Troubleshooting
Fix ControlLogix tags missing from Ignition OPC UA browse
Short answer
Restart the OPC-UA module, delete the stored /data/drivers/.tags file on the Ignition server, then refresh the browse from the device page — not the OPC Browser. The Logix driver resolves only addresses present in its browse results, so a hand-typed tag for a missing address stays Stale until the stored browse data is rebuilt.
On this page
- Why do some ControlLogix tags never show up in the Ignition browse?
- Why does a typed tag stay Stale when the address is correct in the PLC?
- How do you prove the connection is healthy before touching anything?
- Could the redundant ControlLogix pair be hiding the tags?
- Which tag setting in the PLC hides tags from every browse?
- How do you clear the driver's stored tag list?
- Why refresh from the device page and not the OPC Browser?
- How do you verify the rebuilt browse is complete?
- What do you do when the missing tags come back?
- Questions asked
The device shows Connected, part of the tag tree browses, and the rest never arrives: some branches are missing and others show up as folders where tags should be. Refresh does nothing, waiting does nothing, a full Ignition restart does nothing, and RSLinx on the same IP address sees every tag. The fix is the driver's reset sequence: restart the OPC-UA module, delete the stored /data/drivers/.tags file on the Ignition server, then refresh the browse from the device page — not from the OPC Browser. The case that documents this fault ran an L55-M13 processor pair behind a 1756-ENBT at firmware 15.60 with Ignition 7.4.2 (b953), missing both global and program tags.
Why do some ControlLogix tags never show up in the Ignition browse? #
The driver's stored browse data is incomplete, and that stored file survives a gateway restart: whole branches of the tag tree never populate, and nodes that should hold tags appear as folders instead.
The Logix driver keeps its browse results in a file on the Ignition server and works from that copy. When branches are missing from the stored set, the OPC Browser view, tag creation, and address resolution all inherit the gap. Waiting cannot fill it in, because nothing re-reads the controller on its own. A gateway restart cannot clear it, because the file persists on disk. Diagnose the pattern before you clear anything:
| What you see | What it means | Where you go |
|---|---|---|
| Only program-scoped tags missing | Program-scope folders did not enumerate | Test one controller-scoped tag, then the reset |
| Global and program tags both missing | Stored browse data incomplete | The reset sequence below |
| Branches show as folders, not tags | That node's browse data never populated | The reset sequence below |
| Same missing set after refresh and a gateway restart | Persisted state on the server, not timing | The reset sequence below |
| Missing tags blocked by External Access | Hidden from external reads by design | Fix the setting in RSLogix 5000 |
Missing tag trees are not unique to this driver. When Aveva Edge cannot browse S7-1500 OPC UA tags, the same rule holds: prove the connection first, then distrust the browse data.
Why does a typed tag stay Stale when the address is correct in the PLC? #
The Logix driver resolves an address only if that address exists in its browse results, so a tag typed by hand for a missing address sits in Stale forever, no matter how correct the name is in the controller.
Hand-typing the address is not a workaround — it is a dead end. The name can be verified in RSLogix 5000 and still never resolve, because resolution runs against the driver's stored set, not against the controller. Confirm what the tag actually reports by right-clicking it in the Tag Browser and choosing View Tag Diagnostics; that view shows the status behind the quality. The decision: fix the browse data, never the address. Retype nothing — the tag reads the moment the rebuilt browse returns its address.
How do you prove the connection is healthy before touching anything? #
Drag one browsable tag from the device browse into the designer and read its quality: a Good quality on a browsed tag proves the IP path, the bridge module, and the CIP connection work, and rules out cabling and addressing.
Work through it in order:
- Drag one tag from the device browse and create it in Ignition.
- Quality Good: stop looking at cabling, IP settings, and module configuration. The connection is healthy and the fault is the tag list. In the documented case a browsed tag read Good, which is exactly what pointed at the stored data.
- Quality Bad or Stale on a browsed tag: you have a connection fault, not a browse fault. Check the device status, the IP address, and the routing path. Documented connection-side causes of partial browses include IP conflicts and DHCP collisions — a duplicate address on the network produces exactly this half-working behavior.
- When the device itself drops in and out, a SCADA PLC communication alarm for link loss puts the failure on the operator desk instead of burying it in a quality column.
Could the redundant ControlLogix pair be hiding the tags? #
No. Redundancy is the obvious suspect and the wrong one: the driver talks to one IP address, the redundancy hardware moves that IP between the Ethernet modules at changeover, and the driver never knows a second processor exists.
The driver never merges tag lists from two controllers, so there is nothing on the redundancy side to miss. Verify it in one step: point RSLinx at the same IP address Ignition uses.
- Every missing tag present in RSLinx on that IP: the controller is exonerated. Do not spend time on redundancy configuration or tag sharing.
- Tags absent on that IP: you are looking at the other chassis, or at a project mismatch between primary and secondary. Fix that first — that is a different fault.
The "shared tags will not show" idea belongs to GuardLogix, not to a redundant L55 pair. Redundancy touches this fault in one way only: after a changeover the driver may still hold browse data built against the previous primary, which is cached state — the reset below handles it.
Which tag setting in the PLC hides tags from every browse? #
In RSLogix 5000, a tag's External Access has to be Read Only or Read/Write for that tag to appear in an external browse, so check this on the missing tags before you touch the driver.
Open the tag's properties and read the External Access field. A tag set below those two values is invisible to Ignition, to RSLinx, and to every other external reader on purpose — that tag was never the driver's fault. Change the setting and download. If RSLinx shows the tag, its access is already sufficient and the fault is the stored data, so proceed to the reset. Running this check first keeps the reset's verdict clean: anything still missing afterward is either access-restricted or a genuine recurrence.
How do you clear the driver's stored tag list? #
Restart the OPC-UA module, delete /data/drivers/.tags on the Ignition server, then refresh the browse from the device page: that three-step sequence, in that order, is the documented fix, and no step is optional.
- Restart the OPC-UA module in the gateway. Do this first — deleting the file while the module still holds the stale set in memory risks the same data being written back.
- On the Ignition server, delete /data/drivers/.tags.
- Refresh the browse for the device from its device page, not from the OPC Browser.
Then re-browse the previously missing branches and confirm they populate as tags rather than folders. One caution, once: restarting the OPC-UA module interrupts every OPC client this gateway serves, so screens stop updating until it returns — run the sequence in a window where stale values on the floor are acceptable.
Why refresh from the device page and not the OPC Browser? #
The device-page refresh tells the driver to re-read the controller's tag list; an OPC Browser refresh only re-reads what the server already holds in memory, so it can never bring a missing branch back.
This distinction explains the no-help result you already saw. A refresh executed in the OPC Browser never reaches the driver, so the incomplete stored set stays exactly as it was. After the reset, the device page is also where you confirm the rebuild: the driver's own browse results are the evidence that the stored set is complete.
How do you verify the rebuilt browse is complete? #
Compare the Ignition browse against an RSLinx browse of the same IP address, branch by branch, then re-create the tag that was stuck in Stale from the browse and confirm it reads Good with a value that matches the controller.
Count, do not eyeball. An illustrative example with real arithmetic: the RSLinx browse returns 1,850 tags in controller scope and the Ignition browse shows 1,522. The gap is 1,850 − 1,522 = 328 tags, and 328 ÷ 1,850 = 0.177, so 18 percent of the scope never arrived. Any gap above zero means the stored browse is still incomplete; the size of the gap only sets how deep the spot-checks go. With a project planning thousands of tags, verify several program scopes as well — a tree that looks right at the root can still be missing whole scopes lower down. Finish by re-creating the original Stale tag from the browse and confirming Good quality and a matching value.
What do you do when the missing tags come back? #
Record the fault state, run one clean reset, then call the vendor's support while the fault is still present: the behavior is intermittent, was never reproduced in-house, and a live case is worth more than a cleared one.
Work the recurrence in this order:
- Record the fault state in full before you delete anything — the evidence is gone once the file is cleared.
- Run the reset sequence once, cleanly, in order.
- Still missing after one clean reset: stop. Repeated resets add nothing. Open a case with Inductive Automation support while the fault is present, carrying the recorded state.
- If a changeover window is available, repeat the branch-by-branch comparison after the pair swaps. A tag list that survives a changeover rules redundancy out as the trigger.
Questions asked
Does deleting the .tags file remove the tags I already built in Ignition?
No. The file caches the driver's browse results, and your project tags live elsewhere. Tags that point at addresses the rebuilt browse returns start reading again on their own. A tag still Stale afterward points at an address the browse still does not return, so create that one from the refreshed browse.
Could a newer Ignition release make this unnecessary?
The documented fix for this fault is the reset sequence, and no release note in the available records names a version where the intermittent browse loss stopped. Run the sequence on the build you have, capture the build number in your notes, and report any recurrence so the case file carries it.
Does this fault happen on a non-redundant ControlLogix too?
Nothing in the mechanism depends on redundancy. The driver stores one browse result set per device, and that stored set is what goes incomplete. The documented case ran a redundant L55 pair, but the reset is written against the stored data, not against the pair, so the same sequence applies.
What should I write down before I delete the file?
The time, the controller model and firmware revision, the bridge module, the Ignition build number, and the exact list of missing branches. The fault is intermittent and the evidence disappears once the file is cleared, so that record is everything a support case will have.
Why does a tag write return Good but the value snap back to the old number?
That is a different fault. A Good write followed by the previous value means the controller program is writing the same address from somewhere else, typically function-block logic or another routine, so audit the rungs that write the tag. The browse reset changes nothing about competing writes.