I assume this is your own server?
If so, I personally would do this as the root user:
rm /usr/sbin/rclone
rm /usr/bin/rclone
curl https://rclone.org/install.sh | sudo bash
Once done, re-run:
which rclone
from the path there, run the rclone version
Yes I am the root user
Will this also delete my remotes and configs?
No, it's only removing the binary as the rclone.conf is stored elsewhere.
Thank you!
rclone v1.53.3
Should I now look up what is in the DEBUG log?
If that's all good now, you should retest with the vfs-cache-mode full.
Thanks man! That helped a lot 
Buffering frequency significantly dropped.
Now the movie starts very quick, however, it still buffers every 5 mins.
No transcoding (but it never transcoded before)
Is there any chance for optimization? I would not have any problems to wait a few seconds more until the movie starts so that it doesn't buffer in the following.
(Would it be useful to pause the movie immediately after starting/resuming it, for a while? But this would be - in fact - only a workaround)
Is the server / player in the same location? That sounds more like a bandwidth issue as full mode really takes rclone out of the mix in playing things assuming you have enough bandwidth to sustain the bitrate you are playing.
No. Server is at a data center near to my location (home). My local bandwidth is 500mbit what is used by the player (Nvidia Shield). Server has 1Gbit uplink.
Yeah, that sounds more like an internet issue than anything else.
For play high bitrate things, you need a constant level of banwidth as most plex players have very tiny buffers are very susceptible to a drop in bandwidth.
The problem being that many movies are constant and they have big dips but big ups as well.
My plex server and clients are in my house so any issue is handled by rclone as it reads ahead and would have no issues with a burst.
I have a ~95Mb/s movie going on now:
You can see a nice spike when it started to read ahead and it bounces around while being played.
Okay. Thank you. However, it is a bit strange because I've played movies with even higher bitrates with no problems. This occurred only since some days. Furthermore, my local bandwidth (500 mbit) should be more than sufficient. I can only speculate whether it comes from the present plex (beta) version or something else... but I cannot imagine that my local bandwidth is the problem. 
Finally, you would say that there is nothing that I could do?!
EDIT:
should i post a Screenshot of the dashboard (like you) when streaming the content?
There is a difference between capacity and a constant rate/stream of traffic as internet blips happen every so often. I have a fiber gigabit connection but rated a gigabit levels but that does not mean I always get that throughput and it depends on where I am going and how good the peering is to that.
If it's every 5 minutes on the button, that would make me something might be spiking/limiting it on a server/client.
I've been running Plex for a few years and haven't hit any Plex issues in my setup so I can't speak to that. There have been a few issues/bugs with the NVidia Shields along the way which is why I never got one. If you search the plex forums, you'd see some articles.
All in all, if rclone is behaving, you'd see a pretty constant flow on your server and the file would be cached pretty much anyway as it reads ahead a bit so it would be strange for rclone to be the issue.
Thank you for clarification! That sounds convincing, even if it makes me sad as there is no nice workaround.
Please let me finally ask you one more question as I found out that the only two movies I noticed this issue are "different" from all other of my stored movies:
The beforementioned movie has two custom mixed lossless high bitrate audio lines (first lossless TrueHD Atmos 7.1, second lossless DTS-HD-MA 7.1).
So there is much data that has to be transmitted and processed at the same time ...
Could it be a solution to remove these audio lines from the movie and replace it with the standard version?
However, this is a forum about rclone, so forgive me my last question.
Thank you for spending your time regarding my concerns 
Feel free to ask anything.
As for audio, it's usually pretty small in terms of size. In a 60G file, the TrueDolby track is only 2.5G in my own example. If you have the shield transcoding or something, that could cause issues but from what I understood, the Shield direct plays just about everything.
If you check the Dashboard details, you can see what it's doing or check something like Tautulli.
Thank you!
Further testing today revealed something very interesting: Could my problem be "time-of-streaming"-dependent?
Here are the dashboard details:
This is from streaming now (2:30 p.m. local time).
Whenever i noticed my problem, it was around 8 or 9 p.m. (local time). Currently, it could be an artifact of the covid-19 pandemic since there might be even more traffic (in general). But I certainly do not have a good explanation why the traffic rises at night 
However, as you might see the bitrate is relatively stable after the first peak. And I do not get any streaming buffering.
So, I might watch such movies with my wife earlier 
For this night I'll test two things:
First same settings (later this day) looking at this dashboard info. Second, this certain movie is with Dolby Vision.. I also have a HDR10 x265 encode version of it. The bitrate is strongly reduced (around 30 mb/s). This might result in more smoother streaming.
If you're interested, I certainly would be happy to share this information with you
The scale on that is a bit tough to read as it looks like you are jumping around 30-50Mb/s at spots. The challenge with movies as they are bursty at times and you see those ups and downs and they are usually high parts of action that require more bitrate for a short period of time.
If you have a small hiccup, that's how you get a buffering message since the plex server is remote.
I'd guess on the server, you'd see a fairly similar curve with rclone.
If you are sure it's direct playing, that would really eliminate rclone from the mix imo and I'd focus on lowering the bitrate.
More traffic/congestion would make sense but that's super tough to track down. I'm spoiled and have fiber to my house and run everything local so my use case is very tuned for what I am doing and I have a lot of control to see what's going on and where the problem.
The challenge with a remote server is accurately diagnosing a problem between the server / client as it goes over the Internet to get to you.
I think there is nothing more that you can do for me, but as I promised here are some details of streaming:
here you can see the details:
After a first peak the streaming is relatively constant... here the details:
Unfortunately, in this particular case there was no buffering
So... I guess regarding your assumption that the problem is not rclone-related, you are absolutely right. It might comes from my connectivity
Could a 1Gbit internet connectivity upgrade of my ISP be a solution? Currently I have 500mbit, next level would be 1Gbit.
Finally, I would thank you, for your assistance. I think any further details would be better discussed at the Plex forums. 
That feels a bit like overkill as if you think of it like a water pipe. You have plenty of capacity at 500 but you can't change how fast water flows down the pipe as that's fixed (gravity). If you go to gig, you can a larger pipe, but still haven't changed gravity.
Good luck as debugging internet connectivity is not fun!
Hello!
i have a similar problem trying to view 4k videos from RClone using Plex.
The problem does not come from network connectivity, it comes from RClone itself.
Long story short: Mount is super slow compared to Copy.
I do not know how to fix the issue, some post suggested using RClone beta as it has a workaround a Kernel regression but that did not work for me. Any help with this would be much appreciated thanks you.
Screenshot attached
Best to start a new post and fill out the help and support template.
thanks for suggestion, ended up re-shuffling my setup completely and I see enough improvement to stream the videos now
FYI I switched to DropBox, my theory behind it is that the COPY is sequential but mount is not so much, and therefore either too many login requests OR the Mega encryptions kicks in too often from cold because of implementation details from go-mega. Either way, with Dropbox it mitigate the issue it seems
