R
RunPod•4mo ago
MushyPotato

Pod Downsized, with Pictures

I have been having issues with the A6000's, my pods keep getting downsized. I wasn't able to catch a snip of the first couple times but I did here. As you can see I am renting 2 A6000's but when I try to start it I am only renting 1. What is happening here? this keeps on happening so I have to terminate and start new pods repeatedly, my entire audit logs is me restarting new A6000's and I keep losing money trying to get the service to give me what I bought. (ID:wp0ofb7xs94uf0)
No description
No description
Solution:
Hi @MushyPotato - when you rent a pod on spot and it gets stopped (really, when a pod gets stopped for any reason) the GPUs in that machine are made available for other customers, and by the time you can start it again there may be fewer GPUs available. Setting up a network volume will allow you to deploy to any GPU within that data center, so you will not be limited by whatever the GPU rental status in your specific machine is....
Jump to solution
4 Replies
MushyPotato
MushyPotato•4mo ago
I will not terminate this one and I will let it go idle , here is a extra snip for further proof
No description
Solution
Finley
Finley•4mo ago
Hi @MushyPotato - when you rent a pod on spot and it gets stopped (really, when a pod gets stopped for any reason) the GPUs in that machine are made available for other customers, and by the time you can start it again there may be fewer GPUs available. Setting up a network volume will allow you to deploy to any GPU within that data center, so you will not be limited by whatever the GPU rental status in your specific machine is.
MushyPotato
MushyPotato•4mo ago
I understand
Madiator2011
Madiator2011•4mo ago
The machine that host your pod only have 1 free gpu
Want results from more Discord servers?
Add your server
More Posts
I'm pretty sure I've been getting pods where "/" lives on a network diskWhich makes data reading impossible and the pod useless. I have a little 'see hQuestion about graphql APIIn https://doc.runpod.io/recipes/view-gpu-types-info If securePrice is zero does that mean that the Create new pod with runpodctlI'm trying to create a pod with runpodctl. It appears by reading the --help that I cannot create a pCommunity cloud servers repeatedly fail to correctly download containersA100 PCIe 80GB servers repeatedly fail to download parts of containers. Easiest template to reproducUrgent: All new gpu pods are brokenHi, our existing pods and new pods we are creating are having all same issue where they cannot find CPU Pods NOT WORKINGI'm having issues with my template that worked for months and also with your template: `runpod/base:Streaming is not quite workingHi I was trying to deploy a simple huggingface model where my inference code looks like thisKnowing Which Machine The Endpoint UsedHi, I configured my endpoint to opt for multiple types of gpu like the image attached. When I run a how to know about serverless api region informationI want to get region where the serverless api deployed in the rp_handler.py for upload file to choosGPU usage when pod initialized. Not able to clear.Tried nvidia-smi -r, restarting, and reseting. There is still usage on one gpu in the pod.