Closed ednxzu closed 3 years ago
Did you configure the Transcode path in Plex's configuration as mentioned in the readme? Meaning:
From the logs it seems that the Worker is indeed transcoding and also reporting progress back to Plex. My suspicion is that Plex is not finding or being able to read the transcoded segments which are written to the trascoding shared path.
Can you enable Plex logging in the configuration and monitor the output in Plex's Console? There should be errors or warnings there popping up.
Well I must say I'm really confused lol. Plex is logging that everything works fine when it just loads the thing to infinity.. the only error I'm getting is when I stop the transcode because plex tells me that it can't remove the transcode folder due to it not being empty..
That's odd! Have you confirmed that playback does indeed work when running without the clusterplex dockermod or using the standard linuxserver plex image? You can just comment out the dockermod or change the base image, everything else can stay the same.
If you try playing content in a browser with DevTools open, what sort of errors or network status codes occur when requesting segments?
It really is ! The playback works fine if no mods are on and even with the custom images if pms is handling transcode (transcode mode : both). however the network console is giving me some 404 on header for things that look like transcode chunks ??(see screenshot) and some other error cause I dont have a token but that shouldnt be an issue, the weird ones really are the 404 on headers.
and this is from my LAN, I don't get errors (they might be caused by cloudflare proxy then)
So from your LAN you get no errors but playback still doesn't begin?
I'd say be careful with proxying Plex traffic through CF, I believe proxying video is against their T&C and they tend to block accounts that do.
It's pretty puzzling, I mean, ultimately clusterplex doesn't really do anything in regards to playback, Plex should be picking up the segments from the shared path and streaming those.
Could you try the "Convert" functionality as a test? Meaning getting any random media and Optimizing it for another media type, just to see if that conversion indicates progress and finishes correctly? That one should use the remote transcoder as well.
From LAN I dont get any errors while playing transcoded content, but GET requests to the transcoded chunks just never post any status (see screenshot). From Internet (without cloudflare in between) I get header 404 and same playback problem where GET on chunks never give a result.
It's probbly the weirdest problem I've had in a long time lol I don't even have any idea what might be going on... everything looks fine, but yet it's not..
And direct play works fine
So weird! Now, those GET requests with no status have no info at all there in devtools? Could it be a permissions issue when PMS is attempting to read the chunk from /tmp? any chance to grant +777 to the source of the /tmp mount?
The /tmp is already 777. GET request have nothing except a header (following screenshot)
and a stack trace that doesnt tell much
I just noticed chunks are generated with 644 perms is that normal behaviour ?
Yup, that looks about right:
Have you tried connecting through https://app.plex.tv/ instead of the LAN IP? To see if there's any difference in how the player behaves.
In my Network section of Plex's configuration I have defined several settings, such as LAN Network with my real LAN subnets and a few docker subnets, as well as the Custom server access URLs (i've included my custom domain there, as well as the LAN ip url). Also in my network i've configured "List of IP addresses and networks that are allowed without auth" to the following (my LAN and docker network subnets): 192.168.0.0/16,10.0.0.0/16,10.255.0.0/16,172.18.0.0/24 I haven't documented those since I believe they aren't mandatory for it to work, but reviewing the config might be worth a try.
See https://github.com/pabloromeo/clusterplex/issues/31#issuecomment-717208287 to see if this user's solution works.
I kept trying to figure out what I was missing but I can't for the sake of me, make it work. I'm just going to give up for now. You can close this ticket or leave it open if someone's facing a similar issue and wants to share it here.
Best,
Sorry to hear about that! I'll be closing it for now, but will use it as reference in case someone is having a similar issue. Thanks!
Describe the bug When remote transcoding, everything seems to be running fine, but the actual pms never plays back content even tho worrkers seem to transcode properly.
To Reproduce Brand new install with recommended setup, and NFS share mounted across all swarm nodes
Desktop (please complete the following information):
Additional context When starting a playback that is supposed to be remotely transcoded, web player loads to infinity, while worker is transcoding and orchestrator waits on
in progress
status :LOGS FROM PMS :
LOGS FROM ORCHESTRATOR :
LOGS FROM WORKER :
And of course the stack file
All this time the CPU load crearly shows that the worker is in fact transcoding. Thanks for everything so far this project is insanely interesting !!!