Server Version#: 18.04.1 LTS
Player Version#: 1.13.8.5395
My physical server is a Dell r710 with dual Hex 3.06GHz processors. I have an esxi hosting my VMs. This VM has 32GB and 8 cores dedicated. I have NFS shares off of a Synology 1815+. My NFS with the media is read only, I have separate NFS set as read/write for trans-coding and converting stuff if needed. I removed cloud-init and added tools for NFS, PowerShell. VMTools and some other minor changes, but other than that, everything is bare-bones.
The problem that I’m facing is that all my RAM ends up being used, I started with 8GB, upped it to 16GB and now 32GB. I keep reading that this is just a false-positive, but it is not a false-positive in my case. Videos streaming start to lag (watching through my xbox One S), I pause, then play again and it catches up, after an hour of doing this, my VM server crashes (I have to hard reset, the terminal becomes unresponsive).
This is my second build; I experienced similar RAM issues with Ubuntu desktop, but that was limited to when I was scanning the library files (not even a full refresh on metadata). I had the same issue with the previous Plex build (I don’t remember the version number). I never experienced these issues when I used the Synology Plex app. The reason that I created the VMs is that I wanted to offload the service from the Synology because it has trouble playing 10-bit media as well as 4k.
Does anyone know of a quick-fix for this? I’ve seen many posts about RAM usage but the answers always indicate that Linux doesn’t use up RAM like that, but a close friend of mine is going through the same issues and ended up abandoning Linux and building out a Windows Server VM because of all of the frustration. I don’t want to go through a third build (as well as licensing costs).
This is my first time ever posting on a Forum anywhere, so I hope I didn’t breach forum etiquette. ChuckPA tutorials have been wonderful, hopefully this is an easy fix for a Linux buff.
One last caveat, I didn’t restart after I updated to the newest Plex (which was earlier today), but my VM crashed in a similar fashion last night on the previous Plex, and the VM had been restarted yesterday afternoon.
Might help to stick something like https://my-netdata.io on the box to monitor when things go south. Never heard of RAM issues like this myself… got a decent size library and I’m running Plex via a docker on Linux.
Normal Linux operating philosophy is to find a way to use all installed memory in some way.
What confuses most folks is “free memory”. It looks like all the memory has been used when, in fact, Linux has used silently keeping everything it reads from & writes to disk also in RAM (like a SSD R/W cache). This way, if it’s needed again in the not-to-distant future, it’s already in memory.
Reclaiming Cache memory is a zero-cost/delay step.
My desktop has 16GB of RAM installed.
The graphical tool shows: Program 30%, Buffer: 0%, Cache:70%
Here is what my status bar shows.
All of memory is in use in some way, which makes sense… Find some way of using memory to improve performance.
If I may ask why Docker when,
Linux direct portability exceeds docker by any and all measurement standards
All the complications of docker are removed
HW transcoding is directly available without need of hacking inside the container.
Windows will show you the 30% number and report the rest (70%) as free. It is performing the same caching silently in the background. Linux is showing you what’s actually going on without filter.
A Linux VM is no different than the real host. Linux in the VM will behave the exact same.
Care must only be given to make certain enough base memory has been allowed for Linux to function without swapping.
I’ve monitored the RAM on the host and the client; I understand that vmware gives a “false-positive” as far as actual usage but within the VM I was getting results that showed my memory usage was pretty high… I wasn’t monitoring it when it crashed though so I’m not sure the direct cause.
I just had another crash and now the performance summary isn’t even showing over 11GB, so there is definitely something else going on
I really don’t want to build out a third VM, especially if I’m just going to have the same results that I’ve been having. I never had these issues on my Synology (although that is pretty idiot-proof).
Strange, I looked into that log the other day, and it hadn’t updated since the 4th. My system crashed again some time last night, looked in that log and it doesn’t have anything new either. I may have to look in a different log somewhere. I’m getting the feeling that it may be the VM tools causing a crash, but I haven’t been able to substantiate that hypothesis.
Well, this is strange. For some reason Plex added "/media/yaseen/Miso Hairy Dove/" to almost a full season. I tried to refresh the metadata but it still says this. Maybe Linux isn't my thing, I may need to stick with what I have all my certifications in... Microsoft. hahaha
