Problem:
Not sure exactly when this problem started but it has never behaved like this before. The issue is that when i play something it starts ok but after a while (few minute) and sometime even from start in some cases the played (on my LG Tv) starts to show buffering.
Right now when i play a movie it plays 2 second and stops 1 second like its “hacking”. As if the data is not being transcoded or delivered to the player in time or something.
Not fully sure when the problem started. Could be a few plex server updates before. Could also be the player itself on the tv which was also updated recently. Sadly i cant downgrade.
I know for sure its not performance on the QNAP since i use barely 25% of the CPU and i have atleast 8 GB of RAM free.
As an example i played this:
I removed all the logs since there were allot of files and played the movie when the issue happened. Also restarted Plex server on NAS + TV to be sure. Logs.7z (149.7 KB)
I read also that subtitles cause issue here (looked into other threads too but didnt find anything that helps):
Although i never had any issue until just recently. I have the “Burn Subtitles” on Automatic and tried “only image” as well but didnt help. But for sure i dont get the lag or buffering when i remove the subtitle. Considering my PC has lots more power and i have plenty of RAM i dont get why it lags.
I connect everything using cable so there shouldnt be any issue there either.
Just for the sake of testing i also used an AppleTV (new ones, think 3rd gen) to play the same movie with subtitle and it works like a charm. Actually much much better. I kept pushing with fastforward and all and it never slowed down.
I hope someone can figure out what the issue is here. I really like to use my TV for plex as i have done for the last year.
The logs seem to be from Sept. 9th, but they refer to EAE can’t write to it’s directory errors.
That is the Easy Audio Encoder program from Plex.
We should get some Plex Media Server logs from the QNAP.
Can you open a web browser to https://app.plex.tv/desktop and go into Settings - Troubleshooting - Download Logs please? Zip up what it gives you
and attach it with your next reply.
Thanks for looking at the files. I actually used sftp and copied them directly from
/share/CACHEDEV1_DATA/.qpkg/PlexMediaServer/Library/Plex Media Server/Logs
But here comes the zip file from where you pointed out.
May i ask which file(s) you look into? i try to read some but the info is overwhelming.
Today i just borrowed my daughters appletv and it works just fine. Wish i could use my LG as i use to.
Hope the new logs can show something more. I have not done any changes at all except updating the software. Plex Media Server Logs_2019-10-13_11-32-28.zip (1.5 MB)
I sort by the most recently modified, and look at the
Plex Media Server.log
com.plexapp.system.log
which usually go back far enough for me to gather evidence by searching for
error
fail
exception
warn
sqlite3
network
decision
EAE_ROOT
that usually points me to sections of the logs I need to then read the raw stuff around the results.
In your case with low CPU usage, it’s not clear what it will be besides things interrupting the network traffic.
It looks like Plex is aware of this issue and has released a new update. My problem is, I do not know how to manually update unraid because this update will not work automatically. Plex does not have a direct support avenue as they rely on the community supporting the product.
I have downloaded the new update, but does anyone know where I should copy it to or how to manually update PMS in unraid?
Installed Version 1.18.0.1944
Even rebooted the NAS.
Playing Live TV and DVR shows still shows pixellation for about 45 seconds when starting any recorded or live show.
This pixelation only appears when playing through PLEX, all other clients play with zero pixellation.
So I’m looking at the last logs you uploaded. I can only find references to you playing videos on an AppleTV.
If you still need help with stuttering from PMS to the LG, let us know.
If so, we could use a restart of PMS, wait 6min, express the failure on the LG, wait 30sec, then please use the web app to Settings - Troubleshooting - Download Logs
I thought it should have the LG playback since i just switched yesterday but sure i will try it tomorrow and upload new files. Thanks allot for the help!
Sir! Thank you for the timing and the new logs. This is what I saw the player and server decide:
Oct 16, 2019 20:20:03.252 [0x7f0003159700] DEBUG - MDE: The Art of Self-Defense (2019): selected subtitle cannot be converted to a compatible format, burning into video stream
So it wants to burn the PGS subs in that file, as they are image based subs. The better type of subs to have would be SRT. Anyway subtitle burns require the video to be transcoded on the CPU.
Could you try changing your Audio & Subtitle settings from Automatic it Manual?
It’s in Settings - Account. That should help you try to stop it.
I’m curious what these addresses are that it’s finding and why they have these names:
Oct 18, 2019 23:18:08.747 [0x7fc421a04740] DEBUG - Detected primary interface: 192.168.2.100
Oct 18, 2019 23:18:08.747 [0x7fc421a04740] DEBUG - Network interfaces:
Oct 18, 2019 23:18:08.747 [0x7fc421a04740] DEBUG - * 1 lo (127.0.0.1) (loopback: 1)
Oct 18, 2019 23:18:08.747 [0x7fc421a04740] DEBUG - * 6 bond0 (192.168.2.100) (loopback: 0)
Oct 18, 2019 23:18:08.747 [0x7fc421a04740] DEBUG - * 11 br0 (192.168.2.120) (loopback: 0)
I think my QNAP uses en0 and lo.
I didn’t see anything terrible in your logs. There were a a few errors, but not at the time you mentioned or nearby. It looks like it’s just doing a transcode and burning in subtitles.
br0 = first (and pre-created) virtual bridge adapter
QTS 4.3 and above use eth for the adapter and qvs for the virtual switch names.
I have the 4 onboard (eth0 - eth3), the 10 GbE at eth4. eth4 feeds the virtual switch and IP is attached to qvs0.
Had you not stated this is QNAP, I would have presumed Synology or other distribution.
To the matter at hand, given both bond0 and br0 are on the same subnet, default TCP packet handling is going to try and use both if the destination is also on 192.168.2.x
It does this because the configuration defaults to a metric (cost) of 1. Linux will always favor the path with the lowest cost. If cost is the same for two paths, it will assume both are valid and use both.
This could easily be the lag you’re seeing because it’s attempting to send on br0 which likely goes nowhere. Not being able to resend, it jumps back to bond0 and then ultimately does get sent.
Thanks for the findings. Not sure i understand every detail but to explain about my network setup i can share this :
Basically my NAS uses 2 ETH port in trunked mode (192.168.2.100).
My VM’s uses the other address 192.168.2.120 (as you say feeds my virtual switch). So Plex should always be using 192.168.2.100.
I never changed anything and it use to work before but maybe my network configuration is wrong to begin with. Appreciate some tips. I have QTS 4.4.1.
I have looked through the settings and cant see what i should change to be honest.