Files
multica/server/pkg/db/queries/runtime.sql
Bohan Jiang a73336dcf8 feat(daemon): persistent UUID identity + legacy-id merge at register-time (#1220)
* feat(daemon): persistent UUID identity + legacy-id merge at register-time

daemon_id is now a stable UUID persisted to `<profile-dir>/daemon.id` on
first start, replacing the hostname-derived id that drifted whenever
`.local` appeared/disappeared, a system was renamed, or a profile
switched — each of which used to mint a fresh `agent_runtime` row and
strand agents on the old one.

To migrate existing installs without operator intervention, the daemon
reports every legacy id it may have registered under previously
(`host`, `host` with `.local` stripped, and `host[-profile]` variants
for both). At register-time the server looks up each candidate row
scoped to (workspace, provider), re-points its agents and tasks onto
the new UUID-keyed row, records which legacy id was subsumed in the
new `legacy_daemon_id` column for audit, and deletes the stale row.
Result: users running `xxx.local`-keyed runtimes today transparently
land on the new UUID row on next daemon restart.

The hostname-prefix `MigrateAgentsToRuntime` / `daemon_id LIKE '...-%'`
compatibility shim is no longer needed and has been removed along with
the handler call that invoked it.

* fix(daemon): handle bidirectional .local drift and case drift in legacy merge

Review on #1220 flagged two gaps in the legacy-id migration candidate set:

1. Reverse .local: LegacyDaemonIDs only added the stripped variant when the
   current hostname ended in `.local`. The opposite direction — DB has
   `foo.local`, current host is `foo` — was missed, so runtimes registered
   under the `.local` variant stayed orphaned after upgrade. Now both
   variants (`foo` and `foo.local`) are always emitted, regardless of what
   `os.Hostname()` currently returns, plus their `-<profile>` suffix forms.

2. Case drift: os.Hostname() has been observed returning different casings
   on the same machine across mDNS/reboot state. A case-sensitive `=`
   comparison stranded rows like `Jiayuans-MacBook-Pro.local` when the
   daemon later reported `jiayuans-macbook-pro.local`. FindLegacyRuntimeByDaemonID
   now uses `LOWER(daemon_id) = LOWER(@daemon_id)` on both sides, so casing
   differences merge rather than orphan. The (workspace_id, provider) prefix
   still bounds the scan to a tiny set of rows so the non-indexed LOWER()
   comparison has negligible cost.

Tests: TestLegacyDaemonIDs gets the mixed-case + reverse-direction cases;
daemon_test.go adds TestDaemonRegister_MergesLegacyDaemonIDRuntime_ReverseDotLocal
and TestDaemonRegister_MergesLegacyDaemonIDRuntime_CaseDrift.

* fix(daemon): consolidate every case-duplicate legacy runtime, not just the first

Follow-up review on #1220: after switching to `LOWER(daemon_id) =
LOWER(@daemon_id)`, the single-row lookup still only merged one legacy
row per candidate. If a machine already had two rows in the DB that
differed only in casing (e.g. `Jiayuans-MacBook-Pro.local` AND
`jiayuans-macbook-pro.local` coexisting because earlier hostname drift
already minted a duplicate), only one of them got consolidated and the
other stayed orphaned — violating the "no duplicate runtime per machine
after backfill" acceptance.

- FindLegacyRuntimeByDaemonID → FindLegacyRuntimesByDaemonID (:many)
- mergeLegacyRuntimes iterates every returned row and dedupes across
  overlapping legacy candidates so `foo` and `foo.local` both resolving
  to the same stored row don't double-process

Test: TestDaemonRegister_MergesAllCaseDuplicateLegacyRuntimes seeds two
case-duplicate rows with one agent each and confirms both rows are
deleted and both agents end up on the new UUID-keyed row.
2026-04-17 15:10:38 +08:00

135 lines
4.8 KiB
SQL

-- name: ListAgentRuntimes :many
SELECT * FROM agent_runtime
WHERE workspace_id = $1
ORDER BY created_at ASC;
-- name: GetAgentRuntime :one
SELECT * FROM agent_runtime
WHERE id = $1;
-- name: GetAgentRuntimeForWorkspace :one
SELECT * FROM agent_runtime
WHERE id = $1 AND workspace_id = $2;
-- name: UpsertAgentRuntime :one
INSERT INTO agent_runtime (
workspace_id,
daemon_id,
name,
runtime_mode,
provider,
status,
device_info,
metadata,
owner_id,
last_seen_at
) VALUES ($1, $2, $3, $4, $5, $6, $7, $8, $9, now())
ON CONFLICT (workspace_id, daemon_id, provider)
DO UPDATE SET
name = EXCLUDED.name,
runtime_mode = EXCLUDED.runtime_mode,
status = EXCLUDED.status,
device_info = EXCLUDED.device_info,
metadata = EXCLUDED.metadata,
owner_id = COALESCE(EXCLUDED.owner_id, agent_runtime.owner_id),
last_seen_at = now(),
updated_at = now()
RETURNING *;
-- name: UpdateAgentRuntimeHeartbeat :one
UPDATE agent_runtime
SET status = 'online', last_seen_at = now(), updated_at = now()
WHERE id = $1
RETURNING *;
-- name: SetAgentRuntimeOffline :exec
UPDATE agent_runtime
SET status = 'offline', updated_at = now()
WHERE id = $1;
-- name: MarkStaleRuntimesOffline :many
UPDATE agent_runtime
SET status = 'offline', updated_at = now()
WHERE status = 'online'
AND last_seen_at < now() - make_interval(secs => @stale_seconds::double precision)
RETURNING id, workspace_id;
-- name: FailTasksForOfflineRuntimes :many
-- Marks dispatched/running tasks as failed when their runtime is offline.
-- This cleans up orphaned tasks after a daemon crash or network partition.
UPDATE agent_task_queue
SET status = 'failed', completed_at = now(), error = 'runtime went offline'
WHERE status IN ('dispatched', 'running')
AND runtime_id IN (
SELECT id FROM agent_runtime WHERE status = 'offline'
)
RETURNING id, agent_id, issue_id;
-- name: ListAgentRuntimesByOwner :many
SELECT * FROM agent_runtime
WHERE workspace_id = $1 AND owner_id = $2
ORDER BY created_at ASC;
-- name: DeleteAgentRuntime :exec
DELETE FROM agent_runtime WHERE id = $1;
-- name: CountActiveAgentsByRuntime :one
SELECT count(*) FROM agent WHERE runtime_id = $1 AND archived_at IS NULL;
-- name: DeleteArchivedAgentsByRuntime :exec
DELETE FROM agent WHERE runtime_id = $1 AND archived_at IS NOT NULL;
-- name: FindLegacyRuntimesByDaemonID :many
-- Looks up runtime rows keyed on a prior (hostname-derived) daemon_id. Used
-- at register-time to find rows owned by the same machine under its old
-- identity so agents/tasks can be re-pointed at the new UUID-keyed row.
--
-- Comparison is case-insensitive because os.Hostname() has been observed to
-- return different casings on the same machine (e.g. `Jiayuans-MacBook-Pro`
-- vs `jiayuans-macbook-pro`) across reboots/mDNS state changes. A case-
-- sensitive `=` would strand the old row; LOWER() on both sides handles drift
-- without forcing the daemon to enumerate cased permutations.
--
-- Returns many rather than one because case drift may have already minted
-- duplicate rows historically (e.g. `Foo.local` AND `foo.local` under the
-- same workspace+provider). A single-row lookup would consolidate only one
-- of them and leave the rest orphaned. Callers must merge every returned
-- row into the new UUID-keyed runtime.
SELECT * FROM agent_runtime
WHERE workspace_id = @workspace_id
AND provider = @provider
AND LOWER(daemon_id) = LOWER(@daemon_id);
-- name: ReassignAgentsToRuntime :execrows
-- Re-points every agent referencing old_runtime_id at new_runtime_id.
UPDATE agent
SET runtime_id = @new_runtime_id
WHERE runtime_id = @old_runtime_id;
-- name: ReassignTasksToRuntime :execrows
-- Re-points every queued/running/completed task referencing old_runtime_id.
-- Required before deleting the old runtime row because agent_task_queue has
-- an ON DELETE CASCADE FK that would otherwise drop historical tasks.
UPDATE agent_task_queue
SET runtime_id = @new_runtime_id
WHERE runtime_id = @old_runtime_id;
-- name: RecordRuntimeLegacyDaemonID :exec
-- Remembers the most recent hostname-derived daemon_id that was merged into
-- this row. Useful for debugging when tracing back why a given runtime row
-- subsumed an old one, and only overwrites NULL so the earliest merge is
-- preserved.
UPDATE agent_runtime
SET legacy_daemon_id = COALESCE(legacy_daemon_id, $2)
WHERE id = $1;
-- name: DeleteStaleOfflineRuntimes :many
-- Deletes runtimes that have been offline for longer than the TTL and have
-- no agents bound (active or archived). The FK constraint on agent.runtime_id
-- is ON DELETE RESTRICT, so we must exclude all agent references.
DELETE FROM agent_runtime
WHERE status = 'offline'
AND last_seen_at < now() - make_interval(secs => @stale_seconds::double precision)
AND id NOT IN (SELECT DISTINCT runtime_id FROM agent)
RETURNING id, workspace_id;