ai-memory  /  v0.7.0  /  hive-1461-baseline
Peer-reproducible baseline · #1461

hive-1461 — federated DigitalOcean baseline.

A deterministic, idempotent 0→60 build of the ai-memory v0.7.0 federated test fleet on DigitalOcean. Terraform stands the infrastructure up; a push-based SSH toolkit brings every node to a verified federated state; a verification harness proves it. Everything a reviewer needs to reproduce both the environment and the results ships inside release/v0.7.0 under deploy/hive-1461/. Seven nodes across two regions, all traffic TLS+mTLS, every step re-runnable.

VERIFIED · 40/40 validate · 22/22 full-spectrum · ZERO FAIL
Headline

The numbers.

40/40validate PASS
22/22full-spectrum PASS
7droplets
2regions (nyc3+sfo2)
3peer quorum mesh
W=2/N=3write quorum
5test groups
0FAIL across 62 checks

One entrypoint reproduces the whole thing: make seed up provision validate test builds the fleet, brings every node to a verified federated state, runs the verification harness, then runs the full-spectrum P3 suite. make down tears it all back down. Every probe in both harnesses goes over the real TLS+mTLS path; the canonical green reports are committed at deploy/hive-1461/results/.

Topology

7 nodes, function-encoded hostnames.

Hostnames encode each node's function: hive-<function>-<region>-<NN>. The 3 peers form a W=2-of-N=3 write-quorum mesh; agents and ctrl are pure mTLS clients of that mesh — they hold a client cert only, never a server cert, and run no inbound HTTPS daemon.

HostRoleRegionSizeRuns
hive-peer-nyc3-01peernyc3s-4vcpu-8gbfederated serve + PG16/AGE + Ollama embedder
hive-peer-nyc3-02peernyc3s-4vcpu-8gbfederated serve + PG16/AGE + Ollama embedder
hive-peer-sfo2-01peersfo2s-4vcpu-8gbfederated serve + PG16/AGE + Ollama embedder
hive-agent-nyc3-01agentnyc3s-1vcpu-2gbxAI grok-4.3 NHI client (mTLS API client)
hive-agent-nyc3-02agentnyc3s-1vcpu-2gbxAI grok-4.3 NHI client
hive-agent-nyc3-03agentnyc3s-1vcpu-2gbxAI grok-4.3 NHI client
hive-ctrl-nyc3-01ctrlnyc3s-8vcpu-16gbloadgen / chaos / orchestration (mTLS client)

Account stays well within the 15-droplet cap (7 in use). All fleet traffic is TLS + mTLS: the peer HTTPS port enforces client_auth_mandatory — a connection is accepted only if the SHA-256 of the client cert's DER bytes is on the mtls-allowlist.txt (fingerprint pinning, the SSH known_hosts model). Outbound quorum/API clients verify peer server certs against the campaign CA, whose SAN pins each peer's public IP.

0→60 flow

Five make targets, eight provision steps.

Every target is idempotent and re-runnable. Provisioning is push-based over SSH; nothing is installed on the operator host beyond the CLI prereqs (terraform jq openssl ssh scp curl). The provision steps run in this exact deterministic order:

StepScriptEffect
0000_render_inventory.shproject terraform output fleetinventory.json
0505_wait_ssh.shblock until every node accepts SSH
1010_binary.shfan out the golden binary; assert version + sha256
2020_pg_age.shper-peer PG16/AGE/pgvector container + schema-init (v55)
2525_ollama_embed.shper-peer CPU Ollama sidecar serving nomic-embed-text (768-dim)
3030_config.shrender + push per-role config.toml + secret EnvironmentFile
4040_tls.shcampaign CA + per-node leaf certs + mTLS allowlist fan-out
5050_federation.shper-peer systemd unit; start the quorum mesh; health-gate
Verification harness

make validate — 40/40 PASS.

The verification harness exercises the live fleet over the real TLS+mTLS path and emits a JSON + tabular report. Exit status is 0 iff every check is green. Per node it asserts binary sha256 + --version; per peer it asserts /api/v1/health, storage_backend == postgres, db_schema_version == 55, single-instance, and systemd-active; and a fleet federation-convergence probe writes a collective-scope marker to one peer and reads it back by id on another over the encrypted path.

Canonical green baseline committed at results/verify-baseline.{json,tsv}: TOTAL=40 PASS=40 FAIL=0.

Full-spectrum testing

make test — 22/22 across 5 groups.

The P3 suite runs against the live fleet, gated behind a green make validate. Every probe goes over the real TLS+mTLS path and authenticates with x-api-key; throwaway markers land in the _test / _verify namespaces and are best-effort deleted, so the baseline corpus is never mutated. Canonical green report committed at results/test-full-spectrum.{json,tsv}: TOTAL=22 PASS=22 FAIL=0.

Group 1 · 6 checks regression CRUD roundtrip; semantic search (exercises the nomic embedder end-to-end); namespace isolation; private-scope owner visibility (a private memory is invisible to a different caller). 6/6 PASS
Group 2 · 7 checks crypto Negative TLS/mTLS + authz: no client cert refused (000); non-allowlisted client cert refused (000); wrong server CA refused (000); privileged endpoint without x-api-key401; with key → 200; /health exempt → 200; admin endpoint as non-admin → 403. 7/7 PASS
Group 3 · 3 checks federation Write to peer-1; converge on peer-2 (same region) and peer-3 (cross-region nyc3→sfo2) within the catch-up window. 3/3 PASS
Group 4 · 3 checks a2a Agent-to-agent E2E: agent-alpha (an mTLS client node) writes a collective memory to a peer over the network; agent-beta (a different client identity on a different node) reads it back on the write peer and on a federated peer. 3/3 PASS
Group 5 · 3 checks ai_nhi The grok-driven NHI loop: agent-alpha drives a live xAI/grok expand_query decision over the mesh, commits the LLM-derived term as a collective memory, and the decision converges on a federated peer — a full NHI decision → commit → federate loop. 3/3 PASS

The live AI-NHI loop is the capstone: in the canonical run, grok returned the term paxos algorithm, the harness committed it as a collective memory on hive-peer-nyc3-01, and it converged on hive-peer-nyc3-02 — a real LLM decision crossing the encrypted mesh and federating cross-node.

What "reproducible" means here

Pinned, deterministic, idempotent, verifiable.

Reproduce it

Build, prove, full-spectrum test.

deploy/hive-1461/README.mdThe runbook — prerequisites, secrets, 0→60 flow, security model, verification report deploy/hive-1461/MakefileSingle entrypoint: make seed up provision validate test (teardown: make down) deploy/hive-1461/provision/Push-based 0→60 toolkit: steps 00–50 + lib.sh pinned-constant source + pg-age/ results/verify-baseline.tsvCanonical green verification report — TOTAL=40 PASS=40 FAIL=0 results/test-full-spectrum.tsvCanonical green full-spectrum report — TOTAL=22 PASS=22 FAIL=0
Provenance

Reproducibility contract.

Campaign
#1461 — hive-fleet peer-reproducible baseline
Operator
justin@alpha-one.mobi
Authoring agent
Claude (Opus 4.7), AI NHI under operator authority
Cloud
DigitalOcean — nyc3 + sfo2 (7 droplets, 15-droplet cap)
Binary
golden ai-memory v0.7.0, features sal,sal-postgres,sqlite-bundled
Binary sha256
5e86de3ab0be6a19e02f760390651d9c425ba7ca73b9b8c7db0ce3b6f25a0aa7
Schema
v55 (PG16 + Apache AGE 1.6.0 + pgvector)
Embedder
nomic-embed-text (768-dim) via CPU Ollama sidecar per peer
NHI driver LLM
xAI grok-4.3 (agents); peer chat LLM via OpenRouter
Quorum
W=2-of-N=3 write-quorum mesh across 3 peers
Transport
TLS + mTLS everywhere; client_auth_mandatory fingerprint-pinned allowlist
Branch
release/v0.7.0
Build commits
219a0e39a (P3 harness) · 10a9d97eb (docs + canonical results)
Verification
40/40 PASS · full-spectrum 22/22 PASS · zero FAIL
Tracking issue
#1461 — hive-fleet install campaign
Related

Other v0.7.0 campaigns.

2026-05-28 — Ship campaign (SHIP-CLEARED)Integration sweep + dogfood + postgres+AGE 8,028 + docs/Pages drift + CoALA + AI-NHI v3 2026-05-22 — Release-gate final (7,321 / 0 / 0)22 release-gate defects closed; SHIP-RECOMMENDED Heterogeneous AI-NHI assessmentThe independent fresh-agent assessment series v0.7.0 release notesWhat changed since v0.6.4