Repository navigation
Create overview.mdx - #864
lavanya-gunreddi wants to merge 11 commits into
Conversation
|
Preview deployment for your docs. Learn more about Mintlify Previews.
💡 Tip: Enable Automations to automatically generate PRs for you. |
prestwick
left a comment
There was a problem hiding this comment.
Handful of suggested changes and nitpicks.
| In the GPU picker, find the **External compute** section. It lists the GPU types available on your organization's online machines. Select the GPU you want. | ||
| </Step> | ||
| <Step title="Configure and deploy"> | ||
| Configure the rest of the Pod the way you normally would, including the template, storage, and environment variables. Then click **Deploy**. |
There was a problem hiding this comment.
Is there a doc with the workflow we reference that can be linked here?
|
|
||
| ## How placement works | ||
|
|
||
| When you deploy on external compute, Runpod places the Pod on one of your organization's online machines that matches the GPU you selected. Pods deployed on your machines stay within your organization. Other Runpod customers can't deploy on them. |
There was a problem hiding this comment.
| When you deploy on external compute, Runpod places the Pod on one of your organization's online machines that matches the GPU you selected. Pods deployed on your machines stay within your organization. Other Runpod customers can't deploy on them. | |
| When you deploy on external compute, Runpod places the Pod on one of your organization's online machines that matches the GPU you selected. Pods deployed on your machines stay within your organization. Other Runpod customers can't deploy to them. |
| In the Runpod console, go to **Resources → External compute** and click **Add compute**. | ||
| </Step> | ||
| <Step title="Describe the machine"> | ||
| Select your machine type, such as an on-prem server, enter a name for the machine, and click **Next**. |
There was a problem hiding this comment.
In my testing there was some limitation on the hostnames we could give the machines. Our generated hostnames use underscores by default, and I had to change those to hyphens to get things working. I'm going to escalate this to Mike and Audry, but this may require some mention.
From https://man7.org/linux/man-pages/man7/hostname.7.html
Each element of the hostname must be from 1 to 63 characters long
and the entire hostname, including the dots, can be at most 253
characters long. Valid characters for hostnames are ASCII(7)
letters from a to z, the digits from 0 to 9, and the hyphen (-).
A hostname may not start nor end with a hyphen.
|
|
||
| This page covers the most common problems when you connect and use external compute, and what to do about each. | ||
|
|
||
| ## Verification failed |
There was a problem hiding this comment.
There is also a "verification timed out" state. When I was testing this was because sshd wasn't listening on all interfaces and therefore Nebula the Runpod control plane couldn't make an SSH connection to the machine. This likely could be generalized as the machine couldn't complete the network connection to the control plane.
There was a problem hiding this comment.
Added a “Verification timed out” section, written generally as Runpod not being able to reach the machine, with the sshd fix as the first step. Also added it to the status table on the setup page.
|
|
||
| ## Requirements | ||
|
|
||
| Runpod Hybrid Cloud is available to organizations only. The Runpod team must enable External Compute for your org account before the feature appears in your console. [Contact us](https://www.runpod.io/everywhere) to get set up. |
There was a problem hiding this comment.
the term 'external compute' is being debated and might be changed, fyi. TBD if it changes to 'private compute'.
| ## How it works | ||
|
|
||
| <Steps> | ||
| <Step title="Add your machine"> |
There was a problem hiding this comment.
Let's use compute. Lines up better with the planned terminology changes.
| <Step title="Add your machine"> | ||
| In the console, go to **Resources → External compute** and click **Add compute**. Select your machine type (on-prem server, AWS EC2 instance, or any cloud VM), give it a name, and click **Next**. | ||
| </Step> | ||
| <Step title="Install the host agent"> |
There was a problem hiding this comment.
will need to align on how to refer to "host agent". i think we might refer to it as "Runpod agent" preferrably, but let's get a decision memorialized then we can mass update. @prestwick
There was a problem hiding this comment.
Current recommendation is Host connector.
| In the console, go to **Resources → External compute** and click **Add compute**. Select your machine type (on-prem server, AWS EC2 instance, or any cloud VM), give it a name, and click **Next**. | ||
| </Step> | ||
| <Step title="Install the host agent"> | ||
| Copy the install command that appears and paste it into an SSH session on your machine. This installs the Runpod host agent, a small program that connects your machine to Runpod over an outbound-only encrypted connection. No inbound ports are opened. Set your public and private IPs, open the required firewall ports, then reboot. For full instructions, see [Set up external compute](/hybrid-cloud/setup). |
|
|
||
| - An [organization](/organizations/orgs-overview) with external compute enabled. During alpha, the Runpod team enables this for you. | ||
| - An organization role that allows you to add compute. See [Roles](/organizations/org-roles). | ||
| - A single machine with at least one GPU. Multi-node machines connected by InfiniBand or a similar fabric aren't supported yet. |
There was a problem hiding this comment.
@maltmanRP you good with this language? thinking we could leave it more open ended like "talk to runpod" if they need multinode usecase.
There was a problem hiding this comment.
I like this phrasing. Customer feedback on this will help with prioritization.
| 2. Check that the machine can make outbound connections to the internet. | ||
| 3. Confirm that you rebooted the machine after you configured networking. | ||
|
|
||
| ## Machine stuck in verification in progress |
There was a problem hiding this comment.
fyi @prestwick it will timeout after 15 min and retry once, then if that fails again, will end in Verification failed state. don't know if we need to expose all the nitty gritty but we have sweepers and retries that will make sure these land in some terminal state after ~30 minutes.
There was a problem hiding this comment.
Good to know. Maybe we state here that verification should complete within 30 minutes. Otherwise contact support.
| - At least one machine with the **Online** status. See [Set up external compute](/hybrid-cloud/setup). | ||
| - Membership in the organization that owns the machine, with permission to deploy Pods. See [Roles](/organizations/org-roles). | ||
|
|
||
| ## Deploy a Pod |
There was a problem hiding this comment.
The completeness of the setup workflows for serverless are mixed at the moment. For example the quick deploy and deploy from hub options currently lack the External compute option.
| | Docker disk | A dedicated disk of at least 600 GB. | | ||
| | Disk throughput | At least 150 MB/s. | | ||
| | Outbound network | All TCP, UDP 4242 (Runpod control plane), UDP 53 (DNS), UDP 123 (time sync), and ICMP echo. | | ||
| | Inbound network | TCP 10000–60000. | |
There was a problem hiding this comment.
| | Inbound network | TCP 10000–60000. | | |
| | Inbound network | TCP 10000 and 20000. | |
Apologies my original guidance was incorrect. I've validated we don't need this large range. Only these two ports are checked during verification.
| 2. Check that the machine can make outbound connections to the internet. | ||
| 3. Confirm that you rebooted the machine after you configured networking. | ||
|
|
||
| ## Machine stuck in verification in progress |
There was a problem hiding this comment.
Good to know. Maybe we state here that verification should complete within 30 minutes. Otherwise contact support.
| - At least one machine with the **Online** status. See [Set up external compute](/hybrid-cloud/setup). | ||
| - Membership in the organization that owns the machine, with permission to deploy Pods. See [Roles](/organizations/org-roles). | ||
|
|
||
| ## Deploy a Pod |
There was a problem hiding this comment.
The completeness of the setup workflows for serverless are mixed at the moment. For example the quick deploy and deploy from hub options currently lack the External compute option.
|
|
||
| ## Get started | ||
|
|
||
| Runpod Hybrid Cloud is available to early customers, and the Runpod team works with you directly to set up your machines. [Request a demo](https://www.runpod.io/hybrid-cloud) to get started. |
There was a problem hiding this comment.
Maybe nitpick, I feel this "early customers" reads a little bit wierd.
How about "Runpod Hybrid Cloud is in alpha and available to select customers, ..."?
No description provided.