[Provider Audit]: provider.1legion.cloud

[dvargasracero]

Provider information

Provider owner address (akash1…):
akash1xgwhu8s7umxcqn7905ellqxlvqgt4yzatk3f46

Provider domain:
provider.1legion.cloud

HTTPS URL of the audit deployment (ingress on port 80 — share the link auditors should open):
http://mjhogr7jlpajv11uiv62vqiatg.ingress.1legion.cloud

Note:
We will keep the audit deployment running until an auditor confirms they are done.

Discord & email (required)

Discord username (as it appears in Akash Discord — confirm membership and Provider role):
dvargasracero

Email:
support@1legion.com

Contact (optional)

• Organization / Name:
  1Legion

• Website:
  https://1legion.com

Prerequisite checks

1. Provider attributes and contact info are configured on-chain

• region: ca-montreal
• host: akash
• tier: community
• organization: 1Legion
• cuda-version: 13.0
• capabilities/gpu/vendor/nvidia/model/rtxpro6000blackwellmaxqworkstationedition: true
• capabilities/gpu/vendor/nvidia/model/rtxpro6000blackwellmaxqworkstationedition/ram/95Gi: true
• capabilities/gpu/vendor/nvidia/model/rtxpro6000blackwellmaxqworkstationedition/interface/pcie: true

Provider info:

• email: support@1legion.com
• website: https://1legion.com

2. Provider host and ingress DNS resolve correctly

• provider.1legion.cloud -> 38.108.83.87
• *.ingress.1legion.cloud -> 38.108.83.87

3. TLS / public provider endpoint

• Provider endpoint is live at:
  https://provider.1legion.cloud:8443

4. Hardware

• 8x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition
• 95 GiB VRAM advertised per GPU
• 192 vCPU
• 527 GiB RAM allocatable
• ~1.7 TiB ephemeral storage currently allocatable
• Location / region: Montreal, Canada (ca-montreal)

5. Provider operating mode

• Community provider
• GPU-enabled
• Persistent storage is intentionally not enabled yet
• Public rentals are currently targeted at stateless / ephemeral workloads

6. Smoke test / live validation

We successfully validated the provider with a real mainnet smoke test:

• deployment created
• provider bid received
• lease created successfully
• manifest submission passed
• deployment then closed cleanly after validation

[Zblocker64]

does these cards have 96gb of ram?

[Zblocker64]

you also most likely need to open an issue or submit a PR here

https://github.com/akash-network/provider-configs/tree/main

[dvargasracero]

yes these cards have 96GB RAM each

[Zblocker64]

then this is incorrect

capabilities/gpu/vendor/nvidia/model/rtxpro6000blackwellmaxqworkstationedition: true
capabilities/gpu/vendor/nvidia/model/rtxpro6000blackwellmaxqworkstationedition/ram/95Gi: true
capabilities/gpu/vendor/nvidia/model/rtxpro6000blackwellmaxqworkstationedition/interface/pcie: true

[dvargasracero]

I created a ticket. I believe this GPU is not listed here. Would you like me to change the name to something shorter?

[dvargasracero]

I am not able to see my server in the pricing of the server in the akash console, I created the server in CLI. Is there anywhere in the GUI I can see the transactions?

[Zblocker64]

please follow these instructions

https://github.com/akash-network/provider-configs/blob/main/README.md

[dvargasracero]

Thanks — I reviewed the README and checked the live node again.

The GPU on our worker is:

• vendor: 10de
• device: 2bb4
• model: NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition

Live verification from the node:

• nvidia-smi --query-gpu=name,memory.total,driver_version,pci.device_id --format=csv,noheader
  • NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition, 97887 MiB, 580.126.20, 0x2BB410DE
• lspci -nn
  • NVIDIA Corporation GB202GL [RTX PRO 6000 Blackwell Max-Q Workstation Edition] [10de:2bb4]

From the current provider-configs gpus.json, I can see:

• 10de:2bb1 -> pro6000we
• 10de:2bb5 -> pro6000se

but not 10de:2bb4.

If you are okay with it, I would propose the canonical slug for 10de:2bb4 as:

pro6000maxq

Once confirmed, I will update my provider attributes accordingly.

[dvargasracero]

and on the Audit issue, Thanks — I reviewed the provider-configs README and rechecked the live GPU node.

You were right that the attributes I posted are not canonical.

The live GPU is:

• vendor/device: 10de:2bb4
• model: NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition
• VRAM class: 96Gi

The current official provider-configs database does not yet include 10de:2bb4, so the long-form attributes I posted earlier were fallback/manual values and should not be treated as correct.

I have already opened the corresponding provider-configs issue here:
akash-network/provider-configs#78

Once the canonical mapping for 10de:2bb4 is confirmed there, I will update the provider attributes and correct this audit request accordingly.

[Zblocker64]

you have still failed to read the instruction in the provider-configs repo. Please create and issue or PR there with our gpu ONLY. do not include audit details or it will be closed

[dvargasracero]

done submitted the GPU part separately

[dvargasracero]

Hi there, shall I rename the GPU to pro6000maxq?

[Zblocker64]

pro6000mq

[dvargasracero]

Hi, I have changed the model number, please could you take a look if we can pass the audit now.