Why does my Plex server get the hiccups when the transcoding directory is close to 1GB?

General Discussions
1

Every time my Chromecast player starts pausing, I check the logs and everything’s fine, but the transcoding directory for the session always has between 900 and 1000mb of files.
I’m thinking there’s a problem somewhere, but I don’t know how to fix it, because the transcoding options just use seconds as a measuring unit.

3

But when I restart Plex, the old files are deleted automatically. Wouldn’t it be unable to delete them?

5

It looks like this actually helped. Can’t say for certain, since not all shows I watch generate 1gb of transcode files, but so far it’s been good!
Thanks!

Edit: Spoke too soon. Still have the same problem.

7

Could be anything, I suppose. I’ll do that. I presume you don’t want the debug logs?

9

Yeah, I don’t know what the difference is.
https://www.dropbox.com/s/v6bdo9lc3pd4l5w/Plex%20Media%20Server%20Logs_2017-08-29_15-43-36.zip?dl=1
I’ll delete the file once you reply.

10

I’m helping out Trumpy.

First, your Plex database needs some serious cleanup / maintenance. This will cause numerous problems with everything.

  1. Hover over the Libraries section to expose the ellipsis.
  2. Click it to open the drop down menu
  3. Click Optimize Database
  4. Next please go to Settings - Server - Scheduled Tasks.
Aug 28, 2017 10:06:29.128 [0x7fc0b9bff700] INFO - Library section 2 (TV Shows) will be updated because of a change in /data/tvshows/Game of Thrones/Season 07
Aug 28, 2017 10:06:32.734 [0x7fc0b9bff700] WARN - SLOW QUERY: It took 750.000000 ms to retrieve 6 items.
Aug 28, 2017 10:06:33.339 [0x7fc0b9bff700] WARN - SLOW QUERY: It took 280.000000 ms to retrieve 1 items.
Aug 28, 2017 10:06:33.570 [0x7fc0b9bff700] WARN - SLOW QUERY: It took 250.000000 ms to retrieve 1 items.
Aug 28, 2017 10:06:34.010 [0x7fc0b9bff700] WARN - SLOW QUERY: It took 470.000000 ms to retrieve 1 items.
Aug 28, 2017 10:06:34.553 [0x7fc0b9bff700] WARN - SLOW QUERY: It took 230.000000 ms to retrieve 6 items.
Aug 28, 2017 10:06:34.753 [0x7fc0b9bff700] WARN - SLOW QUERY: It took 250.000000 ms to retrieve 1 items.
Aug 28, 2017 10:06:35.557 [0x7fc0b9bff700] WARN - SLOW QUERY: It took 410.000000 ms to retrieve 1 items.

I’m also seeing numerous occurrences of

Aug 28, 2017 16:33:55.615 [0x7fc092ff9700] INFO - [Chromecast] Load: {"currentTime":0,"requestId":6,"autoplay":true,"media":{"metadata":{"metadataType":0},"contentId":"/library/metadata/63900","customData":{"offset":0,"server":{"address":"192-168-100-19.b71345ec827c443792a2aac8d2230bde.plex.direct","machineIdentifier":"7dc42af6930eaa00adc5ce21ff19a8245fd17d1c","isVerifiedHostname":true,"protocol":"https","transcoderVideo":true,"port":32400,"version":"1.8.1.4139","accessToken":"transient-0253a52f-ba41-4d3b-bf1a-a84701b9df71","transcoderAudio":true},"audioBoost":0,"resolution":"1920x1080","directPlay":true,"directStream":true,"subtitleSize":100,"containerKey":"/playQueues/5321?own=1&window=200","bitrate":20000,"quality":100,"user":{"username":"tofagerl"}},"contentType":"video","streamType":"BUFFERED","duration":4872.202}}
Aug 28, 2017 16:33:56.819 [0x7fc0917f6700] INFO - [Chromecast] [MDE] Augmented profile: { "directPlay": { "mp4": { "mimeType": "video/mp4; codecs=\"avc1.42E01E\"", "video": { "codecs": { "h264": {} }, "maxBitDepth": 8, "maxFrameRate": 60, "maxLevel": 42, "maxWidth": 1920, "maxHeight": 1080, "maxBitrate": 20000, "maxVideoResolution": "1080" }, "audio": { "codecs": { "aac": { "maxSamplingRate": 48000, "maxChannels": 2 }, "mp3": {} } } }, "mkv": { "mimeType": "video/webm", "video": { "codecs": { "h264": {}, "vp8": {} }, "maxBitDepth": 8, "maxFrameRate": 60, "maxLevel": 42, "maxWidth": 1920, "maxHeight": 1080, "maxBitrate": 20000, "maxVideoResolution": "1080" }, "audio": { "codecs": { "aac": { "maxSamplingRate": 48000, "maxChannels": 2 }, "mp3": {}, "vorbis": { "maxChannels": 2 } } } } }, "directStream": { "video": { "codecs": { "h264": {} }, "maxBitDepth": 8, "maxLevel": 42, "maxWidth": 1920, "maxHeight": 1080, "maxBitrate": 20000, "maxVideoResolution": "1080" }, "audio": { "codecs": { "aac": { "maxSamplingRate": 48000, "maxChannels": 2 }, "mp3": {}, "opus": {} }, "maxChannels": 6 } } }
Aug 28, 2017 16:33:56.831 [0x7fc093ffb700] ERROR - get - invalid frameRate value: 23.976
Aug 28, 2017 16:33:56.942 [0x7fc091ff7700] INFO - [Chromecast] [MDE] Starting analysis of 1080 (mkv, h264, eac3, 40, high)
Aug 28, 2017 16:33:57.246 [0x7fc0b1bfd700] WARN - [Chromecast] [MDE] Unable to validate videoResolution; re-analyze the media if possible
Aug 28, 2017 16:33:57.348 [0x7fc092ff9700] INFO - CodecManager: starting EAE at /tmp/pms-e37e79bd-1cfa-42fc-a12e-3504b34d4b43/EasyAudioEncoder
Aug 28, 2017 16:33:57.383 [0x7fc0907f4700] WARN - [Chromecast] [MDE] Unable to validate videoResolution; re-analyze the media if possible
Aug 28, 2017 16:33:57.656 [0x7fc0947fc700] INFO - [Chromecast] [MDE] Finished analysis of 1080 (mkv, h264, eac3, 40, hig

Take note of the frameRate. 23.976 is a valid framerate but, when PMS tries to work with it, it’s unable to validate it. This is telling me it sees something wrong.

At this point, A controlled test is needed:

  1. Select an item to play
  2. Get the XML (Hover over it, click ‘Get Info’, and then ‘View XML’)
  3. Play back the item (attempt it)
  4. When it fails, wait 15 seconds
  5. Go to Settings - Server - Help - Download Logs
  6. Attach the XML and the ZIP file it has given you here with your next post.

Also, in your next post, Can you tell me which distro you’re using (including version) and how much memory is installed in the system?

11

I think I got it right, but it was not easy getting the XML. In fact, I had to reboot the docker container.
XMLs: https://www.dropbox.com/s/fwk82o6lquk17ya/21102.xml?dl=1
https://www.dropbox.com/s/28wlkoej2e3jb98/21273.xml?dl=1
Logs: https://www.dropbox.com/s/c3aq6m1l5h6inl4/Plex%20Media%20Server%20Logs_2017-08-30_17-06-31.zip?dl=1

I run the Docker server in a VM on Freenas 11, but it has 8gb of ram, and plex is the only thing running slow. The Docker VM runs Ubuntu 16.04. The full server has 16gb, and the most I’ve seen the Docker image use is about 2gb. CPU stays low as well, including when this happens. If anything, it’s an IO thing, and I don’t really know how to find IO statistics.
Both config files and media files are served from the Freenas on NFS.

12

Now I know this is Docker, may I ask why you’re adding complexity and layers when it’s might not really be necessary?

DockerPlex is an abtraction of Ubuntu. If I’m understanding correctly; You’re essentially running Ubuntu in a VM (Docker) on FreeBSD (VM) on Ubuntu ???
If so, that’s a significant performance penalty (the I/O problem you speak of) but worse, you’re adding serious constraints to how everything works.

Please clarify for me?

  1. You have your files housed inside the FreeNAS layer?
  2. Which means they are also accessible at the base Ubuntu layer (it has the physical media) ?
13

What are you talking about? I’m running Ubuntu in a VM so I can run docker. There is no more abstraction, Docker is native in Linux.
There is no base ubuntu layer, I’m running FreeNAS.

14

Are you using Plex’s docker image? If so, you’re running another Ubuntu layer albeit thin… it’s a container… and an abstraction.

Would you share with me how you have everything layered? outline / line item form?

15

I’m using LinuxServer.io’s Plex image, but I really don’t think you understand Docker. There is no Ubuntu kernel running in the Plex image, it’s just running one process - and that’s Plex.

I simply have FreeNAS installed on my NAS, and in a VM I have Ubuntu installed so I don’t have to use FreeNAS’ horrible plugins. Before this, I was using FreeNAS Corrall, with a semi-native bhyve host running a FreeNAS-specific Docker Guest image. That was better, because I didn’t have to have a whole Ubuntu VM running, but it doesn’t explain this problem.

I think you’re getting sidetracked.

16

Thank you for qualifying who’s Plex image you’re running. Plex’s Docker image provides that basic abstraction and is a minimalist Ubuntu.

Returning to your OP of why does it hiccup at the 1GB boundary, Somewhere you most likely have a 2GB memory allocation OR some other 2 GB limit.
In Linux, 1/2 of memory is available to /tmp (tmpfs = ram)

I believe your most viable option is to modify the container and place /tmp on a fixed location on the HD which no longer subjects it to TMPFS limits

17

No, there’s a separate /transcoder volume. It’s on the same disk as everything else, but shared separately.