I’ve given up, I’m in the process of moving to Emby. 
I am running 18.04 Plex 1.18.2.2058 and it is using 35-40% of the RAM.
I am running another box (my old one that is still running but doing nothing, absolutely nothing–other than waiting to be shut down and archived) running 14.04 Plex 1.18.2.2029 it is using 5% of the RAM.
DLNA is disabled (and was never enabled) so that is not the issue.
These are identical boxes (hardware) running ONLY Plex and Samba (for file sharing).
Very odd…
I’m also getting this issue, from 1.6 GB to 7 GB in minutes then again later to 12 and back settling at 10 GB of RAM. What is weird is that during this time at 0200 AM everyone is asleep so it’s doing this by its own.
I don’t have remote access enabled only accessible from local network, nor DLNA and the library is small, around 100 items in Movies and TS.
2a is when daily maintenance tasks are run by Plex.
EDIT: This support article explains more:
Good to know. What about 3:45ish where it spiked to 12 GB? Another task?
But even when daily maintenance task is run, I understand using the mem but also should release it when it’s done, no?
I’m having a memory leak issue as well since the last two updates (1.18.2.2058 and 1.18.3.2156). I’m running PMS on a Synology DS918+. Stopping and starting the PMS service solved this on .1.18.2.2058 But after a time it gets back up to above 90% of the total memory.
I’m running 1.18.3.2156 now. And I just stopped the Plex services (which was using above 3Gb of memory - I could see that in the task manager). But for some reason the memory is not freed when stopping the service. Under the processes tab I can see one Plex service still active, and lots of others in sleep mode?
I tried stopping the service “pkgctl-Plex Media Server” with SSH but apparently it was already disabled and stopped.
I had to reboot my Synology.
B.t.w. after rebooting I checked whether DLNA was turned on. The service was not (checked) (but DLNA server time line checks - translated from dutch) was.
Is anybody at Plex looking into this??? This is getting pretty annoying.
I just checked again. And the Synology is now up to 74, just after appr. 20 hours. The 1.18.3.2156 seems to have made it worse!
same issue here, After a few hours plex dlna uses up all of my memory and is getting killed .plexmediaserver_1.18.3.2156-349e9837e_amd64.deb
And now it’s up to 93%. I would really appreciate it if somebody from Plex would reply to this thread!!!
Also - all the other threads regarding memory leaks have been closed automatically.
I’ve been using Plex for a couple of years now and I’m a loyal life time Plex Pass user. I’ve never had problems like this before. And this is a really high critical issue, because the problem is on the current version only fixable by a hard reboot of the NAS, and then only for a short period of time. I don’t appreciate having to reboot my NAS every 24 hours.
I don’t know how it works on your NAS, but I can regain the memory by restarting the docker service, no need to reboot. But I’ve given up now anyway.
Perenor,
Try stopping the DLNA service in Plex. Under Settings, DLNA. Uncheck “Enable the DLNA Server”. On my QNAP NAS this immediately reduced memory usage by Plex to less than 500MB and I no longer get the memory leak issue.
I’d like this fixed too as I use DLNA on some of my devices but it appears the Plex Support team are ignoring the problem even though it’s been around since version 1.15.x. Apparently on new installs of Plex with the current version, DLNA is disabled by default, so rather than fix it they’ve just turned it off.
I was considering buying a Plex pass myself but given the bad support on this issue I’m not sure I want to waste my money.
Cheers,
Frank
That worked with the previous version. Not on the latest. I tried to stop the service yesterday, but iin the services tab of the task manager it still showed as being active. I also saw a lot of instances of the Plex Server service in sleep mode. When checking in a SSH session it showed as disabled/stopped.
DLNA is actually disabled in Plex Server on my Synology. I checked this yesterday. after a server reboot. And still it got up to 93% in approximately 24 hours. “DLNA server timeline reporting” was enabled though. So I disabled that as well.
I rebooted again this morning. Lets see what happens.
I went back to the most stable version i know of. No memory leak with this version.
I run Plex Server on an old Windows 2008 R2 box. The machine started hanging if a film was being watched locally and I didn’t know what was causing it. Eventually I realised that it was Plex causing the issue. I went right back to 1.13.5 and the PC seems stable again now. Think I’ll leave it on this old version for a while.
Any news on this?
You can’t just go back to an older version. The synology NAS doesn’t allow that. And I don’t feel to excited about deinstalling Plex and reinstalling it and having to configure everything from scratch.
Are you running native or docker? I am running native and I don’t see this issue. I have tons of ram in my synology units (16GB in my real (backup) unit and 64GB in my xpenology box)
my whole xpenology box is sitting at 1.4GB reserved, 2.6GB used, 800MB buffer, 55GB cached, and 4.2GB free, and thats with 5 docker containers running (not plex in docker thou)
in resource manager, under task manager, services plex media server is only using 585 meg
sure under processes there are 38 sleeping and 1 running processes running for Plex Media server, and about 30 sleeping processes of plex Tuner service, but these sleepers don’t eat real ram (each PMS process i have lists the same exact 482.2MB and those alone would add up to ~ 18GB, and from my first number you can see I am not using more than 5 GB for everything (cached is just consumed when ram is not used for anything more useful and not plex related))
runtime on NAS is 14 days and plex has been running from 2019-12-17.
I’m also experiencing this problem. I run plex on docker with the official image plexinc/pms-docker:1.18.3.2156-349e9837e, my system has 48G of memory, and i don’t have DLNA enabled. I’m limiting the maximum memory allowed with docker as:
mem_limit: 4G
and even though docker stats is showing it as limited
CONTAINER ID NAME CPU % MEM USAGE / LIMIT MEM % NET I/O BLOCK I/O PIDS
60d21be95d18 pms_plex_1 15.87% 551.4MiB / 4GiB 13.46% 2.91MB / 403MB 48GB / 90.1kB 121
as soon as the Plex Media Agent is “alive”, all my memory is used and it will not go back to normal. The strange part is as the memory is completely full i will run drop_cache (on linux ubuntu 18.04 server LTS)
sync && echo 3 | tee /proc/sys/vm/drop_caches
and the memory will clear, just to be fully used again. this will happen untill Plex Media Agent is running. As soon it ends doing whatever it have to do, the drop_caches will clean it for good and i will stay at memory usage it is intended to.
As this is a very old problem and i think is not getting the deserved attention, i will stick with a minute cronjob to clean the memory cache. It will be a hard hit on overall performance, but i wont get an unresponsive system or get memory starvation.
Please, give this the attentions it needs.
edit:
when i’ve mentioned that the memory is completely used, i meant that it surpasses the limit of docker that i setup. it will grow untill the complete physical memory available to the system 48G…
Same issue running native plex package on synology. DLNA service eats up memory and may be causing kernal panic (not confirmed - still testing with DLNA service disabled). With DLNA disabled, memory seems stable enough. I’m using 1.18.4.2171 on an older synology - DS1512+
I’m also experiencing this. Running native on Windows Server 2012 R2. I have 32GB RAM and it’s getting almost completely filled over the course of a couple of days. DLNA is disabled.
Version 1.18.4.2171
