Cache Mount Samba Share Issue?


#1

So I am having issues /w my samba share and I am not sure where the issue is. I can access the root level, but nothing any lower levels.

I am able to crreate files and all at \MEDIASERVER\local-media.

This is my samba share:
[local-media]
writeable = yes
guest account = dulanic
path = /local-media
public = yes
force user = dulanic
force group = dulanic

Here is my mount systemd service:
[Unit]
Description=rclone mount
After=network-online.target

[Service]
TimeoutStartSec=10
ExecStart=/opt/torrent/rclone mount gd-crypt: /local-media --config /home/dulanic/.config/rclone/rclone.conf \
        --gid=1000 --uid=1000 --tpslimit=4 --cache-workers=5 \
        --cache-chunk-path=/rclone/chunks --buffer-size=0M --cache-db-path=/rclone/cache \
        --cache-tmp-upload-path=/chunks/rclone/uploads --umask 000 \
        --allow-other --allow-non-empty --cache-total-chunk-size=250G --cache-chunk-size=10M \
        --dir-cache-time=160h --cache-info-age=168h \
        --log-file=/var/log/rclone.log --log-level=INFO
ExecStop=/bin/fusermount -uz /local-media

[Install]
WantedBy=multi-user.target

Here is my rclone.conf:

[gd]
type = drive
client_id = xxxx
client_secret = xxxx
token = xxxx

[gd-crypt]
type = crypt
remote = cache:
filename_encryption = standard
password = xxxx
password2 =
directory_name_encryption = true

[cache]
type = cache
remote = gd:
cache-dir = /chunks/rclone
cache-mode = full
plex_url = http://127.0.0.1:32400
plex_username = dulanic
plex_password = xxxx
chunk_size = 10M
plex_token = xxxx

Debug log, but I think this might be the wrong type of log to be viewing:

https://pastebin.com/JYJAKT6b


#2

What version of rclone are you using?

This sounds very similar to Rclone 1.39 vs 1.40 mount issue which should be fixed in 1.41 or the latest beta.


#3
rclone v1.41
- os/arch: linux/amd64
- go version: go1.10

#4

Try becoming the samba user (or whatever the user is called that runs the smb daemon on your linux box. eg sudo su - smb -s /bin/bash then check to see if you can see inside the mount and inside the subdirectories.

It looks like you’ve set all the things you need to like --umask 000 and --allow-other so I’m not sure why it isn’t working.


#5

The user is dulanic AKA user 1000, so it can deff access and is the labelled owner. Unless you mean something else? Samba itself is run by root and it can deff access.

dulanic@mediaserver:/local-media$ ls -l
total 1
drwxrwxrwx 1 dulanic dulanic 0 Apr 29 00:00 Media
-rw-rw-rw- 1 dulanic dulanic 2 May  5 23:23 mountcheck.txt
drwxrwxrwx 1 dulanic dulanic 0 May  3 19:03 torrent

dulanic@mediaserver:/local-media$ sudo ps all -A | grep .mbd
5     0 14243     1  20   0 338292  3296 poll_s Ss   ?          0:00 /usr/sbin/smbd -D
5     0 14247 14243  20   0 329900  1628 poll_s S    ?          0:00 /usr/sbin/smbd -D
5     0 14255 14243  20   0 338816  3020 poll_s S    ?          0:00 /usr/sbin/smbd -D
5     0 14323     1  20   0 240140  3520 poll_s Ss   ?          0:01 /usr/sbin/nmbd -D
5     0 24884 14243  20   0 345600 16004 poll_s S    ?          0:00 /usr/sbin/smbd -D

#6

Here is something else to try… Try increasing this to 60s and see what happens. Increasing it from 0 to 1s fixed this issue last time - maybe it will help?

  --attr-timeout duration              Time for which file/directory attributes are cached. (default 1s)

#7

Adding that at 60s seems to have corrected the issue. Strange fix, but hey if it works it works! Thanks for the idea.


#8

I’m glad that works. I don’t have a clue why though! Probably something to do with file system traversals through the kernel…


Rclone, mounting and samba