Carry the profile fingerprint to the agent and back
CI / test (push) Successful in 1m39s

The stale-profile flag could not fire where it matters. The control plane
keeps no record of a running replica across its own restart — it rebuilds
them from what the agents report — and adoption left the fingerprint empty,
so every replica read as unknown afterwards. That is precisely the moment
the catalog is most likely to have been edited: change the configuration,
restart the server, and nothing would ever say the fleet disagrees with it.

The fingerprint now travels with StartShard, the agent keeps it beside the
pool and fabric it already echoes back, and adoption takes the replica's
version from the agent rather than from the catalog it is about to be
compared against. An agent too old to report it sends nothing, which still
reads as unknown rather than as matching.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
This commit is contained in:
Ilker ManapandClaude Opus 5 committed 2026-09-27 23:22:44 +02:00
1 parent 40f5c930c4
commit 03e38e7dda
7 files changed
+122 -11

No files matched your search

+1
View File
@@ -515,6 +515,7 @@ func (a *Agent) runningShards() []*llmgwv1.RunningShard {
Profile: spec.Profile.Name,
Pool: spec.Pool,
Fabric: spec.Fabric,
ProfileFp: spec.ProfileFP,
LeaderAddr: a.cfg.MgmtAddr,
LeaderPort: int32(spec.LeaderPort),
})
+5
View File
@@ -72,6 +72,11 @@ type ShardSpec struct {
// not used when building a command.
Pool string
Fabric string
// ProfileFP identifies the version of the profile this shard was started
// from. Carried for the same reason as Pool and Fabric, and never read: it
// lets the control plane tell, after its own restart, whether a shard that
// is still running matches the profile the catalog holds now.
ProfileFP string
}
// Leader reports whether this shard serves HTTP for the replica.