コンテンツにスキップ

`POST /api/v1/pools/{name}/takeover` — point this pool's members at one that can poll them.

POST
/api/v1/pools/{name}/takeover
curl --request POST \
--url https://example.com/api/v1/pools/example/takeover \
--header 'Authorization: Bearer <token>' \
--header 'Content-Type: application/json' \
--data '{ "to": "example" }'

The second of the two things an operator can do about a pool with no live poller. The first — bring the site back — no endpoint can do for them: the site’s own box has to be handed a new bundle, and this deployment cannot reach it. This one is reversible; restore_pool puts every member back where this found it, including the ones that were inheriting rather than assigned, which is why the record is a table and not a column.

🚨 This is never the right thing to do merely because a pool is uncovered. A site poller usually exists because core cannot reach those devices; covering them from a host that cannot see them turns one accurate pool alert into N false unreachable ones — worse than the silence it replaces, and indistinguishable from a real outage. It is offered to a person looking at the alert, who can test reachability from this host first (ADR-107 Inc.4 decision 4). Nothing calls it automatically and nothing defaults to it.

name
required
string

The pool that has no live poller

Media typeapplication/json

Where a covered pool’s members should be pointed.

object
to
required

The pool to point them at — one that has a live poller, usually the co-located one.

string
Examplegenerated
{
"to": "example"
}

How many nodes and folders were re-pointed

Media typeapplication/json

What one call re-pointed.

object
folders
required

Folders re-pointed.

integer format: int64
nodes
required

Nodes re-pointed.

integer format: int64
Examplegenerated
{
"folders": 1,
"nodes": 1
}

The destination is not a usable pool name, or is this pool

Not signed in

ManageSystem required

No admin state