If you edit Preferences.xml when PMS is shut down, and the syntax is correct, it will persist.
LogNumFiles=“20”
Ah… Thanks. I think I made the change and then stopped/started the service. I’ll keep that in mind for next time.
I have already reverted back to 1.14. Please let me know if the data I provided is insufficient, and I will upgrade again and collect more data.
My xmltv file is created by epg123.
xmltv.zip (14.6 MB)
Also, just a little about my setup. I have two tuners in plex. One is running through xteve and the other is from my hdhomerun quatro. Both are using this xmltv from xteve.
Thanks for the snapshots and logs. I could only use the 4th-5th May logs as the 7th May logs were for time when memory hardly changed - only increased from 3.7Gb to 3.9Gb.
But, the evidence from the first set of logs and snapshots for the 4th and 5th May are fairly conclusive it is the same DVR EPG issue - leaking about 200-300 Mb of memory every 4 hours
There was only 1 refresh that did not lead to a memory increase - all others did
I will add this to the existing evidence and the advice is the same. Schedule periodic restarts of the server whilst the issue remains
Seeing the exact same problem as of a few releases ago. Was working fine under a 1.5GB VM, then expanded to 2GB, now 4GB and the OOM kicks in every few days. Also using DVR.
I’m seeing a similar issue. 32gb of ram and it’s almost full within 2 days. Is everyone else importing a tv xml? I’m wondering if it’s that.
It looks like it might be the xml file being updated. I am using my own xml file from schedules direct. I have to reboot daily to make sure everything works fine. If I let the memroy build too much then I experience with my DVR. If the memory is full then my recordings will be the wrong channel. As long as I restart daily, I don’t have any issues with my recordings.
i can also confirm its related to xmltv epg data …
my experiences are pointed here, actually it came more or less after the fix for the file usages in /tmp …
now no more docker size ballooing, therefore RAM 
both tests with same tuner setup, so only difference is the guide source …
of course the PMS guide was not as complete (cause only 1 source allowed) but i left it running for some days and Ram did behave normally.
I see in the release announcements that there were several bug fixes starting in 1.16 for server crashes and memory issues. Does that include the issue that was discussed in this thread?
Well, it appears that the memory utilization issue is still present in 1.16. I am in the process of collecting logs using the methodology above and will start a new thread once all the data has been collected.
I’m still having memory issues as well. Think I will do the same and provide some more info.
I had and interesting phenomenon happen when I went on vacation for a week. I came after letting Plex fun for about 7 days straight and the memory wasn’t as full as usual. I usually reset every 24 hours to keep things working correctly. If I go beyond 24 hours I experience messed up recordings.
Maybe I will try to let it go for a week and see what happens. Just hope I can contain myself and not reset it when I start noticing abnormalities. We shall see how it goes.
@sa2000 @ChuckPa
Here is an update to my current memory issues. I let it run for a week to see how it would handle. It looks like the memory would completely fill and then reset and slowly start filling again. Attached are my logs as previously collected.
Plex Version 1.16.1.1273
OS: Ubuntu 18.04.2 LTS
Logs:
Plex Media Server Logs_2019-07-06_12-34-26.zip (12.8 MB)
mem_snapshots.log (899.8 KB)
This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.
