Files
lmxopcua/docs/Uns.md
T
Joseph Doherty b8208b3312 test+docs(v3-batch4-wp5): 2-node dual-namespace harness tests + address-space docs
Tests:
- OpcUaServer.IntegrationTests/DualNamespaceAddressSpaceTests.cs (NEW, over-the-wire,
  offline-safe): both namespace URIs registered + distinct; Raw + UNS subtrees browse
  and read; UNS variable Organizes-references its raw node; single-source fan-out parity
  (identical value/quality/timestamp on both NodeIds); HistoryRead via either NodeId ->
  GoodNoData under the shared tagname; WriteOperate gate symmetric across both NodeIds.
- Host.IntegrationTests/EquipmentNamespaceMaterializationTests.cs (extended): full deploy
  -> persisted-artifact -> ParseComposition round-trip carrying both realms, sealing across
  the redundant 2-node cluster (redundancy non-interference). In-memory harness, offline.

Docs (dual-namespace reality):
- CLAUDE.md: new "v3 OPC UA Address Space (Batch 4)" section + Batch-4 testing paragraph.
- docs/Uns.md: address-space projection (two namespaces, Organizes edge, effective-name leaf).
- docs/Historian.md: dual-registration (both NodeIds -> one tagname); updated CLI examples.
- docs/ScriptedAlarms.md + docs/AlarmTracking.md: multi-notifier fan-out, ConditionId=RawPath.
- docs/ScriptEditor.md: dual-namespace clarification (script tag-path semantics unchanged).

Claude-Session: https://claude.ai/code/session_01LVneM3eh1UtJxEisFXgmox
2026-07-16 13:11:34 -04:00

21 KiB
Raw Blame History

UNS — Global Unified-Namespace Management

The UNS page (/uns in the AdminUI) is the single surface for managing the Unified Namespace across the whole fleet. It replaces the old per-cluster UNS, Equipment, and Tags tabs and the standalone virtual-tags list — those have been removed; everything now lives in one global master tree.

The tree

The page shows every layer of the UNS as one expandable tree. Equipment is a leaf in the tree — tags, virtual tags, and alarms live on a dedicated equipment page, not inline:

Enterprise            (read-only grouping — ServerCluster.Enterprise)
└─ Site / Cluster     (read-only grouping — a ServerCluster row)
   └─ Area            (editable — UnsArea)
      └─ Line         (editable — UnsLine)
         └─ Equipment (leaf — Equipment; opens its own page)

Enterprise and Site/Cluster can be deleted from the tree (see Cluster and Enterprise delete below). Other than deletion, they are configured on the Clusters pages (/clusters) — on a cluster row the ⚙ settings link jumps to that cluster. Editable UNS entities start at Area.

Count badges next to a node show how many direct children it has (for equipment, the combined tag + virtual-tag count).

Navigating

  • Expand all / Collapse all toggle the structural levels.
  • Filter by name does a case-insensitive substring match on the names of a node's direct children.

Creating and editing

Every editable row has inline actions:

Node Actions
Cluster + Area, Delete
Area + Line, Edit, Delete
Line + Equipment, Edit, Delete
Equipment Open, Delete

A + Child action pre-fills the parent for you (e.g. + Line on an area opens the Line modal with that area already selected). Build a branch top-down: Area → Line → Equipment. Tags, virtual tags, and scripted alarms are managed on the equipment page (see below).

Equipment page (/uns/equipment/{id})

Clicking Open on an Equipment row (or + Equipment on a Line) navigates to the equipment page. The page is organised into four tabs:

Tab Content
Details Equipment identity — name, description, OPC-40010 namespace fields, driver binding, and served-by cluster. (This is the former EquipmentModal, now a full page.)
Tags Equipment-bound driver tags. Each tag uses the driver-typed config editor (same editors as before — Modbus, S7, AB CIP, etc.) with the same client-side validation.
Virtual Tags Virtual tags driven by C# scripts. The inline Monaco script editor with Roslyn IntelliSense, completions, and live diagnostics is available here per tag.
Alarms Scripted-alarm definitions bound to this equipment. Create, edit, or delete predicates here; the Monaco editor is available for each predicate script.

"Add equipment" under a Line uses the URL /uns/equipment/new?lineId=..., pre-filling the parent line. Saving redirects to the new equipment's page.

Served-by cluster

An area's cluster assignment is its "served-by" cluster — the cluster node that runs it. It's set when you create the area (under a cluster) and changed by editing the area's cluster in the Area modal, which moves the whole branch. There is no separate "served-by" concept and no migration — it is simply UnsArea.ClusterId.

Tags — reference-only (v3)

v3 (Batch 3): UNS equipment no longer authors or binds tags. Device I/O is authored once in the Raw project tree (/raw — see Raw.md); an equipment's Tags tab holds references to those raw tags. The old driver-bound Tag modal on this tab is retired.

The Tags tab is a list of UnsTagReference rows. Each row shows the effective name, the backing tag's RawPath, its inherited DataType and AccessLevel (read-only — they come from the raw tag), and an editable display-name override. The effective name is the override else the raw tag's Name.

"+ Add reference" opens a raw-tree picker (the /raw tree in picker mode) scoped to the equipment's cluster — cross-cluster tags are structurally unreachable, and the service also enforces tag.cluster == equipment.cluster on commit. Tick individual tag leaves, or use a device / tag-group's "Select all tags below" menu to pull many at once.

Effective-name uniqueness: within an equipment the effective name must be unique across references, VirtualTags, and ScriptedAlarms (they share the equipment's UNS NodeId space, {EquipmentId}/{EffectiveName}). This is enforced both at authoring (a readable rejection naming the colliding source) and at the deploy gate (DraftValidator, UnsEffectiveNameCollision) — the deploy gate is what catches rename-induced collisions a raw rename produced after the reference was authored, naming both colliding sources and the equipment.

Removing a raw tag in /raw is blocked while a UnsTagReference points at it (the error names the referencing equipment); renaming a raw tag or any ancestor warns when a beneath-it tag is historized (no historianTagname override), UNS-referenced, or named by a script literal — see Raw.md.

Galaxy / AVEVA System Platform points are now ordinary raw tags on a GalaxyMxGateway driver in /raw, referenced into equipment like any other raw tag. There is no separate alias concept or SystemPlatform-kind namespace.

OPC UA address-space projection (v3 Batch 4 — dual namespace)

v3 (Batch 4): the server now exposes the address space under two OPC UA namespaces instead of the old single https://zb.com/otopcua/ns:

Namespace URI Subtree NodeId s= scheme
https://zb.com/otopcua/raw the /raw device tree (Folder → Driver → Device → TagGroup → Tag) the node's RawPath (e.g. Plant/Modbus/dev1/Speed)
https://zb.com/otopcua/uns the UNS tree (Area → Line → Equipment → signal) the slash-joined Area/Line/Equipment/EffectiveName

Every device value has exactly one source — the raw tag's node in the Raw namespace. A UnsTagReference projects that raw tag into an equipment as a UNS-namespace variable whose NodeId leaf is the reference's effective name (the display-name override else the raw tag's Name). The UNS variable does not bind a driver of its own: it carries an Organizes reference to its backing raw node and mirrors it. A single driver publish for a RawPath fans out to the raw NodeId AND every referencing UNS NodeId with identical value / quality / source-timestamp — the two NodeIds never drift.

  • Reads / subscriptions work through either NodeId and return the same data.
  • Writes route through either NodeId to the same backing driver ref under the same WriteOperate gating (a UNS write is neither more nor less privileged than the raw write it fans from); a failed device write reverts both NodeIds via the shared fan-out.
  • HistoryRead works through either NodeId and returns the same series — both NodeIds register the same historian tagname (see Historian.md).
  • Native alarms materialize once at the raw tag (ConditionId = RawPath) and fan via SDK notifiers to the raw device folder AND every referencing equipment folder — see ScriptedAlarms.md / AlarmTracking.md.

Note the two distinct identity strings: the wire NodeId is the path Area/Line/Equipment/EffectiveName, while the effective-name uniqueness key above is {EquipmentId}/{EffectiveName} (the logical per-equipment collision space the guards enforce). They are related but not the same string.

Virtual tags

A virtual tag is bound to an equipment and driven by a script (no driver). Add and edit virtual tags on the equipment page's Virtual Tags tab; the data type is chosen from the standard OPC UA type list and the Monaco script editor is available inline. Scripts may read the equipment's references relative-to-equipment with ctx.GetTag("{{equip}}/<RefName>") — the token resolves through the equipment's UnsTagReference rows (by effective name) to the backing raw tag's RawPath at deploy; an unresolved <RefName> is a deploy error and a live Monaco diagnostic (editor accepts ⇔ publish accepts). See ScriptEditor.md.

Galaxy tags

GalaxyMxGateway is an Equipment-kind driver — Galaxy points are ordinary equipment tags authored through the standard Tag modal, exactly like Modbus or S7 tags. The Galaxy reference is stored as TagConfig.FullName (tag_name.AttributeName). The Galaxy address picker on the Tags tab lets you browse the live Galaxy hierarchy to select an attribute; after selecting, set the tag Name, DataType, and AccessLevel (default: read-only). There is no alias concept, no SystemPlatform-kind namespace, and no relay→alias converter.

Typed TagConfig editors

The Tag modal dispatches a driver-typed config editor for the following drivers:

Driver Fields in the typed editor
Modbus Register type, address, data type, word order, etc.
S7 Data block, offset, data type, etc.
AB CIP Tag path, data type, etc.
AB Legacy (DF1/DH+) Address, data type, etc.
TwinCAT Symbol path, data type, etc.
FOCAS PMC address, data type, etc.
OpcUaClient FullName (the remote OPC UA node id string)

OpcUaClient was previously raw-JSON fallback only; it now has a first-class typed editor that exposes a single FullName field (PascalCase JSON key, consistent with the Galaxy editor convention). It is registered in TagConfigEditorMap and TagConfigValidator; unknown keys in the stored JSON blob are preserved on round-trip.

The historian backend is the external HistorianGateway (no OtOpcUa-side tag driver / tag-config editor). See Historian.md.

Drivers not yet listed above (e.g. Galaxy — which uses the Galaxy address picker described below) still use the generic raw-TagConfig-JSON textarea.

"Build address" pickers in protocol-driver editors

The Modbus, S7, AB CIP, AB Legacy, TwinCAT, and FOCAS typed editors include a Build address button. Clicking it opens the driver's existing address-builder UI inside the shared DriverTagPicker overlay; confirming a selection writes the fully constructed address string back into the editor's address field. This means you can visually compose a register reference (e.g. select "Holding Register → 100 → Int16") and have it serialised into the correct JSON without hand-editing.

Historizing tags (first-class controls)

The Tag modal exposes Historize this tag and Historian tagname (override) as explicit controls that work for all drivers — including protocol drivers (Modbus, S7, etc.) that use the typed editor and raw-JSON drivers (Galaxy) alike.

Control JSON key Type Behaviour
Historize this tag checkbox isHistorized bool When checked, the OPC UA node materialises with Historizing=true and the HistoryRead AccessLevel bit set.
Historian tagname (override) textbox historianTagname string (optional) Explicit tagname the historian backend will query. When left blank, the server defaults to the tag's driver FullName.

These fields are merged into the TagConfig JSON blob via the pure TagHistorizeConfig helper, which preserves all other keys byte-stable. The server's OPC UA HistoryRead dispatch already consumes these keys from the TagConfig blob — see Historian.md for the full server behaviour, continuation-point paging, and aggregates.

Note — native-alarm HistorizeToAveva: a tag that carries a native "alarm" object has a separate opt-out field alarm.historizeToAveva (a checkbox labelled "Historize to AVEVA" on the Tag modal's alarm section). That field controls whether the alarm's transition events are written to the AVEVA historian — it does not affect tag-value history (which is controlled by isHistorized). See ScriptedAlarms.md §Native driver alarms for details.

Array tags (1-D)

A tag can be made a 1-D OPC UA array node by setting two keys in its TagConfig JSON blob. The controls are exposed as first-class UI fields in the Tag modal (an Is array checkbox + an Array length numeric input), available for all drivers — typed editors (Modbus, S7, etc.) and the raw-JSON textarea (Galaxy) alike.

Canonical rule

Condition OPC UA node shape
isArray: true AND arrayLength >= 1 1-D array node (ValueRank = OneDimension, ArrayDimensions = [arrayLength])
isArray: false (any arrayLength) Scalar node (default)
isArray absent Scalar node (default)

A single-element array (isArray: true, arrayLength: 1) is valid and materialises as a [1] node.

Fields

Field Type Required Description
isArray bool no When true (and arrayLength >= 1), makes the node a 1-D array. Absent or false → scalar.
arrayLength uint (≥ 1) when isArray: true The element count for the OPC UA ArrayDimensions declaration. Must be ≥ 1 when isArray is set.

UI validation rejects arrayLength = 0 when isArray is checked.

Examples

{"FullName":"PLC1.TemperatureArray","isArray":true,"arrayLength":10}
{"Register":"HR100","DataType":"Float32","isArray":true,"arrayLength":5}

Combined with historization (values are arrays — history of the whole array snapshot):

{"Register":"HR200","DataType":"Int16","isArray":true,"arrayLength":20,"isHistorized":true}

Per-driver read mechanism and live-verify status

Driver Read mechanism Live-verify
Modbus Contiguous FC03/FC04 block (arrayLength × registers-per-element); String and BitInRegister array modes also supported Mac-verifiable (sim 10.100.0.35:5020)
S7 Contiguous ReadBytesAsync block over the declared address span + per-element decode loop Unit-proven (sim fixture down)
AB CIP libplctag native array read (atomic and UDT member arrays) Unit-proven (sim fixture down)
AB Legacy PCCC multi-element file read via libplctag (cap 256 elements) Unit-proven (sim fixture down)
TwinCAT ADS native array symbol read against the declared SymbolPath Unit-proven (sim fixture down)

Out of scope (named deferrals)

  • Array writes (inbound client→device write of an array value) — tagged for a follow-up phase.
  • Multi-dimensional arrays (ValueRank > 1) — not supported; all arrays are 1-D.
  • Array historization — a historized array tag materialises with the correct Historizing flag, but the historian backend treats the value as an opaque blob; per-element history is out of scope. (Continuous historization is numeric-analog only — array / non-numeric values are not recorded.)

See the individual driver docs under docs/drivers/ for per-driver implementation details.

Writable equipment tags (writable key)

An equipment tag is writable by default — the materialised OPC UA node is created CurrentReadWrite and inbound client writes are routed to the device (subject to the WriteOperate LDAP role gate). Setting the optional "writable": false key in a tag's TagConfig makes the node read-only at the driver seam. The six typed tag editors (Modbus, S7, AB CIP, AB Legacy, TwinCAT, FOCAS) expose this as a Writable checkbox (checked by default); it can also be hand-authored in the raw-JSON textarea.

Condition Node
writable absent Writable (driver default) — existing configs are unchanged
writable: false Read-only
writable: true Writable

The key is honoured by Modbus, S7, AB CIP, AB Legacy, and TwinCAT. FOCAS forces read-only regardless — FOCAS PMC writes are unsupported in the wire client, so its editor shows the checkbox disabled/unchecked and any authored writable:true is ignored (and flagged at the deploy gate below). Node-level authz (LDAP WriteOperate) remains the effective write gate above this flag.

Deploy-time TagConfig validation (Deployment:TagConfigValidationMode)

The deploy gate inspects every equipment tag's TagConfig per its bound driver type and surfaces strictness warnings — a present-but-invalid enum value (e.g. a typo'd dataType, which the lenient runtime silently defaults to the wrong width with Good status), a structurally-unparseable TagConfig, or (FOCAS) an authored writable:true.

The behaviour is controlled by Deployment:TagConfigValidationMode in appsettings.json:

Mode Behaviour
Warn (default) Non-blocking: warnings are logged and appended to the deployment result message; the deploy still succeeds. Existing configs never break.
Error Blocking: warnings fold into the draft-gate reject list — a config with a typo'd enum (or unparseable TagConfig) cannot be re-deployed until fixed. Opt-in per environment once the fleet deploys warning-clean.

Running servers are untouched either way — the gate only sees re-deploys. Unmapped drivers (Galaxy, OpcUaClient) are skipped. Unknown keys are never warned on: the TagConfig blob is a shared namespace (platform intent keys coexist with per-driver address keys and forward-compat keys the typed editors preserve), so there is no single schema to validate against — parsers read known fields and ignore the rest by design.

Note (runtime strictness is deferred). In this release the runtime parsers stay lenient (a typo'd enum still silently defaults) — only the deploy gate reports it. Flipping the runtime parsers to strict (a typo'd enum ⇒ BadNodeIdUnknown instead of a wrong-width Good) is a deliberate future phase, to be taken only after fleets have run Error mode clean. Tracked in archreview/plans/STATUS.md.

Galaxy address picker — native-alarm pre-fill

When the Galaxy address picker selects an attribute that is itself an alarm (IsAlarm == true in the Galaxy hierarchy), the Tag modal automatically seeds a default alarm object in the tag config:

{"alarmType":"OffNormalAlarm","severity":700}

This lets the operator author the native alarm in a single picker pass without hand-editing JSON. The pre-fill never overwrites an alarm object that is already present — if the tag already has a custom alarm section, the picker leaves it untouched.

Cluster and Enterprise delete

Cluster delete

A Delete action is available on Cluster rows in the UNS tree. The server refuses the delete if the cluster still has any Areas (children) — the same refuse-if-children guard used by Area and Line delete. Remove all Areas (and their descendant Lines/Equipment) first, then delete the cluster.

No RowVersion concurrency check. ServerCluster does not carry a concurrency token, so the delete does not have the last-writer-wins protection that Area/Line/Equipment deletes have. A follow-up migration will add the token; for now, coordinate cluster deletes manually.

Enterprise delete

An Enterprise row is a read-only grouping label (the Enterprise column of ServerCluster) — it is not a separate entity. Deleting an Enterprise row deletes all clusters whose Enterprise matches that label, all-or-nothing:

  • If any cluster under the enterprise still has children, the entire delete is refused and no clusters are removed.
  • If all clusters are empty, every cluster under that enterprise is deleted in a single transaction.

Remove all Areas under every cluster in the enterprise first, then delete the enterprise label.

Create-new-script inline (virtual-tag panel)

On the equipment page's Virtual Tags tab, when a virtual tag is not yet bound to any script, the inline script panel shows a Create new script button. Clicking it:

  1. Generates a new blank Script record with an auto-generated SC-… id.
  2. Binds the virtual tag to that script.
  3. Expands the Monaco editor inline so you can begin authoring immediately.

This removes the previous two-step flow (create a script on a separate page, then attach it to the virtual tag); the entire lifecycle now lives on the equipment page.

Hosts page per-driver-instance rows (deferred)

Phase 6 did not implement per-driver-instance status rows on the Hosts page. This item (H7-runtime) is F7-runtime-blocked: the runtime plumbing needed to surface per-instance health rows is not yet in place. It remains on the backlog.

Bulk import

Import equipment CSV (toolbar) bulk-creates equipment across many lines and clusters in one pass. After an import the whole tree reloads.

Applying changes

Edits here change the configuration only. As the page header notes, changes apply on the next deployment — run a Deploy (Deployments page) to push them into the running address space.

See also