docs: record clean RC bootstrap and lifecycle acceptance

Co-authored-by: GPT-6 <gpt-6@openai.com>
This commit is contained in:
ZeusinaandGPT-6 Bot committed 2026-10-11 10:59:25 +03:00
1 parent 11f9197ba8
commit 8826c591dc
4 files changed
+76 -28

No files matched your search

+4
View File
@@ -4,6 +4,10 @@
## Текущий статус
- **2026-10-11 clean RC acceptance — PASS:** source RC `d8316e695bc36e1243c111e19bb1e527355907b6`, tag `zapret-autotune-managed-target:rc-d8316e6`, immutable image `sha256:5636aa9031fd0f6a0513f2e22cc0c39b1c806e9714b7f9ac040c8713d6a26c6c`, built from the source git archive with pinned inputs. These source/image identifiers predate this documentation update; a later docs commit is not the tested image source. Private source root: `/home/agent/zapret2-autotune-rc-d8316e6`. Additional clean target root: `/home/agent/zapret2-managed-clean-d8316e6-45bb14ba`, Compose project `zapret-autotune-clean-d8316e6-45bb14ba`, SOCKS TCP/UDP `127.0.0.1:11081`; left running for inspection. The existing target on `11080` was not replaced.
- **Clean bootstrap/evidence:** state was empty before launch (root-owned directory mode `0700`, independent 64-hex secret, `.env` mode `0600`); real `blockcheckw` binary, no mocks or copied config/candidates. Fresh init was observed, but fast bootstrap prevented capturing an exact generation-0 status snapshot. Generation2 committed frontend ID1 and GGC ID2 for exact `rr17---sn-n8v7kn7d.googlevideo.com`; both candidates have valid raw TLS1.2/TLS1.3 evidence, 3/3 each. Actual root config uses the verified layout-safe fake + multidisorder frontend seed. Periodic `lastProbeAt` advanced with failure counter0 and idle recovery.
- **Clean lifecycle gates:** monitor SIGKILL changed host PID522765→525271 inside the same container/PID1, DockerRestartCount0; IDs[1,2] and config hash `57e639c98420be73a8ec4754b20ba975d81aced44215ed547b1ac029c89c14de` unchanged. Graceful stop exited0 in 0.61s, with no worker leaks. Forced recreation (`5af686e2ad94…`, host PID525625) preserved generation2/IDs/hash. Initial post-recreation failure counters1 then2 are historical observations with unproven cause: a fresh-discovery GGC mismatch sample does not prove the actual probe error cause. The next default probe at 07:54 reset the counter to0 with no active recovery. Player API OK and host SOCKS curl to `www.youtube.com` HTTP200; no media playback/SABR claim.
- **Scope/limits:** both older targets' images, container IDs, PIDs, start times and ports stayed unchanged (existing managed PID518298, old SS PID266922); no host-manager or Xray changes. Private final PASS report: `/home/agent/zapret2-managed-clean-d8316e6-45bb14ba/final-cleanrun-report.json`, raw artifacts in `final-evidence/`. This is isolated clean-run acceptance, not blanket production approval, a playback test, or a real 90-minute watchdog wait. The full 90-minute wall-clock expiry was not waited out; deadline behavior remains covered by shortened/injected tests. The working-target entry below describes the unchanged existing target, not the additional clean RC.
- **2026-10-11 container-owned control, committed/deployed:** `5258ae2`, `6467108`, `fe23eec` replace external supervision with PID1-owned runtime-cycle retries (three starts per rolling six hours, cancellable 15s/30s backoff capped at 60s). Guard/config/DB/rollback/cleanup failures remain fail-closed. Reconciliation precedes activation; fresh-shell watchdogs avoid EXIT/TERM trap deadlock; private monitor sessions allow cleanup of orphaned scan workers before restart. SS uses private JSON, never exported password/argv. Graceful SIGTERM joins/restores before stopping owned daemons; Compose has no restart policy and 240s grace, monitor join 45s. Host/container-platform lifecycle belongs to the operator, not a generated host manager. Historical transient acceptance units were unloaded and the tracked systemd unit removed.
- **2026-10-11 strategy absence:** genuine validated no-qualified-strategy results persist `max(2h, configured cooldown)` in SQLite, preserve longer existing deadlines and other leases, and survive monitor/container startup. Lightweight probes continue; heavy saved-first/scan/apply waits until expiry. Other ordinary outcomes retain 30m cooldown. Malformed reports and explicit nfqws2 infrastructure failures are vetoed before both positive and negative selection; they are never disguised as absence. Qualification remains both TLS protocols 3/3. The separate 90-minute whole-job watchdog plus bounded restoration remains abnormal, not normal absence.
- **Verification:** full Go suite, targeted race tests and shell/static/Python fixtures pass. Negative-only checker fault injection demonstrated a persisted 7200-second pause, unchanged checker count 7→7 through internal monitor restart, same container/PID1, DockerRestartCount0, zero candidates and graceful exit0. Active-check crash fixture demonstrated old separate-PGID workers gone before replacement activation, unowned session peer preserved, direct unchanged PID1/container and exit0. Private reports: `/home/agent/container-absence-acceptance-roqcpxq1/actualreport.json`, `/home/agent/active-scan-crash-5z2a6xb_/actual-report.json`. These validate control flow, not new network qualification; real two-hour wall-clock expiry was not waited out (expiry covered by injected-clock tests). Numeric PID signalling retains the usual read/signal PID-reuse race; strict reuse-proof signalling would require pidfds.