docs: record final managed RC field acceptance and remaining gates

Record supplied fd25cb8 managed-image watch, representative post-warmup media transfers and degraded recovery continuation without claiming avatar or durable family qualification. Require explicit Dockerfile.target and positive entrypoint/marker checks before image promotion; preserve the user README badge.

Co-authored-by: GPT-6 <gpt-6@openai.com>
This commit is contained in:
ZeusinaandGPT-6 Bot committed 2026-10-12 23:34:38 +03:00
1 parent fd25cb8939
commit 9984f21466
3 files changed
+111 -24

No files matched your search

+35 -19
View File
@@ -1,4 +1,4 @@
# Fresh managed-test target contract (not deployed)
# Fresh managed-test target contract (isolated acceptance, not production approval)
`managed init|reconcile|status|active-ids|apply` is exclusively for a **new dedicated container**
which owns its entire config. The existing/dev/production container must refuse
@@ -27,8 +27,9 @@ deliberately forging a provisioned runtime.
The managed `config` is a standalone **full config only in this fresh volume**.
Never replace, adopt, or symlink over `/opt/zapret2/config` or any existing
service config. The new deployment entrypoint must ensure the base daemon
actually honors the exact `ZAPRET_CONFIG` path; compatibility remains unproven.
service config. The deployment entrypoint must ensure the base daemon actually
honors the exact `ZAPRET_CONFIG` path; isolated RC acceptance does not waive
that check for a new image.
Existing services would require a separately designed managed **include
fragment**, not this full-config adapter. No such integration is implemented.
@@ -104,15 +105,15 @@ Directories must not be writable by untrusted processes; hostile concurrent
filesystem mutation is outside the transaction boundary. Shell config safety
and exact GGC filtering are unchanged.
The isolated [Compose definition](deploy/managed-target/README.md) now supplies
the marker, mount and ordered startup; it has not been built or deployed.
No Docker socket, remote/container operations, old target paths, Xray edits,
or host firewall provisioning are included. Before deployment,
The isolated [Compose definition](deploy/managed-target/README.md) supplies
the marker, mount and ordered startup; isolated deployments are recorded below.
No Docker socket, old target paths, Xray edits or host firewall provisioning
are included. Before each new deployment,
provisioning must verify the pinned image honors `ZAPRET_CONFIG`, supports all
listed variables, and starts ss-local while the baseline leaves nfqws2 disabled.
Start the fresh target with that baseline, then verify active apply installs
target-local nft NFQUEUE rules and starts nfqws2, and rollback restores both.
These image/startup/routing gates remain unverified: if any is unsupported or
If any image/startup/routing requirement is unsupported or
uncertain, deployment must fail closed, not substitute `restart-daemons` or
alter an old config. Persistent volume durability is a provisioning
requirement; mountinfo alone cannot establish the storage lifecycle.
@@ -131,9 +132,9 @@ Cancelled/failed downtime restores the committed configuration and runs fixed
target `start` with an independent 30-second cleanup deadline. Failed cleanup or
rollback is fatal to the supervisor; unresolved journal state is never started.
The callback caller must hold the monitor singleton; the wiring repeats runtime
identity/SOCKS guards. Fake-only checks are added, but tests/build are deferred
to the orchestrator. `managed monitor` now binds those callbacks and only the
dedicated target entrypoint starts it; no live/runtime validation has occurred.
identity/SOCKS guards. `managed monitor` binds those callbacks and only the
dedicated target entrypoint starts it; local and isolated runtime validation
results, with their limits, are recorded in [PROGRESS.md](PROGRESS.md).
## Dedicated monitor integration (isolated clean-run acceptance passed)
@@ -211,6 +212,8 @@ Example (replace registry and release; never reuse a release tag):
```sh
IMAGE=registry.example/team/zapret-autotune-managed-target:2026-10-10-r1
docker build --pull -f Dockerfile.target -t "$IMAGE" .
test "$(docker image inspect --format '{{json .Config.Entrypoint}}' "$IMAGE")" = '["/bin/sh","/opt/managed-target-entrypoint.sh"]'
docker run --rm --network none --entrypoint /bin/sh "$IMAGE" -ec 'test "$(cat /etc/zapret-autotune/target-id)" = zapret-autotune-managed-v1; test -f /opt/managed-target-entrypoint.sh'
docker push "$IMAGE"
docker image inspect --format '{{index .RepoDigests 0}}' "$IMAGE"
```
@@ -230,6 +233,11 @@ docker compose --env-file deploy/managed-target/.env -f deploy/managed-target/co
docker compose --env-file deploy/managed-target/.env -f deploy/managed-target/compose.yaml pull target
```
Do not omit `-f Dockerfile.target`: root `Dockerfile` builds the proposal/prototype,
not the PID-1-managed target. Assert both entrypoint and provisioned marker on the
exact selected image before promotion; the assertions above are image-contract
checks, not live resource qualification. A managed-looking tag is insufficient.
Start the selected container image with Compose `up --no-build`;
no host service installation is required. Keep the selected digest and
private configuration together for upgrades and rollback. Never run a build
@@ -453,13 +461,21 @@ is not image-transfer qualification or a guarantee of full-family coverage.
`googlevideo_scope` to its JSON. An experimental Googlevideo profile can be
covered yet qualification-unresolved; one image profile does not establish full
asset-inventory success. Status is read-only and reports live health NOT CHECKED.
Reported `a60f62d` field checks passed restart/recreation and status coverage checks.
A live upgrade hit runner `BrokenPipe` and automatically rolled back to the same
last-good image; successful autonomous media recovery remains unaccepted. Fresh
Range media passed once on TLS1.2, then failed on repeat; no TLS1.3 media PASS is
claimed, and the field failure's cause is unknown. Reported watch metadata had an
allowed avatar URL absent from the homepage, but this new conditional fallback
is source-only and neither field-deployed nor avatar-image-qualified. See
[PROGRESS.md](PROGRESS.md) for the remaining gates. Generic production
Historical `a60f62d` restart/recreation and status checks passed, but its upgrade
hit `BrokenPipe` and rolled back; that run did not accept media recovery. Later
supplied final field results for source `fd25cb8`, managed image
`sha256:73e955e7fdc6efaef88c62ebe4f7afeab71c29d29ba3c88b2a2b7547bb647da5`
on SOCKS11082 completed a 15-minute watch with restart count0 and a live monitor.
Validated rejection/no-qualified-strategy recovery continued with naturally
advanced cooldown and unchanged experimental profiles; it did not repair avatars.
Three fresh media hosts across two roots passed post-warmup 3/3 on each TLS
protocol (18 actual HTTP206/65536-byte MP4 transfers). This is field evidence,
not persisted family qualification, a cold/all-host claim or candidate promotion.
Watch fallback discovered allowed avatar metadata, but both-TLS curl28 transfers
had zero samples: avatar evidence/profile qualification is still absent and
overall health is **DEGRADED / NOT READY**. Automatic durable multi-host family
qualification remains unimplemented. See [PROGRESS.md](PROGRESS.md) for exact
hosts, separate warmup results, timestamps, private proof and remaining gates.
No new network checks accompany this documentation. Generic production
media behavior and signed-URL/SS secret privacy remain unchanged, as do the
persisted minimum two-hour strategy-absence pause and 90-minute whole-job budget.