Troubleshooting
Why WebIQ 2.13 Publishing Hangs After a 2.12 Project Update
Short answer
The never-moving publishing screen is a workspace lock on the ctrlX, not a slow transfer. Run the migration with Designer 2.13 against a 2.13 runtime, confirm the connection target with --ws-host and --ws-port, and read %APPDATA%/webiq-designer/designer.log for the active-workspace ApiError. The release is a restart of the complete ctrlX device; uninstalling Designer leaves the ctrlX workspace directory untouched.
On this page
The publishing dialog that never moves is not a slow transfer: the ctrlX is refusing the project load because it still holds an active workspace for that project. Set the migration up correctly first — Designer 2.13 connected to a 2.13 runtime, with the 2.12 project updated before it can start. If "Publishing HMI project" still freezes after that, the fault is on the controller: read %APPDATA%/webiq-designer/designer.log for "there already exists an active workspace for it" and release the workspace by restarting the complete ctrlX device. Uninstalling WebIQ Designer on the PC does nothing here, because that uninstall leaves the workspace directory on the ctrlX untouched.
Match the versions before you publish again #
Two failure states get confused in this migration, and they have different fixes:
- The runtime message "projet version not compatible... see WebIQ designer manual chapter project list" means the 2.12 project has not completed its update for the 2.13 runtime. Run the update again in a matching Designer and publish the result.
- A publish that reaches "Publishing HMI project" and never changes means the update ran and the write to the runtime is what is failing.
Version matching is strict in both directions. Designer 2.12 against a 2.13 runtime, or Designer 2.13 against a 2.12 runtime, is not a valid connection for this migration; a 2.12 project can be uploaded to a 2.13 runtime with Designer 2.13, but it cannot start there until the update has been published. Read the installed Designer version from its own version display instead of inferring it from the project file, and read the runtime version from the installed WebIQ app on the ctrlX.
The update itself runs in Designer 2.13 on the engineering PC. If licensing is what blocks opening the project there, start with activating a WebIQ Designer trial license for project conversion.
Confirm which server receives the publish #
Designer can work against its local WebIQ server or connect directly to the ctrlX, and the two destinations are easy to mix up mid-migration. The switch lives in the launch arguments --ws-host and --ws-port: with them, the Designer on the PC talks to the WebIQ server on the controller; without them, publishing and loading happen locally. A successful local publish proves the project file and the Designer installation — and nothing about the ctrlX workspace, which is where this fault lives. In the reported case the project loaded and published locally on both 2.12 and 2.13 while every publish to the ctrlX hung.
Each connection target writes its own Designer log folder. One installation had the default %APPDATA%/webiq-designer/designer.log alongside a connection-specific folder named %APPDATA%\webiq-designer192.168.1.1-10123; that folder name identifies that installation, not a universal path. Record which runtime receives the publish before importing or updating again, and pull logs from the folder that matches it.
A connection refused at login instead of hanging is a different fault with its own checks; see why WebIQ deployment to ctrlX CORE rejects your login.
Collect the two logs that name the fault #
Collect both files before restarting anything, because a restart wipes the immediate failure context:
designer.logat%APPDATA%/webiq-designer/designer.logon the engineering PC — or the connection-specific folder above when Designer connects directly to the ctrlX.connect.logfrom Runtime Manager, after verifying the configured log path in the Runtime Manager settings.
A readable file is not necessarily the active one: the supplied connect.log in the reported case held only entries from the previous year. Note the time of one controlled publish attempt, then confirm the log shows entries at that time before trusting what it says.
The failing load appears in the connection-specific log as repeated lines under designer.dashboard.uiAction.loadApp:
ApiError: The HMI project could not be loaded as there already exists an active workspace for it.ApiError: The workspace could not be not found.— quoted exactly as the log writes it.
Alongside them, designer.host.state records error retrieving workspace info: [object Object]. That pair is the diagnosis: the runtime is holding an editing workspace for the project and rejecting every load server-side. It also explains why closing the Designer window on the PC accomplishes nothing — releasing the window and releasing the remote workspace are different operations.
What each symptom points to #
| Symptom | Where it points | Do this |
|---|---|---|
| "projet version not compatible..." on the runtime | The 2.12 project has not finished its update for 2.13 | Run the update in Designer 2.13, then publish |
| "Publishing HMI project" never changes | The write to the runtime cannot complete | Collect designer.log and connect.log; confirm the publish target |
"there already exists an active workspace for it" in designer.log |
The ctrlX holds an active editing workspace | Restart the complete ctrlX device |
| A blank 2.13 project publishes, the converted one does not | Connection and path are healthy; the imported project's workspace state is the fault | Release the workspace, then retry the same project |
| Uninstalling and reinstalling changes nothing | The state lives on the ctrlX, not the PC | Stop reinstalling; release the workspace |
A stall while Designer loads the imported project file rather than at the publish step is a different failure with its own path; see fix WebIQ Designer stuck loading an HMI project ZIP.
Release the stuck workspace #
Work through this in order:
- Close every other Designer session that could be editing the same project.
- Use the normal project close or unload operation and watch whether the project actually unloads. In the reported installation this could be attempted indefinitely without effect.
- Check free disk space on the ctrlX, because a full disk produces the same never-completing publish. It was fine in the reported case, and it is a one-minute check.
- Restart the complete ctrlX device. That is the release step for a workspace that will not unload: in the reported installation it was the only action that worked, and after it the same unload steps succeeded and the update-and-publish sequence completed.
- Retry exactly one controlled update and publish, note the time, and collect the matching log entries.
The restart stops everything else the controller runs, so schedule it as controlled downtime rather than clicking through it mid-shift.
Skip the reinstall path entirely. Uninstalling WebIQ Designer removes the PC software only — the workspace directory on the ctrlX is unaffected by that uninstall. One site went further, removing the WebIQ app from the ctrlX, clearing app data and program data on the PC, reinstalling both ends and rebooting; the publish hung exactly as before.
Worked example: timing the hang from the 09:07 designer.log block #
Inputs, all from the failing installation's connection-specific log and its recorded behavior:
2023-02-13T09:07:35.934Z—loadApperror, active workspace2023-02-13T09:07:37.894Z—loadApperror, workspace not found2023-02-13T09:07:38.794Z—loadApperror, active workspace2023-02-13T09:07:39.821Z—loadApperror, workspace not found2023-02-13T09:10:37.685Z—loadApperror, active workspace- Normal publish time for this project before the fault: under 1 minute
- Time the operator waited on the frozen dialog: 30 minutes
The rule in words: subtract consecutive timestamps to get the retry intervals. Rejections arriving seconds apart mean the server is answering and refusing, not transferring. Set the give-up point as a multiple of the known-good publish time, then compare the wait already spent against it.
The arithmetic:
- 37.894 s − 35.934 s = 1.960 s
- 38.794 s − 37.894 s = 0.900 s
- 39.821 s − 38.794 s = 1.027 s
- Burst span: 39.821 s − 35.934 s = 3.887 s
- Gap to the next rejection: 09:10:37.685 − 09:07:39.821 = 2 min 57.864 s, effectively 2 min 58 s
- Wait already spent versus normal: 30 min ÷ 1 min = 30 times the known-good publish time
Result: four rejections inside 3.887 s, the same error returning 2 min 58 s later and still repeating at 09:10. Decision: the publish is hung, not slow. Call it hung at 10 minutes — 10 × 1 min = 10 min, ten times the documented normal — stop waiting, collect the log, and restart the ctrlX to release the workspace. The half-hour wait was thirty times the normal publish time and changed nothing, because no transfer was in progress that more time could finish.
Keep a 2.12 rollback until 2.13 publishes #
Export the 2.12 project before touching any version, and keep that export unchanged. The rollback that worked in the reported case was returning Designer and the runtime to matching 2.12 and re-importing the export; the project ran again immediately. The cost is the reason for upgrading in the first place — the recorder with a custom trigger that motivated the reported move exists only from 2.13 and will not run on 2.12. Keep the export as the fallback, run the migration where a hung publish costs nothing, and repeat the workspace release after any version change on either end.
Questions asked
Does Publish as Copy or renaming the project get around the hang?
No. One site renamed the project to project_v2 and used Publish as Copy, and the publishing screen froze the same way, because the copy runs into the same stuck workspace on the controller. The test is still worth running for what a blank project tells you: a new 2.13 project publishing fine over the same connection shows the path, login and runtime are healthy, which isolates the fault to the imported project's workspace state.
Will uninstalling and reinstalling WebIQ Designer fix it?
No. The uninstall touches nothing on the controller: the workspace directory on the ctrlX survives it, so the stuck state does too. One site also removed the WebIQ app from the ctrlX, cleared app data and program data on the PC, reinstalled both and rebooted, and the publish hung exactly as before. Stop reinstalling and release the workspace on the controller instead.
How do I know the log file I am reading is the active one?
Check that its timestamps advance across your own test attempt. In the reported case the supplied connect.log contained only entries from the previous year, so it was readable but not receiving new lines. If a controlled publish attempt leaves no matching entries in the file, correct the configured path in Runtime Manager settings or switch to the connection-specific Designer folder for your target.
What do --ws-host and --ws-port actually change?
They point a Designer running on the PC at the WebIQ server on the ctrlX instead of the local server. Without them, loading and publishing happen locally, which can succeed while the ctrlX workspace stays stuck; the reported project loaded and published locally on both 2.12 and 2.13. If your logs show nothing for your attempts, you are likely reading the wrong target's logs.
How long should I wait before calling the publish hung?
Under one minute is the documented normal for this class of project on a healthy setup. Call it hung at ten minutes, ten times that, and move to the logs. The reported half-hour wait produced nothing because the runtime was rejecting the load on a minutes-long retry cycle, not transferring data that more time could finish.