OpenText Captiva Export Failures: Causes and Fixes

OpenText Capture Troubleshooting

Written By Ryan Shallenberger

Director of Marketing. Ryan specializes in communicating Reveille's offerings to the ECM, IDP, and RPA market with over 8 years in marketing/sales.

November 17, 2026

Quick answer

OpenText Captiva export failures come down to one of five causes: the export connection to the target repository or database failed, the unattended export module stopped or cannot receive tasks, the batch never reached the export step because one upstream task is stuck or in error, licensing or a zero batch priority excluded the batch from processing, or an export definition fell out of sync after an upgrade or process change. Start in Captiva Administrator — read the batch’s state and find which step actually holds the pending tasks before resubmitting anything.

01 — What You’re Seeing

The batch finished. The repository is empty.

Confirm you’re on the right page before you touch anything.

One or more of these signatures, on a capture system that normally drains itself:

  • Batches scan, recognize, and index clean, then sit — and the documents never arrive in Documentum, SharePoint, Content Server, or the export database.
  • The batch shows an error state with the failed tasks at the export step — Documentum Advanced Export, Standard Export, ODBC Export, or another export module.
  • The export module reports its tasks done, and the documents still are not in the repository — a failure mode administrators have reported on community forums for years, with no published vendor answer beyond re-export.
  • Export backlog grows quietly: batch aging climbs, no module throws an error, and the business notices before operations does.
  • The failures started right after an upgrade, a CaptureFlow change, or a server move.

A note on names: Captiva Capture became OpenText Intelligent Capture, and most sites still call it Captiva — the capture server is still the InputAccel Server either way. The export step matters more in Captiva than in most platforms, because Captiva is built to hold documents only briefly: per OpenText’s own system documentation, documents typically remain in the system a few hours to a few days, until they are exported. An unexported batch is business data sitting in a system that was never meant to store it. (The full set of what can be watched on a Capture estate is in the Reveille for OpenText Capture technical overview, ungated.)

02 — Causes, Ranked by Likelihood

Five causes account for nearly every case

Confirm before you fix — each cause has a cheap, specific test.

#CauseHow to confirmFix, in short
1Export connection to the target failed — broken at the far endExport tasks in error; the module log names the repository, database, or path; the target rejects the export account’s login, or the target is down.Restore the connection and credentials, then retrigger the failed export tasks from Captiva Administrator.
2The export module isn’t running or can’t take tasks — queued, silentlyCaptiva Administrator shows the module disconnected or idle while export tasks wait; the module’s Windows service is stopped, or it was started with a department that doesn’t match the tasks.Start the service, confirm its connection to the InputAccel Server, check department startup arguments.
3The batch never reached export — held upstreamExport steps commonly trigger at the batch level, so one page stuck at Completion or Identification, or one errored task upstream, holds the entire batch out of export. The batch view shows where the pending tasks actually sit.Find and clear the stalled node — complete the operator task or reprocess the errored one.
4Licensing or priority stopped processing — excluded by designLicensed page allotment exhausted or license expired (check the licensing panel); or the batch’s priority is 0, which excludes it from processing entirely.Extend the license or page count; reset the batch priority above zero.
5Export definition out of sync — after an upgrade or process changeFailures began with a redeployed CaptureFlow, a new process version, or a server change; the module log shows mapping, object-type, or path errors; or the service account lost Windows rights on the target path.Re-run setup on the export step, verify the export profile’s mappings, verify the service account’s rights on the target.

Cause 3 surprises people because the export module looks innocent. In Captiva’s architecture the server pushes a task only when its trigger values are set — and an export step that triggers at the batch level fires only after every upstream task has finished. One document parked in an operator queue, one page that errored in extraction, and the export step never receives work. Nothing at the export end is wrong, and nothing at the export end will fix it.

03 — The Fixes

Work them in repair order

Read the batch before you touch the module.

1. Read the batch, not the module

Open the batch in Captiva Administrator and look at where the pending tasks actually sit. If they are queued at an upstream step, this is cause 3 and the export side is a bystander: clear the operator queue, reprocess the errored task, and the export step fires on its own. If the tasks sit at the export step in error, read the module log — it names the connection, mapping, or path that failed, which separates cause 1 from cause 5.

2. Walk the export connection from the export host

Test the target from the machine the export module runs on, as the account the service runs under — not from your workstation as yourself. Confirm the repository or database is up, the export account authenticates, and the target path exists with write rights. Service accounts are the recurring offender: Captiva relies on Windows rights for file operations, so an export that “worked yesterday” often stopped at a password rotation or a permissions change nobody connected to capture.

3. Retrigger the failed export tasks

Once the underlying problem is fixed, re-export the batch: select the batch in Captiva Administrator, select the export module in the batch’s pane, and retrigger its tasks — the server resends every export task for that batch. This is the community-standard recovery, documented by administrators on the EMC-era support forums (community-reported) for the specific case where the export reports done and the repository disagrees. Verify the result in the repository, never in the module.

4. Check the module, the license, and the priority

If export tasks are queued with no errors anywhere, confirm the boring things in order: the export module’s Windows service is running; the module shows connected to the InputAccel Server; the module wasn’t started with a department argument that excludes the waiting tasks; the licensed page count and connection count have headroom and the license has not expired; and the batch’s priority is above zero — a zero-priority batch is excluded from processing by design. The server queues tasks indefinitely for a module that never becomes available, and it does so without raising anything an operator would see.

5. After an upgrade or process change, re-setup the export step

Each change to a process creates a new version on the server, and batches are created against the latest version — so a CaptureFlow edit, an export profile change, or a platform upgrade can leave the export definition referencing fields, object types, or paths that no longer line up. Re-run setup on the export step, confirm the mappings against the target’s current schema, and run one test batch end to end before reopening the input sources. Then move the stranded batches through — remember the retention reality above: capture is transient by design, and the evidence ages out.

04 — How to Stop It Recurring

Exports fail at the boundary. Watch the boundary.

Every cause above is visible from outside the capture system — if something is looking.

Nothing on this page required inside knowledge. Every failure mode announces itself at an observable edge: export tasks erroring, modules dropping their server connection, batch aging climbing, license counters running down, batches sitting at priority zero. Platform SLA is not workflow SLA — the Capture services can all be running while the documents stop arriving — and the edges have to be watched continuously, because nobody opens Captiva Administrator on a day nothing seems wrong.

That is what Reveille watches, out of the box: Check export database connection and export file share total file count for cause 1; Capture modules (any host on which a module is running) are connected, Capture modules unique connections, Capture server services, and module task levels for single or multiple Capture servers for cause 2; Check batch aging for single Capture server or multiple Capture servers, batches in error status for single or multiple Capture servers, batches with task errors, and hung batch count for cause 3; and Check licensed page and connections count, license expiration date and status, and batches with priority 0 for cause 4. Cause 5 is a change, not a component — Reveille does not read CaptureFlow definitions — but a bad redeploy shows up in the first cycle of the same batch aging and batches-in-error checks, which turns a broken export definition into a same-day alert instead of a week of quiet backlog.

Reveille for OpenText Capture

Reveille for OpenText Capture provides over 50 ‘out-of-the-box’ Capture specific tests and over 40 Capture Dashboard Metrics across versions 20.x, 22.x, 23.x, 24.x, and 25.x — including every connection, batch, and license check above. See Reveille for OpenText Capture →

Supporting client estates as an MSP, SI, or ISV? That’s what Reveille SENTRY is built for — multi-tenant observability for the estates you manage.

Before the backlog ages past its service level

See the Capture export and batch checks run against a live environment.

Enter a business email and the platform you run. The demo environment opens with the export connection, batch aging, and license checks already configured.

No credit card · Nothing to install · Runs on Reveille’s sample estate. Want the list first? Read the OpenText Capture technical overview.

Quick answers

Why are my OpenText Captiva batches failing at export?
The five most common causes, in order: the export connection to the target repository or database failed, the unattended export module stopped or cannot receive tasks, the batch never reached the export step because one upstream task is stuck or in error, licensing or a zero batch priority excluded the batch from processing, or an export definition fell out of sync after an upgrade or process change. Start in Captiva Administrator: read the batch’s state and find which step actually holds the pending tasks before resubmitting anything.
Why does Captiva show the export as done when documents are missing from the repository?
A long-standing, community-reported failure mode: export tasks report complete while the target repository never received the documents — usually a target connection, credential, or mapping problem on the export side. Verify in the repository, never in the module. To re-export, select the batch in Captiva Administrator, select the export module in the batch’s pane, and retrigger its tasks so the whole batch exports again.
How do I restart a failed export in Captiva?
Fix the underlying problem first — the target connection, the credentials, the export definition — or the retriggered tasks fail the same way. Then, in Captiva Administrator, select the batch, select the export module step in the right-hand pane, and retrigger the tasks; the server resends every export task for that batch. If the export module is stopped or disconnected, the server queues the tasks until a module becomes available, so also confirm the module’s Windows service is running and connected.
How do I monitor Captiva export backlog proactively?
Watch the boundary continuously: the export database connection and export file share counts, batch aging across servers, batches in error or hold status, module connections to the Capture server, and the licensed page and connection counts. Reveille for OpenText Capture runs these as pre-configured tests — over 50 out-of-the-box Capture specific tests — so an export backlog raises an alert before the repository misses a service level.

Fixed it? Make sure you see the next stalled export first.

You may also like…

Get the signal on what’s shaping IDP, ECM, RPA, and intelligent automation.