IMD SolutionComputing LibraryShop main store

How-to

Run two Plant SCADA Historian connectors during migration

Edge computing and IIoT Published 5 min read

Short answer

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.

On this page
  1. Can one server run a connector for each historian?
  2. Why does the Client process decide where the second connector connects?
  3. How do you set up the second machine?
  4. How do you match the tag selection between the two connectors?
  5. What do you check when tags are missing in the new historian?
  6. How do you prove both historians store the same data?
  7. When is it safe to retire the old historian?
  8. Questions asked

No — one connector instance per machine is the supported limit, so keeping the old historian fed while the new one fills means a second machine. The Plant SCADA Connector (formerly the Citect to Wonderware Historian Connector) reads tags through the CTAPI Server hosted by a Plant SCADA Client process, so the two connectors must also attach to two different Client processes. Below is the layout, the setup steps, and the comparison that decides when the old historian can be switched off.

Can one server run a connector for each historian? #

No. One instance of the Plant SCADA Connector per machine is the supported configuration, and a single connector writes to a single historian, so the dual feed is one connector per machine: two machines, two connectors, two destinations.

Starting a second instance on the existing SCADA server to feed Historian 2023 is not a supported path, and it is the first thing to rule out when someone proposes it. The target layout for the overlap is:

Machine Connector instance Destination
Existing SCADA server Existing connector, configuration unchanged Historian 2014, remote server
Second machine New connector, release 4.0 Historian 2023

Before adding anything, list the machines that already host a connector; exactly one entry per machine is what a healthy installation looks like. If the existing server shows more than one, correct that fault first.

Why does the Client process decide where the second connector connects? #

Because the connector reads tags through the CTAPI Server hosted by a Plant SCADA Client process, every connector instance is another data consumer on that one process; the load lands on the Client process, not on the connector machine.

The connector does not read the SCADA runtime directly. Point each connector at a different Client process, because two connectors on one Client double the tag-read work on the process that also serves the operator displays.

Worked example, illustrative numbers: the Client process serving the existing connector runs at 18% CPU in normal production and measured 10% before that connector was attached, so the connector's share is 18% − 10% = 8%. The Client process you pick for the second connector sits at 10% today, so plan for 10% + 8% = 18% once its connector attaches. If both connectors landed on that same Client, it would carry 10% + 8% + 8% = 26%. Decision: two connectors, two Client processes, and a recorded CPU baseline for each before the second connector starts.

How do you set up the second machine? #

Put connector 4.0 on a machine of its own that can reach both the new historian and the Client process it will subscribe through, and leave the existing connector and its configuration untouched.

  1. Pick the machine. It needs network reach to the Historian 2023 node and to the Plant SCADA node running the Client process it will use; an industrial mini PC is the usual choice when the second connector lives on the plant floor rather than in the server room.
  2. Install the current connector release, 4.0, on that machine. Do not touch the installation on the SCADA server.
  3. Confirm name resolution and reachability from the new machine to both endpoints: the Historian 2023 node and the Plant SCADA node hosting the target Client process.
  4. Configure the new connector to attach to a Client process the existing connector does not use.
  5. Leave the existing connector pointed at Historian 2014, with its configuration unchanged, so the old archive keeps filling until cutover.

A failure at step 3 is a network or name-resolution fault, not a connector configuration problem; fix it there before touching connector settings.

How do you match the tag selection between the two connectors? #

Duplicate the existing connector's tag selection exactly in the second connector, because a differing set makes the two archives impossible to compare and hides missing tags during the overlap.

A tag that exists in Historian 2014 but not in Historian 2023 then means one thing only: the second connector's selection differs. Before moving on, pick one analog tag with a visibly changing value and confirm it appears in the Historian 2023 tag list.

What do you check when tags are missing in the new historian? #

Separate a tag fault from a connection fault by where the data stops: no tags at all points at the connector's route or credentials, while tags present in the old archive but absent from the new one point at the second connector's selection.

Symptom Layer that owns it What to check
Tag missing in both historians Plant SCADA project Tag is configured for trending and its value updates in a runtime display
Tag in Historian 2014, absent from 2023 Second connector's selection Selection matches the existing connector, tag for tag
No tags at all in 2023 Connection Connector status, Historian 2023 address and credentials, route to the Client process
Values flow but displays lag Client process load Client CPU against the recorded baseline; move the connector to a less loaded Client

A total absence of tags is a connection fault before it is a configuration fault — work it the same way you would when troubleshooting historian trend client connection failures.

How do you prove both historians store the same data? #

Query the same three tags — one fast analog, one slow analog, one digital — in both historians over the same window that starts after the second connector went live, and compare timestamps and values.

Matching change events on all three confirm both connectors receive the same updates from their Client processes. A gap in only one archive points at that connector's link or its Client process, never at the field tag. Anything earlier than the second connector's start time exists only in the old archive, so either close that gap by backfilling trend history from a plant historian or start every comparison window at the connector's start time.

When is it safe to retire the old historian? #

After the sample comparison matches over one full production cycle — several days at minimum, long enough that the slowest tags have written several values — the new archive is proven and the old historian can be switched off.

Retire in order: stop the comparison, stop the old connector, then stop Historian 2014. Through the whole overlap, make no configuration change to the old connector — it feeds the one archive you compare against, and a mid-migration edit risks exactly the history you are trying to preserve.

Questions asked

Does the connector also feed AVEVA Insight, or only historians?

Its documentation covers transferring data from Plant SCADA to both AVEVA Historian and AVEVA Insight. That does not change the rule here: one connector instance per machine and one historian per instance, so the dual feed needs its second machine regardless of which destination each instance serves.

Our project runs a single Client process. Where does the second connector attach?

Both connectors then attach to that one Client process and share the CTAPI Server it hosts. The single process carries both subscriptions, so record its CPU before the second connector starts and confirm the added load stays within the headroom the process has while it also serves displays.

How do you copy the tag list from the old connector to the new one?

Export the existing connector's configuration if your release offers an export; otherwise walk the same tag folders in the new connector and enable the same trending set. The check that matters is a tag-count comparison plus a spot check of the sample tags against the old connector, done before anyone relies on the new archive.

Should the old connector be upgraded to 4.0 before the migration?

No. Install connector 4.0 on the new machine and leave the existing installation on the version it already runs. Upgrading the production feed in the middle of a migration adds a second variable to every archive comparison, and the supported layout is one instance per machine, not matched versions on both.

Keep reading

  • How-to

    Backfilling PI from an existing plant historian

    Split the historian-to-PI backfill into four independent stages — extract, intermediate store, external compression, destination load — and qualify throughput with a measured pilot. Rows divided by elapsed seconds, not tag count, sets the schedule: 12,000 tags from 2020 at 500 rows/s is about 203 days. Check the native Historian 2023 replication route before writing pipeline code.

    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

  • Comparison

    Evaluating Ignition historian with TimescaleDB for your plant

    The Gold Technology Provider announcement is a reason to test TimescaleDB behind Ignition, not proof it replaces your historian. No footprint, compression or performance number is documented for this pairing. Adopt it only after a replay test on your exact Ignition 8.3 build measures growth, ingestion continuity, query response and backup/restore against your current store.

    Edge computing and IIoT