#Trying to create a Spot GPU instance leads to 400 response error

8 messages · Page 1 of 1 (latest)

deep ivy
#

Greetings,
I've recently noticed that whenever I try to create a GPU spot instance, I'm getting a 400 responses error.

I'm trying to spin up a spot instance using the RunPod PyTorch 2.2.10 image, with 2x A6000, a network drive, in the Sweden datacenter.

The payload for the request:

{
    "operationName": "BidPod",
    "variables": {
        "input": {
            "bidPerGpu": 0.49,
            "cloudType": "SECURE",
            "containerDiskInGb": 20,
            "volumeInGb": 0,
            "dataCenterId": "EU-SE-1",
            "gpuCount": 2,
            "gpuTypeId": "NVIDIA RTX A6000",
            "minMemoryInGb": 100,
            "minVcpuCount": 18,
            "networkVolumeId": "XXX",
            "startJupyter": false,
            "startSsh": true,
            "templateId": "runpod-torch-v220",
            "volumeKey": null,
            "ports": "8888/http,22/tcp",
            "compliance": null
        }
    },
    "query": "mutation BidPod($input: PodRentInterruptableInput!) {\n  podRentInterruptable(input: $input) {\n    id\n    machineId\n    __typename\n  }\n}"
}
vivaldi_i4CX0dxwzl.png
olive shoalBOT
#

To help others find answers, you can mark your question as solved via Right click solution message -> Apps -> ✅ Mark Solution

sinful scroll
#

looks like error in query

deep ivy
#

Tried it across various devices, all responding the same :(

#

uhhh

#

It works now again? Was there some change?

ancient swallow
#

Yeah we pushed a fix just a little bit ago

olive shoalBOT