Qlik Sense Reload Task Fails with No Reload Log – OpenApp 1013 After Interrupted App Publication

[JorgeEsteso]

A published Qlik Sense application suddenly stopped reloading from QMC after an application publication/replacement process was interrupted.

The application had previously been working correctly. During an update, the publication process became stuck and the page was refreshed before it had fully completed. The publication was then reverted and attempted again.

From that point onwards, the published application could still be opened, but any reload launched against that specific application failed within a few seconds and no normal reload log was generated.

A duplicate of the same application could be reloaded successfully.

System Information:
- Qlik Sense Nov 2025 Patch 4
- Microsoft Windows Server 2019 Standard
- PostgreSQL 14

Error Description

Several tests were performed to isolate the issue.

A very similar application reloaded correctly, and a duplicate of the affected application could also be reloaded successfully from the Data Load Editor and through a new reload task.

However, the original published application continued to fail even after creating a new task, using "Reload now" from QMC, restarting the Qlik Sense services, increasing the available disk space and replacing the application with both a working copy and an older version.

The Scheduler logs showed that the task started correctly and was sent to the Engine:

Reload sequence failed with exception

Reload sequence was not successful
(Result=False, Finished=False, Aborted=False)

Message from ReloadProvider:
Reload failed in Engine. Check engine or script logs.

The Engine was able to begin loading the application:

Server: App Load: Beginning open of app <APP_ID>

DOC loading: Beginning load of document <APP_ID>

Document Load: The document <APP_ID> was loaded

However, the Engine System log returned:

Command=Open app
Result=1013
ResultText=Error: Unknown error

Global::OpenApp

Because the failure occurred while opening the application, the actual reload script never started and therefore no normal reload log was created.

Root Cause

The issue was caused by an inconsistency in the Qlik Repository database (QSR) for the affected published application.

The following query was used to check the internal objects associated with the application:

SELECT *
FROM public."AppObjects"
WHERE "App_ID" = '<APP_ID>'
  AND "ObjectType" IN ('app_appscript', 'LoadModel');

For the affected application, only the following object existed:

LoadModel

The internal object:

app_appscript

was missing.

The same query was executed against a working application and both LoadModel and app_appscript were present.

Based on the sequence of events, the repository inconsistency appears to have been created when the publication/replacement process became stuck and was interrupted.

This also explains why replacing the application with another version did not solve the problem. The issue was not related to the application content itself, but to the repository information associated with the existing published application.

Resolution

One possible solution would have been to publish the working copy as a completely new application. However, this was something we wanted to avoid because the existing application ID was already being used by multiple tasks, links and other integrations connected to the application.

Fortunately, a previous QSR backup from approximately five months earlier was available, from a point when the same application was working correctly.

The affected application was checked in the backup and the missing app_appscript record was still present.

Before making any changes, a new backup of the current QSR database was created. The Qlik Sense services that could write to the repository were stopped, while keeping the PostgreSQL/Qlik Repository Database available.

First, the missing app_appscript object was extracted from the restored backup using the application ID:

SELECT row_to_json(x)
FROM (
    SELECT *
    FROM public."AppObjects"
    WHERE "App_ID" = '<APP_ID>'
      AND "ObjectType" = 'app_appscript'
) x;

This returned the complete app_appscript record belonging to the affected application.

The recovered record was then inserted into the current QSR using a transaction:

BEGIN;

INSERT INTO public."AppObjects"
SELECT *
FROM json_populate_record(
    NULL::public."AppObjects",
    '<JSON_RETURNED_FROM_BACKUP>'::json
);

SELECT
    "ID",
    "App_ID",
    "ObjectType",
    "EngineObjectId"
FROM public."AppObjects"
WHERE "App_ID" = '<APP_ID>'
  AND "ObjectType" IN ('app_appscript', 'LoadModel');

COMMIT;

Only the missing app_appscript record belonging to the same application and the same App ID was recovered from the previous backup and restored into the current QSR. The complete repository was not restored, and the application itself was not deleted or recreated.

After restoring the missing object, the application once again contained both expected entries:

app_appscript
LoadModel

The Qlik Sense services were started again and both Reload now and the scheduled reload task completed successfully.

This allowed us to recover the original published application while preserving its existing App ID and avoiding changes to the multiple tasks, links and integrations that were already connected to it.

Hope this helps someone experiencing the same behaviour, because it's a pretty weird error.