Repository navigation
Add sled-agent API to report multirack join service status - #11535
Merged
Merged
Conversation
We need a way to discover information about sleds in unadopted racks so that we can show them to operators. The easiest way to do this is to return that information from the multirack join service, which already has it. This provides a sled-agent API endpoint to do just that. It adds a separate watch channel for just the status we care about for the control plane and utilizes that via the `MultirackJoinAccessor`, which was formally the `MultirackJoinCompleter`. Now, the question is: How do we find the sled-agent running the multirack join service? That is going to be solved in a separate PR. The sled-agent running the join service will publish the new rack subnet to announce along with its own IP to ddmd. DDMD will announce this information which will make its way to the proto rack ddmd servers. Nexuses on the proto rack can then poll their local sled-agent to discover the location of the sled-agent running the multirack join service and ask it for this info. Nexus can also complete the multirack join via the same sled-agent when the user adopts the new rack into the cluster. See oxidecomputer/maghemite#939
andrewjstone
force-pushed
the
sa-mr-join-status
branch
from
October 20, 2026 20:28
844482e to
d507f3e
Compare
jgallagher
reviewed
Oct 21, 2026
jgallagher
approved these changes
Oct 21, 2026
jgallagher
left a comment
jgallagher
left a comment
Contributor
There was a problem hiding this comment.
Thanks, all the change LGTM. I assume sorting out #11535 (comment) will be minor, so happy to go ahead and approve.
nickelization
approved these changes
Oct 22, 2026
|
|
||
| /// Return the status of this sled's multirack join service. | ||
| /// | ||
| /// Returns a 404 if no multirack join service is running, or a 409 if the |
Contributor
There was a problem hiding this comment.
Cool, I can't remember ever seeing 409 used before, always fun seeing unusual status codes in places that actually make sense for them :)
| fn from(value: StartSledAgentsStatus) -> Self { | ||
| let sled_id = value.this_nodes_sled_id; | ||
| let rack_id = value.rack_id; | ||
| let rack_subnet = Ipv6Subnet::<RACK_PREFIX_LENGTH>::new( |
Contributor
There was a problem hiding this comment.
Genuinely not sure if this is worth it, but I wonder if it would be a good idea to check that all the sleds have the same prefix and log a warning or something here if they don't?
Contributor
Author
There was a problem hiding this comment.
They all get generated from the same prefix, so they share it by construction. This is just to extract it back out for the API.
While it doesn't matter here, I do think that Nexus should validate this when it receives the response containing this data.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
We need a way to discover information about sleds in unadopted racks
so that we can show them to operators. The easiest way to do this is to
return that information from the multirack join service, which already
has it. This provides a sled-agent API endpoint to do just that.
It adds a separate watch channel for just the status we care about for
the control plane and utilizes that via the
MultirackJoinAccessor,which was formally the
MultirackJoinCompleter.Now, the question is: How do we find the sled-agent running the
multirack join service? That is going to be solved in a separate PR.
The sled-agent running the join service will publish the new rack subnet
to announce along with its own IP to ddmd. DDMD will announce this
information which will make its way to the proto rack ddmd servers.
Nexuses on the proto rack can then poll their local sled-agent to
discover the location of the sled-agent running the multirack join
service and ask it for this info. Nexus can also complete the multirack
join via the same sled-agent when the user adopts the new rack into the
cluster. See oxidecomputer/maghemite#939