I’ve been a Plex user pretty much since the beginning and love it! I have a substantial library which I have shared out to family plus use Ombi & PlexPy alongside it. It’s worked great for the years I’ve had it - until an issue starting occurring this week.
The Issue
Seemingly randomly, my PC becomes unresponsive and I’m only able to even move the mouse every few seconds. Checking Task Manager, I see numerous PlexTranscoder processes maxing out the CPU. What I can do during this time is limited, but I have been able to confirm on multiple occasions when it occurs that there’s 1-2 things playing locally - if any at all. When preparing to type this, I opened task manager and saw many processes running even though the CPU isn’t maxed (screenshot below) and no one is watching anything. Closing PMS stops the processes as does obviously restarting the PC. This occurs 2-3 times a day.
I know that other tasks than watching media can cause a transcode, but have verified that video previews and thumbnails are set to “Never” in both the server settings and per library. One interesting thing I have noticed is that the titles being transcoded are seemingly all ones I’ve recently watched within the past week, though.
Below is what I’ve tried so far to remediate / locate the issue - none of which made a difference:
Installing latest PMS update
Viewing logs briefly - didn’t see anything that stood out
Disabling Ombi & PlexPy
Setting max transcodes to “3”
Disabling Process Lasso which I recently installed and configured
Verified no thumbnails, video previews, or optimizations where enabled (none ever have been)
Screenshots & environmental specs are below - Any suggestions? Thanks!
Environment
PMS installed on Windows 10 x64 (3770k CPU / 32gb RAM / PMS installed on SSD, data files on hybrid drive, media on 42tb disk pool)
PMS Version 1.12.0.4829
Transcode Settings (See screenshot below - If gaming, I will disable feature to use hardware acceleration temporarily - have GTX 970)
Most plays are on LAN using x2 Raspberry Pis and 2 Rokus but also use mobile devices. Most local plays on main TV use Kodi & SMB
Screenshots
Multiple PlexTranscoder processes despite max setting of 3 and nothing actively playing. All are titles I’ve played to some extent in past week. CPU usage is ok here, but usually when this many it’s maxed out
Thanks for the reply…
I’m currently on 1.12.0.4829 as mentioned in post, but it started a few days ago on whichever (PlexPass) version came prior to it. I’m pretty good about keeping PMS up to date so would assume it’s the one that was released immediately prior.
No - it requires authentication. I also monitor things in PlexPy and in Activity view can see what’s being played. All things it’s transcoding are titles I’ve recently watched too which would be a coincidence. I guess I could get off beta, but in the meantime; I wrote a small PowerShell script as a temporary solution. It basically looks to see if there are more than 5 PlexTranscoder.exe processes and if so restarts PMS. I’ve got it running every 15 minutes via Task Scheduler and have linked it below if anyone else is interested.
I realize you’re only using a 3770k (and this isn’t related to the topic) but curious as to why you have your preset set to Ultra Fast. Even with a 5 generation old CPU, I would think you could do Very Fast…
Do you not notice larger transcoded transfer files or slightly lower quality, especially for non direct play formats?
Any news or ideas on this? I’m still experiencing the issue as of writing on 1.12.0.4829. Instead of using the script posted above, I set Process Lasso to only allow 4 concurrent PlexTranscoder.exe processes which helped a lot even though wasn’t a fix. I temporarily disabled this rule this morning to see if the issue persisted and within an hour had about a dozen PlexTranscoder processes despite nothing being streamed anywhere (when checking from Activity tab of web console).
After the issue started, I changed my update channel to only use Public releases, but when checking this morning and seeing the version I used was the current public release, I decided to upgrade to the latest Plex Pass / beta of 1.12.1.4871 to see if it resolves the issue I’ll update once I am sure either way.
Meller86 - Sorry for the delayed reply…
It’s simply set that way as part of the troubleshooting for this issue. I also haven’t monitored the size of the transcoded files so can’t speak to that. Quality seems the same to me on all clients I use, too.
Well… that was fast - I still have the issue after upgrading to 1.12.1.4871.
Shortly after upgrading, I restarted the PC (unrelated) and when working on something else, noticed the PC was sluggish. I looked in task manager to find 5 PlexTranscoder.exe processes - one of which was using ~80% CPU. For any potential doubters out there, I’ve also included a screenshot of “Now Playing” to show nothing was being played, synced, etc at the time this occurred. Once I closedPMS, everything went back to normal. Otherwise, the number of these processes would have grown a bit more.
Can someone please advise on what this is and how to resolve it!?!
It was mentioned in an earlier reply that a similar issue was reported ~5 releases ago. What it a confliction with something I can resolve like uninstalling a specific Windows 10 update?
Happy to submit any specific logs if needed.
Thanks!
Screenshots Personal information and the server name are blurred out on purpose
Task Manager window on top of PMS web interface showing installed version
Which log file(s)? As I’m sure you know, there are well over 50. I don’t have the option enabled to include the token, but other than that; will they be safe to post?
Also, as mentioned previously; any process would require the transcoder has not been started manually, via library settings, or PMS server settings. When examining the path of PlexTranscoder.exe in task manager, I see it’s processing titles I’ve recently watched; but that I’m 100% for sure they or other things are not being played or synced at the time as shown in my screenshot from the 16th.
Overall, the next time I encounter the issue, I’ll compress all log files prior to closing PMS (so their state is saved) then post once you tell me which to send. If any debug option needs to be enabled beforehand, please let me know.
OK - after posting, I was looking at the log files more in depth and it seems that they rollover whenever PMS is stopped and started again. This morning at 10:33am, I noticed the issue so closed PMS via system tray then relaunched it again to resolve the issue.
Based on that, attached is a zip of a pair of each primary log file. I had to close PMS again to zip them all. The notes for the files are:
Logs ending in a “1” are from when I closed PMS (via system tray) when experiencing the issue where there were multiple PlexTranscoder.exe processes using most of the CPU and I verified nothing was playing, syncing, etc. I closed PMS at approximately 10:33am this time
Logs not ending in a “1” are from when everything was working correctly and I closed PMS so I could zip up the logs prior to posting
Important Note:When these logs were recording, I had a rule enabled in Process Lasso which restricts the # of PlexTranscoder.exe processes to 4. This can be the cause of some of the errors in the logs, but not the issue as I put it in to help control the issue. I’m removing it now for troubleshooting. Yes, I know this can be restricted within PMS settings, but this setting did not affect the issue so had to use a 3rd party app.
I just relaunched PMS again so if I need to do anything different the next time it occurs, please let me know. Shouldn’t be too long…
I’ve briefly reviewed them using CMTrace (I’m an SCCM Admin) and see many errors existing in the ones when the issue occurs and no errors in the more recent ones when all was working correctly. Some errors seem to point to certain settings so please review my initial post to see the variety of settings I’ve tried with nothing making a difference.
Your 1 logs show that there were 4 transcoding jobs going on at the time you stopped PMS. Unfortunately, they were in the middle of the process so I can’t tell what triggered them. I’ll need older logs back to the beginning when they started to see what caused it.
Mar 20, 2018 10:19:59.316 [17136] DEBUG - Request: [127.0.0.1:1318 (Loopback)] PUT /video/:/transcode/session/60b0c0c1-b29a-40f3-846e-dd2f8ae34164/be5c94d1-74e7-444e-a4bb-25a2ba77888e/progress?progress=40.1&size=-22&remaining=27425&vdec_packets=25459&vdec_hw_ok=25449&vdec_hw_status=1 (26 live) Signed-in Token (bzowk)
See the progress=40.1, that’s the progress in percents Can you check the older plex media server.log.x files and see if you can find when this was 0 and provide me that file. I need to see about 1 minute before this time, so if it’s at the very beginning of a file, I’ll need the one before it.
Also, someone using your account started watching something at 10:30.
Haunted Collector - S2 E6 - Stirring the Dead & Ghost Writer
If none of this was you, then someone may have access to your account. I would suggest changing your password and resetting all device tokens to disable all clients that have your old credentials.
I can guarantee you that at 10:30, nothing was being played. Not only was no activity listed in PMS settings, but I own all devices and am not playing anything. In fact, that title hasn’t ever been selected to be played - at least not recently.
The issue just occurred again so this time I left PMS running and copied the log files (so i could access open files) then zipped it up and attached it. Below are screenshots I took at the same time showing nothing is being played, synced, etc. i also walked around to each TV/device with access locally and verified nothing was playing. Finally, I monitor bandwidth using a variety of tools including DU Meter on my desktop. As you can see below, despite items seemingly being played, the LAN upload bandwidth doesn’t reflect that. Internet / WAN bandwidth during this time is less than 10% of that hovering at ~4kb/s so it’s not streaming online either.
If you look at the requests coming in, you see something like this:
Interesting bits:
Tons of requests.
All of them HEAD
My suspicion is that you have something running on 192.168.0.50 which is essentially “scanning” the PMS. I would usually suspect DLNA, but the logs are minimal, so it doesn’t appear to be that.
I notice in your initial post you have Ombi installed. Is it possible you that this is Ombi scanning the library to cache episode and movie information?
I have been unable to install Ombi connected to my PMS as when it tries to do this it kills my (1300 movie, 18000 episode server) through I believe a similar issue. Note I have not monitored much when this happens, simply killed the Ombi docker I had.
Many differences on my setup as my PMS is in unRAID as a docker (so on Linux), server is HP with Xeon E3-1220v3, etc…
192.168.0.50 is a Linux VM I have which basically runs Kodi and keeps my mySQL database up to date as it performs an automatic update each hour. Even though it scans the same files, there’s nothing Plex related installed on it - including no Plex Kodi plugins. All files are accessed via SMB - not DLNA or Plex in any way. I still don’t see how it would be at fault for transcoding multiple titles, though; as I’ve never played anything on it before as it acts somewhat headless.
I’ve got Chromium installed on that PC, but haven’t ever played anything from it. I’ll remove all devices associated with it - assuming there are any - from my config tonight just in case.
As far as Ombi, I thought of that and it’s installed locally; but stopped its service (verifying process was gone) and the issue still occurred.
