects all the required information up front so it isn't a 20 question guessing game to what you are seeing and what the logs show. I can't understand at all why people don't use the template as it truly baffles me as I don't know what else to write to make people use it. I guess you just hate puppies.

11

I was using google drive, but then modified my existing gdrive command to work with dropbox.

12

What is the problem you are having with rclone?

I'm currently having an issue with rclone and dropbox, where my mount doesn't seem to be updating. I have two servers, one of the servers downloads files to disk and then uploads to dropbox and the mount is updated immediately.

On my 2nd server, it is a mount only, but is not syncing what's been deleted/uploaded as when doing a 1:1 comparison between both servers, the main server shows files that have been deleted correctly, but on the 2nd server they are still visible. My rclone command is;

rclone mount --allow-other --drive-impersonate admin@jumbo.stream --vfs-cache-mode full --cache-dir /dev/shm/cache/ --vfs-cache-poll-interval 1m --tpslimit 12 --tpslimit-burst 12 --vfs-cache-max-age 24h --vfs-cache-max-size 8G --drive-pacer-min-sleep 5ms --user-agent=JumboStream --multi-thread-streams 1 --vfs-read-ahead 150Mi --buffer-size 300Mi --vfs-read-chunk-size 150Mi --attr-timeout 8700h --dir-cache-time 8760h --poll-interval 1m --no-traverse dropenc: /home/plex/mounts/dropbox/

Is there something that I am missing?

Run the command 'rclone version' and share the full output of the command.

rclone v1.63.0

  • os/version: ubuntu 22.04 (64 bit)
  • os/kernel: 5.15.0-78-generic (x86_64)
  • os/type: linux
  • os/arch: amd64
  • go/version: go1.20.5
  • go/linking: static
  • go/tags: none

Which cloud storage system are you using? (eg Google Drive)

Dropbox

The command you were trying to run (eg rclone copy /tmp remote:tmp)

Paste command here

The rclone config contents with secrets removed.

[dropenc]
type = crypt
remote = dropbox:/DataTest/
filename_encryption = standard
password = ###################################
password2 = ##################################

A log from the command with the -vv flag

Paste  log here
13

I've just filled it out Rclone mount not updating - #12 by Jason_Lucas

14

ok, @Animosity022 knows all about dropbox and mounts, so just follow his/her/its advice.

15

First thing I'd want to do is create a client ID and secret.

Second, I'd clean all the garbage that does nothing and makes no sense. First pass, something like

rclone mount --allow-other --vfs-cache-mode full --cache-dir /dev/shm/cache/ --vfs-cache-poll-interval 1m --tpslimit 12 --tpslimit-burst 12 --vfs-cache-max-age 24h --vfs-cache-max-size 8G --user-agent=JumboStream --dir-cache-time 8760h  dropenc: /home/plex/mounts/dropbox/

Not sure why use

as that gets wiped on reboot and the small 8GB means you keep very little in cache. Anything written that is very large has a chance to consume quite a bit of RAM that route.

Now for testing the issue. Pick one file, one server.

Put the mount in debug log with -vv and --log-file /some/location/rclone.log

and test

rclone copy testfile.txt dropenc:
wait ~1 minute and check the mount with a simple ls on the box

If not there, share the debug log.

16

I am curious what will be this investigation result. I have seen similar tickets in the past and I have suspicion that polling updates do not work for dropbox when there is more that one mount active. Like for some reason only one mount reads polling data and they are not available for the second one. I do not have dropbox account so can't test it myself.

17

I've tested this a bunch and found no issues. I have multiple mounts, tested copy a file directly to the remote and it's usually instant for me as the change notify picks it up.

If there is an issue with the polling, love to see as I've sene the exact opposite as it works better than Google Drive for me anyway.

18

I imagine you have different client_id for each mount. Maybe this is what is needed.

19

Wouldn't imagine that matters as the client ID/secret as just access points to the API.

In any event, when I test, I use the same remote for all the whole process so that's not related in my test case anyway.

20

Oh great. It works!

As @Animosity022 says, I am using the same remote.

As for cache, what would you recommend as a good cache size? I have now moved it out of RAM.