neuron-technologies/neuron
3
0
Fork 0
Code Issues 44 Pull Requests Actions Packages Projects Releases Wiki Activity

Save-side: 46.6% of the graph (36,737 nodes) are byte-identical copies of one record #144

New Issue
Open
opened 2026-08-08 00:33:25 +00:00 by tim.lingo · 0 comments
tim.lingo commented 2026-08-08 00:33:25 +00:00
Member
Copy Link
Copy Source

Found during the ten-iteration retrieval run, on a 78,768-node copy of the live graph. Not a retrieval defect — a save-side one. Filed separately because it is out of scope for #141 and needs its own decision.

Measured: 36,737 of 78,768 nodes — 46.6% of the entire graph — are byte-identical copies of one single record.

Consequences already observed:

  • retrieval result lists returning the same node repeatedly
  • every corpus scan paying ~1.9x for nothing
  • any duplicate-rate or growth metric computed on this graph is meaningless until it is understood

Not established: how they got there, whether they are still being produced, and whether the live graph has the same ratio (measured on the derived corpus at ~/neuron-eval-corpora/snapshot-pre-repair-20260806-embedded.json, taken from the live snapshot).

First step is one measurement, not a fix: count them on the live snapshot, then check whether the count grows between two boots. If it grows, the producer is live and that is the actual bug.

Related: #87 (non-ASCII mangling on the same save route), #67 (arena recycling on the store's write path), inventory row B-32.

Found during the ten-iteration retrieval run, on a 78,768-node copy of the live graph. Not a retrieval defect — a save-side one. Filed separately because it is out of scope for #141 and needs its own decision. **Measured:** 36,737 of 78,768 nodes — **46.6% of the entire graph** — are byte-identical copies of one single record. Consequences already observed: - retrieval result lists returning the same node repeatedly - every corpus scan paying ~1.9x for nothing - any duplicate-rate or growth metric computed on this graph is meaningless until it is understood **Not established:** how they got there, whether they are still being produced, and whether the live graph has the same ratio (measured on the derived corpus at `~/neuron-eval-corpora/snapshot-pre-repair-20260806-embedded.json`, taken from the live snapshot). **First step is one measurement, not a fix:** count them on the live snapshot, then check whether the count grows between two boots. If it grows, the producer is live and that is the actual bug. Related: #87 (non-ASCII mangling on the same save route), #67 (arena recycling on the store's write path), inventory row B-32.
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
BETA-CRITICAL
On the beta critical path — must resolve before ship
 blocks-public-beta
Must be fixed before any public beta
 HELD
Blocked pending an external decision (counsel / dual sign-off)
 ORTHOGONAL
Parallel workstream (testing/hardening) — safe to assign off critical path
 P0
Drop-everything severity
 POST-BETA
Important but not blocking the beta ship
 security
Security vulnerability
No labels
Milestone
No items
No Milestone
Projects
Clear projects
No projects
Assignees
Clear assignees
andre.almeida (Andre Almeida) tim.lingo (Tim Lingo) will.anderson (Will Anderson)
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: neuron-technologies/neuron#144
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?