IMD SolutionComputing LibraryShop main store

Troubleshooting

Citect SCADA connector to Historian shows disconnected

Edge computing and IIoT Published 6 min read

Short answer

A "Disconnected" destination status means the Citect SCADA Connector has no session to Historian and is buffering samples to local disk. SMC tag-discovery and write entries prove the connector's stages, not that values were archived. Stop the connector, verify the destination instance, services and write authorization, reconnect once, and close only when a query returns a sample and the buffer drains.

On this page
  1. What "Disconnected" reports, and what it does not
  2. Four stages, one fault
  3. Working the case: Disconnected on a shared PC, SMC write logged, query empty
  4. Fix the destination entry first
  5. Then services and write authorization
  6. Reconnect once, and leave the buffer alone
  7. Prove recovery with a tag that is actually changing
  8. Questions asked

A "Disconnected" destination status on the Citect SCADA Connector means one thing: the connector's session to the Historian destination is down, while the connector keeps reading Citect tags and storing every sample to local disk. Your data is not being thrown away. The SMC entries showing variable and trend tags found, then a write to Historian, report the connector's own discovery and write stages; they are not proof that the archive holds those values, which is why your query comes back empty. Fix it in this order: stop the connector, correct the destination entry so it names the Historian instance you actually query, confirm the Historian services and the connector's write authorization, reconnect once, and close the fault only when a query returns a fresh sample and the local store drains.

What "Disconnected" reports, and what it does not #

The status on the connector's runtime panel describes the destination connection only. The same status on that panel is documented with Citect Connector V3.0 against Historian 2020, so trust the panel wording for the connector version you run. Two things around it are working: the connector sees your Citect tags, and it is writing samples to local disk because the destination path is unavailable. Every report of this fault shows the same pairing of an offline connection with data stored locally, so treat the buffering as the connector's behavior while the destination is unreachable, not a second fault.

The SMC log then adds a stage the panel does not show: variable and trend tags found, and a final entry that the connector wrote to Historian. That entry is the connector reporting its own write processing. It cannot tell you whether Historian accepted the samples into the instance you query, and a live value on the Citect screen proves less still: a Historian PC can host Citect Connector V3.0, but the value on the screen and the archived sample are different objects on different paths.

Four stages, one fault #

Treat the path as four stages, each proven only by its own evidence:

  • Source discovery — SMC listing variable and trend tags proves the connector sees the Citect tags.
  • Destination session — the connector's connection to the configured Historian endpoint. "Disconnected" fails this stage.
  • Tag binding and write — each discovered tag mapped to a destination identity and written. The SMC entry claims this; it does not prove it.
  • Archive visibility — a query against the intended instance, by the exact destination tag name, over a time range that includes the sample time. Only this proves the data was archived.

Your fault sits at the second stage, and the empty query leaves the fourth unproven. Everything that follows is about the destination session; leave tag configuration alone while the status reads Disconnected.

Working the case: Disconnected on a shared PC, SMC write logged, query empty #

The inputs, each with its source:

  • Destination status: "Disconnected", read on the connector's runtime panel — your report; the same panel status is documented for Citect Connector V3.0 against Historian 2020.
  • Delivery behavior: samples stored to local disk while the status stays Disconnected — your report; independent reports of this fault show the same buffering.
  • SMC log: variable and trend tags found, then a final write to Historian — your report.
  • Query result: no values for those tags in the instance you are querying — your report.
  • Installation: Citect SCADA and AVEVA Historian on one PC — your report; a Historian PC can host Citect Connector V3.0, so co-location is supported.

The rule in words: every observation proves exactly one of the four stages, archival is proven only by archive visibility, and recovery is closed only when three independent outcomes hold together — the destination status reads Connected, the local store has stopped growing and is draining, and a query returns a fresh timestamped sample under the exact destination tag name.

The elimination, stage by stage:

  1. Source discovery: proven by the SMC tag list. One stage proven.
  2. Destination session: failed — the panel reads Disconnected.
  3. Tag binding and write: claimed by the SMC entry, unproven against the archive.
  4. Archive visibility: failed — the query returns nothing.

Closure outcomes met: 0 of 3. The status is not Connected, the local store is still filling, and the query is empty.

Result: the fault is isolated to the connector-to-Historian session, with the write claim unverified and nothing on the source side needing work. Decision: with 0 of 3 outcomes met, run the three corrections below, one change per reconnect, and stop changing anything the moment the count reaches 3 of 3. Below that count the fault stays open, whatever the SMC log says.

Fix the destination entry first #

Read the connector's destination configuration and record exactly what it names: host, instance and authentication method. Compare that with the Historian instance your query targets. Co-location does not make them the same object: a destination entry naming a different instance, or a different form of the host, fails on one PC exactly as it would across two, and it produces precisely your combination of symptoms — offline connection, local storage, write entries, empty query.

Correct the entry to the intended Historian endpoint, and resolve that endpoint from the Historian configuration rather than guessing a port or instance identifier. A documented resolution of this exact symptom starts by stopping the Citect SCADA Connector before the destination change is made; changing the destination under a running connector leaves you comparing states you did not capture.

Then services and write authorization #

With the endpoint verified, check three things in the service-management interfaces, not on the SCADA screen:

  • The Historian services on the shared PC are running.
  • The connector's own runtime or service is active.
  • The connector's configured account authenticates to that endpoint and is authorized to write to it.

A running Citect runtime proves none of these. A connector service can run under an identity Historian refuses, and the panel will still say Disconnected. Fix the endpoint first, the identity second, and make one reconnect per change so each reconnect tests exactly one correction.

Reconnect once, and leave the buffer alone #

After the destination entry is corrected, reconnect through the supported procedure for your version and watch two things: the panel status and the local store.

  • If the status turns Connected and the local store stops growing and starts draining, writes are being accepted; let the backlog deliver.
  • If the status turns Connected but the local store keeps growing, writes are being rejected. Go to the per-tag write and binding diagnostics instead of restarting again — the session is healthy at that point and the fault has moved to stage three.

The backlog reaches Historian as timestamped history, the same shape of data handled when backfilling trend history from a plant historian, so queued samples appear at their original times rather than at delivery time. Do not delete or clean up the local store while diagnosing: it is the only copy of the undelivered samples.

Prove recovery with a tag that is actually changing #

Close on evidence, not on the status line. Pick one tag that is changing during the test, record its source name and a timestamp, confirm its write result in the connector diagnostics, then query the exact destination tag name in the same instance over a range that includes that timestamp. Do not judge the connection on a tag sitting at a constant value: a flat tag can stay empty in the query while the whole path is healthy. Use a changing tag, or force one value change on a test tag, before you judge the connection.

If the connector shows Connected but the query itself fails to open, the fault is on the client leg rather than the connector leg — that is the separate problem covered under historian trend client connection failures.

Questions asked

Citect SCADA and Historian share one PC, so can I skip the host and instance checks?

No. Co-location removes the network hop, not the endpoint settings. The connector still resolves a host and an instance name, and a destination entry pointing at a different instance or host form fails on one machine exactly as it would across two. A Historian PC can host Citect Connector V3.0, so the shared PC is a supported arrangement and not itself the fault.

Is the locally stored data lost while the status is Disconnected?

No. Storing to local disk is what the connector does while the destination is unreachable, and reports of this fault show the same buffering. The real risks are the disk filling and writes being rejected after reconnection, not loss in transit. Delete nothing from the local store while you diagnose; it is the only copy of the undelivered samples.

Why does one tag still show nothing after the status turns Connected?

A documented Historian behavior applies: if Historian was shut down, disabled and then re-enabled, it does not reconnect to tags whose values have not changed. A tag holding a constant value can stay empty while the whole path is healthy. Test with a tag that is changing, or force one value change on a test tag, then query its exact destination name over a range covering the change.

What exactly does the SMC write-to-Historian entry prove?

It proves the connector processed a write for the tags it discovered, which is stage three of four. It does not prove Historian accepted the samples into the instance you query, and it says nothing about archive visibility. Only a query returning the sample under the exact destination tag name closes that gap, which is why the log entry and your empty query can both be true.

Keep reading

  • How-to

    Run two Plant SCADA Historian connectors during migration

    No: only one Plant SCADA Connector instance is supported per machine, and one connector feeds one historian. Put the second connector on its own machine, attach each connector to a different Plant SCADA Client process so the CTAPI Server load does not double, and retire the old historian only after both archives compare equal.

    Edge computing and IIoT

  • How-to

    Backfilling trend history from a plant historian

    Yes, you can backfill trend history from a plant historian into a SCADA server. Extract the stored records through the historian's native export or ODBC interface, map each source point to its destination object with original timestamps and quality, insert the batches through the destination's documented history methods, and prove the import by reading the interval back from the destination. A saved CSV is a bridge, not proof of import.

    Edge computing and IIoT

  • How-to

    Export historian trend CSV at one-minute intervals

    Export the trend's recorded values and you get change-based events, about 17 minutes apart in this case, because the historian stores a value only when it changes. To get one row every minute, query the tag in the Query tool with regularly spaced retrieval at a 60-second interval and export that result; the grid comes from retrieval, not storage.

    Edge computing and IIoT