Subject: Problems NVDEC Hardware Video Decoding Access
I'm running a video processing pipeline for sports analytics on RTX 4090 pods and need hardware-accelerated video decoding (NVDEC) for performance reasons.
TECHNICAL DETAILS:
- Pod ID: y50n3mu0w6z3nr
- GPU: RTX 4090
- Current Issue: Getting "cuvidGetDecoderCaps error 100" when trying to use NVDEC...
Pulling an Image for H100 SXM extremely Slow
We're trying to spin up a container id dnv1q1479mtn93 Pulling a 10gb image is taking us more than 30 minutes currently in H100 SXM (The total image size is 60gb). This is really odd trying to pull an image for a powerful machine like this. We're trying to pull the image https://hub.docker.com/repository/docker/aiorbust/musubi-tuner-h100-v15/general. I feel like the download speed when loading the container is always irregular. Can someone help us take a look at why the speed is so slow?
Hi, im having a compatibility issiue with nvidia drivers. Is there any way to set the correct driver
Getting this error when running the code on a 5090, but on the 4090 it works. Any way to set the correct driver? Or is there another factor
Traceback (most recent call last):
File "/app/ai-engine/matchlytics/main.py", line 1276, in <module>
main(
File "/app/main.py", line 786, in main...
cant download models from hugginface to runpod
i have this error
UDP port for custom TURN server
Hello everyone, I'd like to ask how to expose a UDP port in the RUNPOD VM. I am trying to create a WEBRTC server and a custom TURN server. But I found no documentation for exposing UDP ports right now I am using TCP ports for that, but I am not getting the performance as what i was getting in local env, I am not getting any help from online documentation. If anyone can clarify this. as what I am missing here.
Training Looks Like a Blob? (Pls Help) :(
Hi Friends! So I'm new to all this and maybe doing something wrong (been working with ChatGPT to try and figure things out) - it's helped me get this far. I managed to do sequential training on Kohya starting with 30 face images then 40 body images (all captioned with matching titles) and using the RevAnimate model and a VAE - then saved this as a .safetensors file. Then, I went to A1111 to generate images of her (using her .safetensors model) and see how it turned out with the hopes of generati...
There are no GPUs or CPUs available in US-IL-1, why not?
There are no GPUs or CPUs available in US-IL-1, why not? It seems like the entire region is down... hard to believe even no CPUs are available?
Solution:
It was a cookie issue. Runpod was storing something there that caused incorrect pod availabilities
Pod load up speed is way too slow
I have a container booting up: t3uqnnamr1zctt. But it's taking almost an hour to load the container. How can I increase the container loading speed / are there settings I am missing here?
I can't connect to my ssh on runpod
I open up command prompt and paste in the command it told me
just to get this
C:\Windows\System32>ssh wosut9qqgw53s5-64411afd@ssh.runpod.io -i ~/.ssh/id_ed25519 Warning: Identity file C:\Users\Berb/.ssh/id_ed25519 not accessible: No such file or directory.
wosut9qqgw53s5-64411afd@ssh.runpod.io: Permission denied (publickey)....
wosut9qqgw53s5-64411afd@ssh.runpod.io: Permission denied (publickey)....
Bug in runpodctl
I add runpodctl send . from a directory I want to get all files. It was successful, then pasted runpodctl receive <code> from the directory I want the folder to go into. It got sent and unzipped but could not find the folder.
This is on my MAC. I have done all sorts of
find ~and find / with file name patterns but could not find the files/folder based on the names/name respectively....How do I download a folder to my computer using runpodctl
I'm trying to do runpodctl send /workspace/madapps/SwarmUI/Output/local/raw/2025-09-10 but its just not working, give me a blank message then kicks me back to make another command
Wrong Templates
Something is not working right with the pod templates in the last few days.
I choose a pod template, then I choose my network volume, then GPU and I deploy the pod. And then it does weird things - it begins to pull wrong template, not the one I chose. I need to do everything again, sometimes two times, until I get it to work and it pulls the template I chose....
Can't deploy pod with RTX 6000 Ada
I'm trying to deploy pod with RTX 6000 Ada for several hours, but I can't. But, I can deploy pod with same setting with L40s.
- NV: US-IL-1
- Volume ID: xstaqifwh3
- Template ID: jj1k6pxd6e...
Upload speed
Almost 100% of the time, when trying to upload files, the upload speed is below 50mb/s, very often below 20, however I sometimes must setup a new pod and upload datasets up to 100GB, hox am I supposed to have a faster internet for my pods? i do not want to both wait and pay 10 hours of GPUs just to upload a file, what can I do?
When encoding video with ffmpeg, nvenc does not work.
DC:US-NC-1
GPU:RTX 5090
I have switched data centers to US-IL-1 in addition to US-NC-1, but the results remain the same.
...
pod taking too long to load up
i just opened a new network volume and ive already deployed 2 pods with different computers each and they're taking so long just to even start (stopped them at 15m) any reason why?
downloading containers takes ages
I dont know how in other regions but in EU-RO-1 it takes even an hour today to download a container from dockerhub, I restarted and created completly new pods, no difference
Rest API 500s error
Anyone else experiencing issues with the REST API? Suddenly getting 500s while I haven't touched the code or pod in a while
Solution:
This was fixed, sorry.
Change GPU on stopped POD.
Hello, I have tried to find a way to change gpu on stopped pod bun I'm not sure if its possible. I would like to adjust gpu depend on the case without creating new pod.
Thak you....
8 x RTX 5090 PODs JupyterLab and Terminal not working
I'm having almost constant issue when I try to use 8 RTX 5090s. The POD starts after 6 to 8 minutes only to find that neither JupyterLab nor the terminal are working, either showing a 404 error or an empty white page. I tried to stop and restart the POD but that didn't help. I have to terminate and try again until I get lucky and the terminal works, and 1 out of 10 times the JupyterLab may work. That is something I had to live with which is annoying but doable. But my problem is I'm being charge...