ARCH-01 — Concurrency-safe platform.bootstrap grant succession #823
Open
opened 2026-07-22 19:41:36 -05:00 by jcwalker3
·
0 comments
No Branch/Tag Specified
master
fix/issue-842-conflict-fix-lease-lifecycle
fix/issue-843-cross-role-allocation-handoff
fix/issue-790-slice-a-heartbeat-policy
feat/issue-628-autonomous-handoffs-orchestration
feat/issue-634-readonly-system-health-api
feat/issue-638-webui-app-shell-phase1
fix/issue-840-cross-role-queue-allocation
feat/issue-822-atomic-install-authority-kernel
feat/issue-633-console-authz-audit-model
feat/issue-636-inventory-api
fix/issue-815-preflight-worktree-forwarding
feat/issue-812-publish-unpublished-commit
feat/issue-635-project-registry-api
feat/issue-798-worker-registry-schema
feat/issue-610-live-remote-parity
docs/issue-632-web-console-architecture
fix/issue-760-exact-owner-renewal
fix/issue-787-kill-segment-separators
chore/issue-681-preserve-review-session-wip
v1.1.0
Labels
Clear labels
allocator
anti-stomp
architecture
bug
chore
codex
concurrency
contamination
control-plane
dashboard
database
design
documentation
enhancement
gitea
glitchtip
important
incident
incident-bridge
integration
jenkins
labels
leases
mcp
mcp-health
mcp-menu
multi-project
mutating
nice-to-have
observability
portability
preflight
protected-branch
queue
read-only
reconnect
recovery
refactor
release
reliability
resumable-review
reviewer
roadmap
safety
security
self-hosted
sentry
stale-runtime
status:blocked
status:in-progress
status:pr-open
status:ready
terminal-lock
testing
tracker
type:bug
type:feature
type:feature
type:guardrail
visibility
workflow
workflow-hardening
workflow-hardening
Controller-owned work allocator
Prevent concurrent LLM session stomping
Architecture / structural design
OpenAI Codex client / workflow session surface
Concurrent session safety
Workflow or session contamination incident
MCP control-plane coordination and allocation authority
MCP operational dashboard/queue view
Internal coordination storage (SQLite/Postgres)
Design / investigation, no implementation
Docs / runbooks
New feature or improvement
Gitea MCP workflow
GlitchTip integration
Operational or process incident requiring durable audit trail
Sentry-to-Gitea incident bridging
Integration testing
Jenkins integration
Label taxonomy management
Lease adopt/release/expire lifecycle
MCP server / tooling
MCP namespace and runtime health
MCP menu surface
Work spanning multiple monitoring projects or Gitea repos
Mutating action; requires gating
Observability, metrics, traces, error reporting
Cross-platform / portability
Shared preflight gates before mutation
Protected branch / stable-branch policy concern
Work queue visibility and allocation
Read-only, no mutation
MCP client reconnect/reload recovery path
Recovery paths for stale/foreign leases
Code refactor / restructure
Release / versioning
Reliability / failure handling
Persist and resume prepared review verdicts across sessions
Reviewer workflow tooling
Roadmap / umbrella issue
Safety rails and fail-closed mutation guards
Security / trust boundary
Self-hosted infrastructure integration
Sentry error monitoring integration
Stale backend daemon / runtime-vs-master parity failures
Issue is blocked
Issue is being worked on
Issue has an open pull request
Issue is ready for work
Terminal review lock (#332) path
Tests / test coverage
Issue tracker hygiene / meta
Bug or defect
Feature or enhancement
Feature or enhancement
Safety gate or guardrail
Workflow state visibility for LLMs/operators
Cross-tool workflow
LLM workflow coordination hardening
LLM workflow coordination hardening
Milestone
No items
No Milestone
Projects
Clear projects
No projects
Notifications
Due Date
No due date set.
Dependencies
No dependencies set.
Reference: Scaled-Tech-Consulting/Gitea-Tools#823
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Candidate architecture requiring executable validation; not yet approved, production-ready, or certified.
Parents: #820, #821. Depends on: #822 (kernel +
platform_active_invariant).1. Summary / objective
Define and implement concurrency-safe succession of
platform.bootstrapgrants: grant, replace, (optionally) suspend, (optionally) expire, and revoke, such that at least one active grant always remains and no two concurrent transactions can drive the active count to zero.2. Security/correctness problem
The platform root is the top of the authority lattice. If succession is not serialized on a shared invariant, two transactions can each observe two active grants and revoke different ones (write-skew), leaving zero active roots and bricking all authority. Expiry/suspension add the same hazard.
3. Threat model / failure modes
Write-skew last-grant revocation; concurrent replacement; expiry firing to zero; suspension of the last active grant; grantor-revocation cascade; raw-write reactivation of a revoked grant; resurrection via UPDATE.
4. In-scope behavior
platform_active_invariantrow (or an equivalent global lock), maintainingactive_counttransactionally.[SCHEMA]for theactive_count>=1CHECK;[TRUSTED-SERVICE]/[SCHEMA]for the serialization.platform.bootstrap → platform.bootstrap) permitted only to preserve authorized succession (seeded immutable lattice edge from #822).no_grant_resurrection,no_grant_delete).5. Exclusions
No supervisor-grant succession (separate), no project.admin (ARCH-03), no evidence, no workspaces.
6. Schema/operation contracts
platform_bootstrap_grants(from #822) +platform_active_invariantsingleton. Operations:grant_platform_bootstrap,revoke_platform_bootstrap,replace_platform_bootstrap, optionalsuspend_platform_bootstrap/resume. Each updatesactive_countunder the invariant lock.7. SQLite behavior
BEGIN IMMEDIATE; read-modify-writeactive_counton the singleton row; revoke/suspend rejected when result<1.8. PostgreSQL behavior
SELECT active_count FROM platform_active_invariant FOR UPDATE;SERIALIZABLEwith retry; losing transaction returnsLAST_PLATFORM_GRANTorSERIALIZATION_RETRY. Backend-specific expected errors documented.9. Concurrency / transaction requirements
Two concurrent revocations that each saw two active grants: exactly one succeeds, the other returns
LAST_PLATFORM_GRANT. No interleaving reaches zero active grants.10. Structured results
GRANTED,REVOKED,REPLACED,SUSPENDED,RESUMED,LAST_PLATFORM_GRANT,SERIALIZATION_RETRY,GRANT_RESURRECTION_FORBIDDEN,AUTHORIZATION_DENIED.11. Enforcement classification
active_count>=1CHECK and append-only immutability[SCHEMA]; serialization discipline[TRUSTED-SERVICE]on SQLite /[SCHEMA]-adjacent viaFOR UPDATEon PG.12. Acceptance criteria
LAST_PLATFORM_GRANT.LAST_PLATFORM_GRANT(no zero state).13. Named tests
t_grant(+),t_revoke_nonfinal(+),t_revoke_final(−),t_concurrent_last_revoke(concurrency/write-skew, both backends),t_replace_atomic(concurrency),t_suspend_last(−),t_grant_resurrect(raw-bypass),t_grant_delete(raw-bypass),t_audit_succession(+).14. Audit events / evidence
platform_grant_granted,platform_grant_revoked,platform_grant_replaced,platform_grant_suspended,platform_grant_expired. Test-run output is durable evidence.15. Dependencies / parent
Parents #820, #821; depends on #822.
16. Definition of done
Executable on SQLite and PostgreSQL; ACs pass incl. write-skew on both backends; bounded PR; author does not self-review/merge.
17. Known limitations / deferred
Supervisor-grant and project.admin succession are out of scope. If expiry is adopted, a scheduled reaper is a separate concern and must itself honor the invariant.