Opensearch: Full reindex on new Cloud instance startup hangs at 0% and requires manual cancellation

Problem Statement

When a new dotCMS instance is provisioned in a Cloud environment seeded with the corpsite (starter) data, a full reindex is automatically triggered on startup. This reindex never progresses — it stays at 0% (0 / 8887) with an estimated time of n/a and remains stuck indefinitely.

Navigating to Maintenance → Index shows "A full reindex is in progress." with no forward progress. The only way to recover is to manually Stop Reindex Process, after which a new, manually-triggered reindex completes successfully.

This breaks the expected out-of-the-box experience for freshly provisioned cloud instances: the search index is not usable until an operator notices the stuck reindex and intervenes manually.

Impact: Every fresh instance provisioned with corpsite data on OpenSearch. Workaround exists (stop + re-run), so severity is Medium, but it requires manual intervention on every new instance.

Steps to Reproduce

Video screencast: reindex-on-instance-start.mov — to be attached (drag-and-drop into this issue).

Steps:

1. Provision a new dotCMS instance in a Cloud environment (latest from main).
2. Start it seeded with the corpsite starter data (OpenSearch backend).
3. Once the instance is up, navigate to Maintenance → Index.
4. Observe: "A full reindex is in progress." stuck at 0% — Reindex Progress: 0 / 8887, Time: n/a.
5. Wait — the progress never advances.
6. Click Stop Reindex Process.
7. Trigger a new reindex manually.
8. Observe the manual reindex now progresses and completes normally.

Expected vs. Actual:

Images/Videos

https://github.com/user-attachments/assets/60adb3e4-d898-4065-8290-903516485719

Acceptance Criteria

• The automatic full reindex triggered on a fresh instance startup (Cloud environment, corpsite data, OpenSearch) completes successfully without manual intervention.
• The Maintenance → Index view shows real progress (document count advancing) and a valid time estimate instead of n/a while indexing.
• No manual Stop Reindex Process + re-run is required to obtain a healthy index on a new instance.
• Root cause of the stuck 0 / 8887 reindex on startup is identified and fixed (or documented as environmental with a mitigation if it is not a code defect).

dotCMS Version

Latest from main branch — reproduced on a Cloud environment.

Severity

Medium - Some functionality impacted (workaround: stop the stuck reindex and re-run manually).

Links

NA