I’m noticing the same issue.
Ah good. Well not good, but then it’s not my system =)
First noticed this behaviour with 1.5.1. I downgraded back to 1.4.4 but since the bandwidth cap is also a biggy, I had to upgrade to 1.5.2.
Changing from /dev/shm => local temp drive did solve my issues (but obviously will cost me some performance. Log files only showed something like ‘cannot write to /dev/shm’. When this error occured I was able to manually write something to that location. I concluded it was a Plex issue, but if anyone else has suggestions…
anyone has amazon hostname(not IP) ?
i want just the hostname where it files come from, i will extract ip-list from it and block slow server ip:s
it seems 50% of them are very slow. and because of that the file transfer is very slow…
Are all of the nodes accessible from all endpoints?
it seems so, found the hostname used, its HUGE LIST alot IP:s
i tried one by one IP like 10-20 of them, and there is all have same problem.
the download from cloud comes full speed, but the seek access time is random, sometimes it takes 1sec and sometimes like 30sec to a transfer to start. LOL. im running out of options, the amazon cloud was fast 6months ago, but now it is very slow.
Totally agree. I’ve seen a massive degrade in performance over the past few weeks in particular.
the degradation is about seeking the file, it takes sometimes 1-2sec to access file and sometimes like 30+sec… that is doing this drive slow… this was not like this a month ago… more users mean more lag and more time to access file.
lowering –timeout to 10s fixed some issues, but well… it bring some things along, its under test and tweak…
im using s3ql so its a chunked files, i need the fast file seeking.
Latest plex pass version (Version 1.5.4.3596 ) fixed the problem. Transcoder would crash if inotify was not supported.
2 Likes
Tnx for the update, i’ll set my temp folder back to shm =)
hi there,
do you have any idea why suddenly the rclone ACD mount is “almost” inaccessible slow?
I have two servers, one in Germany and one in the Netherlands and today I’m experiencing very slow folder changes on both
/ACD/.acd
/ACD/acd
/ACD/media
are very slow to query
My rclone mount is very simple, I was using this in the past day without any issues:
rclone mount acd:Plex _/blabla/_ACD/.acd &
As its happening on both servers
I’m thinking either Amazon is extremly slow today, or my rclone mount has some difficulties
Or there are somehow “stuck” processes which are eating up the mount…
I have also 2 times rebooted the servers…
see this post: ACD BAN/Alternatives discussion (error HTTP code 429)
Problem on amazon side for everybody
2 Likes
Is there a standard way to pass several arguments to rclone, at the moment when i use the script from @Ajki github it fails stating maximum 2 arguments.
Cheers, Geof
you may want to paste an example of command.
Geof. I just did this yesterday. If you copy straight from his github into a script it will not work because it is not formatted correctly. This works (or it would if amazon was work at all right now):
rclone mount
–read-only
–allow-other
–acd-templink-threshold 0
–stats 1s
–buffer-size 1G
–timeout 5s
–contimeout 5s
-v
–log-file=/home/rclone/rclone.log \
Do you mind sharing these scripts? I’m trying to setup a similar process and would love a template to work off of.
1 Like
Hey guys,
I have a few mounting (rclone 1.36 without fuse Ubuntu 14.04.5) questions:
-
Is the -acd-templink-threshold flag fuse related? Everytime I try to use it, I’ll get an flag unknown back.
-
with --buffer-size 1G the connection gets lost very often. Dedi with bytsized, 16 GB RAM and Intel Xeon E3-1200
-
Everytime the connection to the mount breaks, after remounting all files that are stored there are unavailable and I have to rescan the library manually. Is there another maybe automated solution for that?
thx 
i hope you know, that a buffer size of 1G actually means, that every simultaneous open file connection gets a buffer of 1GB. how many users do you have?
yes I know that, there are no more than two or three at the same time. Just five users have access to my server, all direct play their streams no transcoding.
+1 for sharing those script, @saitoh183 
I started to work on a plex scanner script to have the values added (by flexget) to it for the update
but if there is any working solution even better
1 Like
I will try to post it later tonight
1 Like