- #890: installer at f75ee7d0 re-applied on both nodes with -DenyPaths root, pki, data and
C:\ProgramData\ZB\kek (the Secrets KEK sits outside the install root); second apply 12/12 unchanged.
- Script-runner canary authored in the AdminUI and deployed (Calculation tag + VirtualTag, 1 s timers):
Good and advancing on both endpoints; kept on QA as a runner monitor.
- #794 step 5: the live worker token is restricted (SeChangeNotify only; identity/Everyone/Users/RESTRICTED)
and is denied read and write on pki, data, app, logs and the KEKs; network sockets none. Finding #891.
- #780: QA runaway drill passed (killed at the 4000 ms hard deadline twice, quarantined, siblings Good,
ServiceLevel 250/240 unchanged); drill removed, redeploy back to the canary revision 82dd0598.
- #780 P4 recorded: InProcess stays the default and the rollback; StuckOrphanQuarantineFactor documented
as in-process only.
Claude-Session: https://claude.ai/code/session_01GT67ACmpMR7b9LGRPLijMi
The installer denied NT SERVICE\OtOpcUaScriptWorker only on -InstallRoot. On the wonder layout the node's keys and
secrets sit beside it (E:\ApiInstall\OtOpcUa\pki with the application certificate's private key, data\ with
secrets.db and LocalDb), inheriting BUILTIN\Users:RX. Users is a restricting SID of the worker token, so a worker
could read them.
- Install-ScriptWorkerIsolation.ps1 -DenyPaths <dir>,...: each folder gets the same inherited deny as InstallRoot
(new SensitiveDirDeny steps). Refuses drive/share roots, script-worker\ and the PE cache; folders must exist;
a comma/semicolon string works for powershell -File.
- The deny is read AND write (0xD0157: ReadData, WriteData, AppendData, WriteEA, DeleteChild, WriteAttributes,
Delete, WRITE_DAC, WRITE_OWNER). Traverse/Execute, ReadAttributes, ReadControl and Synchronize stay so the
loader still reaches script-worker\. The write half stops a worker planting a cert in pki\trusted\certs where
a data drive lets Users create files in subfolders.
- The list is recorded under the anchor service's key (Parameters\DenyPaths, step DenyPathRecord): a re-run
removes the deny from folders dropped from the list, -Uninstall removes every recorded one without repeating
-DenyPaths. -WhatIf/-PlanOnly print every step.
- After an apply, WARN for each protected-DACL item under a denied folder that the inherited deny cannot reach;
without -DenyPaths, WARN for pki/data/logs/backup/config/secrets siblings of InstallRoot.
Tests:
- Dry-run harness 46 -> 86 assertions (QA layout plan, mask bits, refusals, stale removal, record-driven
rollback, host account never denied). Mutations (no sibling deny, deny->allow, mask back to ReadData, no
stale removal, rollback ignoring the record, script-worker not refused) each turn it red.
- Test-ScriptWorkerIsolation.Windows.ps1 (Windows PowerShell 5.1, windev): 34/34, incl. -DenyPaths apply,
11/11 unchanged re-apply, shorter-list re-apply, uninstall from the record, second uninstall unchanged.
- New InstallerLayoutIsolationTests (OTOPCUA_ISOLATION_INSTALLER_E2E=1, elevated Windows): a scratch wonder
layout under the real restricted worker token. Before install and after uninstall the token reads pki/data/
logs/backup/appsettings and plants a trust-list file; installed it can do neither, still reads script-worker\,
a real worker launched from the layout runs a script, and the host still reads everything. Mutating the
-DenyPaths denies to a no-op mask turns it red on pki\own\private\otopcua.pfx (reproduces #890).
QA vq03/vq04 are NOT re-applied; the command is in docs/deploy/qa-upgrade-b8b4c18f-to-fea7de0a.md.
Claude-Session: https://claude.ai/code/session_01GT67ACmpMR7b9LGRPLijMi
Profiled the remaining lmxopcua#789 misses on windev, a VMware VM, using per-thread CPU and context-switch accounting
plus ETW CPU sampling.
- One blocking hand-off costs ~26 us of latency and ~18 us of CPU there, about 10x macOS.
- An idle evaluation makes four hand-offs in a row: host writer -> worker reader -> evaluation thread -> host reader
-> dispatcher.
- 48% of host CPU at 10k/s is in the kernel: the per-RunAsync NtWriteFile, thread-pool wakes, reads and waits.
Spin-before-block was measured at every hand-off and is not kept. It meets 150 us idle p50 (118 us) only by keeping
about two host cores and four worker cores busy for one stream. At 10k/s it reads +250 points.
Harness: new "paced 10,000/s, batched" row. It uses the same rate and keys, with each 1 ms tick submitted as one
RunBatch call, which is the production seams' shape. It runs after the original four phases, so their rows stay
comparable. On windev it halves host CPU, to +36..+45 points against +71..+84 unbatched. That is still above 25.
Design doc:
- 12.8: the record.
- 14 #15: the owner's decision on the Windows budgets. Recommended: measure on bare-metal Windows, then state the two
budgets per platform, with the batched row as the gate. The alternative is a shared-memory transport.
Claude-Session: https://claude.ai/code/session_01GT67ACmpMR7b9LGRPLijMi
Measured on windev (3 alternated sessions) and macOS: a pool resume in place
of the dispatcher, and an inline-on-reader upper bound, with the harness as
committed and with its paced continuation made synchronous.
- pool: 10k/s host CPU unchanged, idle p50 and idle CPU worse (macOS idle CPU
x3 from pool spin); only the closed loop gains. Not kept.
- inline (unsafe for every seam; the ceiling for a VirtualTag-only opt-in):
~8 points of a core at 10k/s raw and ~50 us idle p50 on windev with a cheap
continuation, nothing with the harness as committed. Both budgets still
missed by a wide margin, so the opt-in was not built.
No code change. Records the remaining attribution (harness pool hop ~15
points raw; reader wake + syscalls; worker side untouched).
Claude-Session: https://claude.ai/code/session_01GT67AD99zQQqpLCVm3p7xk
§12.1 gains the 2026-11-11 windev (VMware VM) perf table: idle p50 581 us
(MISS), kill latency after T+G 8-16 ms (PASS), kill both shards 2.12-2.14 s
(MISS by ~6 %), worker RSS 160 MB compiled / 36 MB warm-cache (PASS under the
revised 200 MB budget), host CPU +131 % vs InProcess (MISS, untuned).
It also records the first Windows run of Category=ScriptWorker: 107/0/2 from
C:\build, all 13 WorkerSoftClockTests cases (5 freezing a real worker with
NtSuspendProcess) green with no code change. From a user-profile directory
the restricted-identity isolation test fails because the restricted token
cannot read the worker dll there (profile ACL), which is environment, not a
defect.
Claude-Session: https://claude.ai/code/session_01GT67AD99zQQqpLCVm3p7xk
The owner revised the per-shard worker idle RSS budget from 80 MB to 200 MB
+ scripts on 2026-11-11: compiles run in the worker, so Roslyn's code and
metadata stay resident (measured 159-187 MB) and no GC setting moves it.
The perf harness now reports both RSS rows against the 200 MB constant with
a PASS/MISS mark; A-3 records what the measurement showed.
Claude-Session: https://claude.ai/code/session_01GT67AD99zQQqpLCVm3p7xk
Owner decisions §14 #5/#6 as built (design §12.5):
- Scripting:Worker:Identity (default unset, Windows only, fail-closed validation): each worker
starts under a restricted token derived from the host's. Every privilege but SeChangeNotify is
removed; the restricting SIDs are NT SERVICE\OtOpcUaScriptWorker's SID + Everyone + Users +
RESTRICTED; the token and default DACLs are host + SYSTEM + identity + RESTRICTED. The worker
is created suspended, put in its Job Object before it runs, inherits only its stdio pipes, gets
an ACL'd work dir as TEMP, and runs with DOTNET_EnableDiagnostics=0. The identity is granted on
the host's window station and desktop, because USER32 init made coreclr.dll fail with
ERROR_DLL_INIT_FAILED on windev. The logon SID is not used because the host's process DACL
gives it PROCESS_VM_READ.
- The worker pipe on Windows always has an explicit protected DACL: host account + identity
only. There is no owner ACE, so an elevated host's Administrators gets no access.
- The worker treats access denied on its parent-pid poll as alive; the Job Object covers that case.
- scripts/install/Install-ScriptWorkerIsolation.ps1 + ScriptWorkerIsolation.psm1 are idempotent
and support -WhatIf, -Uninstall and -PlanOnly. They create the disabled anchor service, the
PeCacheDir ACL (protected, no Users, key and markers host-only, pe\ Modify for the identity),
read+execute on script-worker\, a ReadData deny on the install root, the outbound firewall
block on the worker exe, and optionally the host-service env.
- Tests: the dry-run harness scripts/install/tests/Test-ScriptWorkerIsolation.ps1 (46
assertions, mutation-checked); ScriptWorkerIdentityTests; WorkerIsolationTests, which cover
the pipe ACL, token access and a real restricted worker running a script, all green on windev.
Claude-Session: https://claude.ai/code/session_01GT67AEHd2Acv1YPwGfXEoV
ForbiddenTypeAnalyzer now denies System.AppContext, System.Console, GCSettings, ProfileOptimization,
EncodingProvider/CodePagesEncodingProvider, TypeDescriptor and CryptoConfig as types, and the ambient
CultureInfo.{DefaultThreadCurrentCulture,DefaultThreadCurrentUICulture,CurrentCulture,CurrentUICulture}
and TimeZoneInfo.ClearCachedData members (member deny-list now covers properties and member-access reads).
Shadow-mode tests diverge via culture-sensitive string.Compare (worker is InvariantGlobalization) instead of
AppContext.BaseDirectory, with a premise test. Mutation-proved: 21 new cases fail with the analyzer reverted.
Claude-Session: https://claude.ai/code/session_01GT67AEHd2Acv1YPwGfXEoV
Shadow on central-1/2 with all three seams: zero mismatches on ordinary VTs / Calc / alarms, counted mismatches on
exactly the divergent (value) and decimal (unsupported_type) scripts, kill -9 of the shadow workers left every
published value Good. OutOfProcess with ScriptedAlarm listed: alarms raise and clear, a runaway predicate is killed
twice and quarantined (BadInternalError) with its siblings Active and Good, kill -9 under an in-flight predicate
freezes it Active + BadResourceUnavailable and it re-evaluates in isolation. Answers A-7 (Modbus UInt16 boxes as
ushort) and A-8 (a decimal VT publishes BadNotSupported in process). Rig restored to revision e221b179.
Claude-Session: https://claude.ai/code/session_01GT67AEHd2Acv1YPwGfXEoV
- IScriptRunner.RunBatch: one EvalBatch frame per worker for every evaluation ready to run; per-item, streamed results.
The default (InProcessScriptRunner) is RunAsync per item. Used by the Calculation pump (dequeue all, split at a
repeated tag, publish in dequeue order), the VirtualTag seam (ScriptBatchCoalescer, no timer) and the scripted-alarm
engine (whole dirty set). In process the pump and the engine stay one at a time, unchanged.
- Exact per-item deadlines: the worker hands work to an idle evaluation thread or defers it (EvalResults
deferred_eval_ids); the host stops that run's clock and restarts T then G when it starts (started_eval_ids). A run
queued behind a runaway is neither timed out nor struck and is a victim, not the culprit, at the kill.
- Scripted-alarm seam behind Scripting:Runner:Seams; unset Seams now means VirtualTag + Calculation, so alarms run out of
process only when listed. Load still compiles each predicate (to PE bytes only) so a broken one fails the load; passes
batch and apply in the serial order under the gate. Script faults freeze the condition BadInternalError; runner faults
freeze it BadResourceUnavailable and re-evaluate on recovery; a dead worker never clears an alarm. IScriptRunnerProvider
carries the runner to DriverHostActor.
- Perf harness: batch throughput (RunBatch vs RunAsync x n) and alarm-pass latency (in process / batched / serial);
design §12.3 records macOS + windev. On windev the serial out-of-process alarm pass misses 250 ms from ~500 alarms;
batched p99 <= 53 ms at 1,000.
Claude-Session: https://claude.ai/code/session_01GT67AEHd2Acv1YPwGfXEoV
#789 (partial, further tuning deferred): combining inline frame writer, buffered
frame reader, one timer per evaluation for T and T+G, slot release on the reader
before the caller resumes, a non-spinning per-connection result dispatcher, and
blocking pipe I/O on Unix (both ends) so the socket never enters the socket
engine; TerminateJobObject kill on Windows. macOS host CPU at 10k evals/s K=2:
228-240 % -> 79 % (InProcess 69 %, IPC share +10 vs budget 25); windev 153 % ->
123 % (+115, still a miss). Harness reports the "vs InProcess" share, warm-cache
worker RSS, and times kill latency from the soft-timeout answer.
#790: the worker sends LoadStarted; the compile timeout is armed then, so a load
queued behind other compiles is never killed or struck.
#788: one Error with stack per down period, rate-limited Warnings with attempt
count, one Information on recovery.
Design doc §8.2 notes + §12.2 record.
Claude-Session: https://claude.ai/code/session_01GT67AEHd2Acv1YPwGfXEoV
- A self-contained Host publish (-r win-x64 --self-contained, the Windows installs) now publishes the
ScriptWorker self-contained for the same RID into script-worker/. The launcher starts its apphost
directly when runtimeconfig.json lists includedFrameworks. A framework-dependent worker (docker-dev)
keeps running through the dotnet muxer.
- docker-dev/docker-compose.script-runner.yml: OutOfProcess for the central pair's VT + Calc seams.
Live drill recorded in design §12.1 (runaway kill + quarantine, kill -9 recovery, open circuit
with /health/ready still 200, metrics).
- ScriptRunnerPerfTests (Category=Perf, opt-in via OTOPCUA_SCRIPT_PERF=1): InProcess vs OutOfProcess
at K=1/2/4. The accepted 5k evals/s/node gate passes on macOS and windev. The host-CPU, RSS and
Windows latency budget misses are tracked as #789.
- Fix: the single-flight slot is released before the awaiting caller resumes. Before, a caller that
started its next run right after the previous one returned could be refused with "previous
evaluation still running", which the seams publish as Bad.
- Fix: the worker compiles on dedicated threads, not the thread pool. A burst of cold loads starved
the heartbeat, and the host killed the shard (found on windev). Load-timeout queueing is #790.
Claude-Session: https://claude.ai/code/session_01GT67AEHd2Acv1YPwGfXEoV
Second P1 slice of docs/plans/2026-11-09-out-of-process-script-runner-design.md; still opt-in behind
Scripting:Runner=OutOfProcess (InProcess unchanged).
- Compiled-PE cache (design §7, §14 #8): Scripting:Worker:PeCacheDir holds <dir>/pe/<id>.pe entries authenticated
with HMAC-SHA256 under a per-node random key (<dir>/pe-cache.key, created by the host, 0600/0700 on Unix), handed to
the worker in Hello. Entry id = SHA-256(compile fingerprint + kind + source); the fingerprint covers the Roslyn
version, the Core.Scripting build (analyzer, wrapper, stack guard), the sandbox reference set and the runtime. Edited,
truncated, other-key or renamed entries are rejected, deleted and recompiled; the worker never loads unverified bytes.
Pruned by recency at startup (30 d / 50 000). ScriptEvaluator splits into CompileToImage + FromImage. Measured on
40 scripts, restart to every first result: macOS 820 ms no cache / 840 ms empty / 85 ms warm;
windev (Windows) 9.5 s / 9.4 s / 1.0 s.
- Quarantine race: WorkerConnection runs exit attribution synchronously BEFORE completing the evaluations/loads it held
(and before Exited), and the shard passes it straight to the runner — a kill strike is in place before the killed
key's slot is released.
- Supervision: a dead worker restarts by itself after its backoff; Unavailable outcomes carry ScriptRunOutcome.Recovery
(-> VirtualTagEvalResult.RetryWhen) and the VT actor / Calculation driver re-evaluate once against current inputs when
it completes. Isolation worker stops after Scripting:Worker:IsolationIdleTimeout (10 min). Keys compiling when a
worker dies are suspects too (isolation finds a poison compile and quarantines it). Startup orphan sweep by launch
marker: kills only a process whose marker's host is gone and whose pid + start time + image match.
- Crash-loop circuit (Scripting:Worker:CircuitOpenAfter, 2 min): runner Unavailable, one Error, evaluations answered
at once while shards keep retrying at the cap.
- Observability (§10): otopcua.script.runner.state, worker.up/inflight/queue_depth/heap_bytes (shard),
ipc.roundtrip + eval.duration (kind), compile{cache}, isolation.suspects, exits reason=idle. `script-runner` health
check on the ready tier, at worst Degraded (§14 #12: degraded but still ready).
Left on #780: self-contained worker publish, installer (account, pipe ACL, firewall), docker-dev live drill,
scripted-alarm seam, the §4.5 per-seam batch submit, the §8 perf harness.
Tests: SupervisionTests, PeCacheStoreTests/PeCacheWorkerTests (real workers), ScriptRunnerHealthCheckTests, VT + Calc
recovery seam tests; 14 mutations each turned their test red.
Claude-Session: https://claude.ai/code/session_01GT67AEHd2Acv1YPwGfXEoV
Reproduced first, out of process (ScriptRecursionContainmentTests re-launches
the xunit v3 test assembly to run one recursive script through
ScriptEvaluator + ScriptRunGuard, so the runner never executes one): before
this change every shape — local function, mutual local functions,
self-referencing Func<>, anonymous self-application, recursion through LINQ,
an expression-bodied void lambda, an anonymous method — ended the child with
exit 134 (SIGABRT, "Stack overflow."). A StackOverflowException cannot be
caught, so in production one script takes the whole node down.
Fix: ScriptEvaluator.Compile now rewrites the validated wrapper
(ScriptStackGuardInjector, after ForbiddenTypeAnalyzer) so the script body and
every lambda, anonymous method and local function starts with
ScriptStackGuard.Check(), which throws ScriptStackExhaustedException when
RuntimeHelpers.TryEnsureSufficientExecutionStack reports the stack nearly
spent. The script then fails like any throwing script (Bad, script log) and
the process lives.
Why a runtime check, not an analyzer rule: recursion is not decidable from
syntax (anonymous self-application never names itself), and bounded recursion
must keep working. Expression-bodied lambdas/local functions become blocks
(`return e;` or `e;`, chosen from the bound symbol, so the same overload
binds); expression-tree and async lambdas are left alone. Limits (documented
on ScriptStackGuard and in docs/VirtualTags.md): depth depends on the thread's
stack size; a recursion that never re-enters script code is not seen; a script
may catch the exception, which only unwinds it.
Tests: ScriptRecursionContainmentTests (7 shapes, child process each; all 7
fail with exit 134 when the rewrite is disabled), ScriptStackGuardInjectionTests
(bounded recursion and every rewritten lambda/local-function shape keep their
value; the expression-tree case fails if those lambdas are rewritten).
Claude-Session: https://claude.ai/code/session_01GT67AEHd2Acv1YPwGfXEoV
VirtualTagActor called the blocking IVirtualTagEvaluator.Evaluate, and
RoslynVirtualTagEvaluator blocked on the run guard with
.GetAwaiter().GetResult(), so every script run parked a dispatcher thread for
up to its 2 s budget (a first-use Roslyn compile did too). This is the #750 P0
refactor for virtual tags.
- IVirtualTagEvaluator.EvaluateAsync (default: a completed task over the
snapshot overload, so the null evaluator and fakes are unchanged).
- RoslynVirtualTagEvaluator implements it without blocking: passthrough and
compile failures complete synchronously, a first-use compile goes to the
pool, the run is awaited. The blocking overload now delegates to it.
- VirtualTagActor: a completed task is applied in the same turn, as before;
an incomplete one switches to an "evaluating" behaviour that stashes every
dependency/reassert message until a private completion message arrives,
then applies the result and unstashes. Each trigger is still evaluated
against its own inputs, results are applied/published in trigger order and
the dedup/Bad-transition state sees the same sequence. The dependency map is
not touched while a run reads it.
Calculation and scripted-alarm serialisation is NOT relaxed: neither blocks a
thread, and running independent scripts in parallel is not semantics-neutral
(pool consumption bound, cross-tag publish order, timer/pump exclusion for
Calc; state-machine/persist boundary, transition timestamps and the
store-failure path for alarms). Recorded in docs/VirtualTags.md and the #750
design's P0 entry.
Tests: VirtualTagActorAsyncEvaluationTests — on a one-thread dispatcher a slow
script no longer starves a sibling actor and the slow actor answers Identify
(fails when the actor calls the blocking Evaluate); triggers during a run are
evaluated afterwards in order against their own inputs; a reassert during a
run is applied after it; a faulted evaluation takes the evaluator-threw path
(the ordering/reassert tests fail with the stash removed).
RoslynVirtualTagEvaluatorAsyncTests — a slow script and a cold compile return
an incomplete task (fail without the override); cheap outcomes and timeouts
match the blocking overload.
Claude-Session: https://claude.ai/code/session_01GT67AEHd2Acv1YPwGfXEoV
Design only. Documents why today's in-process ScriptRunGuard can abandon but
not stop a runaway script (file:line evidence), and proposes a sharded
worker pool with an isolation worker under a ScriptRunnerActor, a
named-pipe + protobuf frame protocol carrying full input snapshots,
soft/hard deadlines with process kill, kill attribution and quarantine,
worker sandboxing (scrubbed env, pinned assemblies, Job Object, heap
limit), an HMAC'd content-addressed PE cache, performance budgets,
failure-quality semantics, a flag-gated InProcess/Shadow/OutOfProcess
rollout, a test plan, open decisions and an assumptions register.
Claude-Session: https://claude.ai/code/session_01GT67AEHd2Acv1YPwGfXEoV