Skip to content

Pull Problems

Diagnose BusinessObjects pulls that do not start, stall, run slowly or leave stale objects in the MetaMiner Client.

A pull is MetaMiner collecting metadata from one BusinessObjects environment. Each enabled environment pulls in a cycle: changed objects first, then deleted and newly imported objects, then deep processing of Web Intelligence documents and universes. See Environments and Pulls for the model.

Check the pull first

  1. In the MetaMiner Client (MMC), right-click the environment row in the CMS Explorer and choose Data Refresh Statistics. The dialog lists the most recent pulls of that environment.
  2. To start a cycle now instead of waiting for the schedule, right-click the environment row and choose Force On Demand Pull....
  3. On the server, open server\logs\server.log (MMA: File > Open MetaMiner Service Log) and search for the environment name.

Image of the MMC CMS Explorer with the environment row context menu showing Data Refresh Statistics and Force On Demand Pull

../../images/help/mmc-environment-row-menu.png

Pull does not start

No new data appears for an environment

Symptom. Objects changed in BusinessObjects do not appear in the MMC, and Data Refresh Statistics shows no recent pull.

Cause. One of these: the environment is disabled in the MetaMiner Administrator (MMA); the MetaMiner service is not running; the current time falls inside a pull exclusion window for the environment; the object type is excluded from pulls; or the license does not include that object type.

Fix.

  1. In the MMA, check that the environment is enabled and validates. See Validate an Environment.
  2. Check the environment's pull schedule and exclusion windows. See Pull Schedule and Exclusions.
  3. Check MetaMinerServer.properties for cms.pull.exclusions or cms.pull.inclusions left over from an earlier diagnosis. See Server Properties.
  4. Use Force On Demand Pull... and watch server.log.

What to send support. A log bundle and the name of the environment.

The first pull of a new environment takes a long time

Symptom. A newly added environment shows few objects for hours.

Cause. The first pull reads every object in the repository, then deep-processes documents and universes. Later pulls read only what changed. MetaMiner sends an email when the initial pull of an environment completes, if email notifications are configured.

Fix. Let the first pull finish. Configure the CMS database connection for the environment to speed up repository reads. See Audit and CMS Databases.

What to send support. Only if it never completes: a log bundle and the approximate object count of the repository.

Environment validation errors

Web Intelligence REST API validation failed

Symptom. The MMA shows Web Intelligence REST API validation failed for the environment.

Cause. MetaMiner reads Web Intelligence documents through the BusinessObjects RESTful web service. By default it uses the address that the CMS advertises for that service. That address can be an internal or load-balancer name that the MetaMiner server cannot reach.

Fix. In the environment dialog, check Custom next to REST API URL and enter a URL that the MetaMiner server can reach, for example https://<wacs-host>:<port>. Validate again. Make sure the port is open (see Ports).

What to send support. The REST URL you tried, the error detail shown after the message, and admin.log.

Audit DB connection failed / CMS DB connection failed

Symptom. The MMA shows Audit DB connection failed or CMS DB connection failed.

Cause. The database host, port, database name, driver type or credentials are wrong, or a firewall blocks the database port from the MetaMiner server.

Fix. Test the connection with the same account from the MetaMiner server (for example with your database client). Correct the settings and validate again. MetaMiner only reads these databases, so a read-only account is enough. See Audit and CMS Databases.

What to send support. Database type and version, the error detail, and admin.log.

Connection validation failed / Unable to detect the CMS version

Symptom. Validation fails before any database or REST checks.

Cause. The logon to the CMS failed, or the account cannot read the repository.

Fix. See the BusinessObjects entries in Connection Problems.

What to send support. admin.log.

Pull stalls or is slow

A pull starts but never finishes

Symptom. Data Refresh Statistics shows a pull that started long ago, or server.log stops logging progress for the environment.

Cause. MetaMiner is waiting for BusinessObjects to answer: the CMS is overloaded or partly down, or a processing server is not responding. Each call to the BusinessObjects SDK is cut off after the ORB request timeout (ooc.orb.policy.request_timeout, default 4 hours), so a hung call eventually fails and the next cycle starts. While MetaMiner waits, other environments on the same BusinessObjects version can also wait.

Fix.

  1. Check the health of the CMS and its servers in the CMC.
  2. Open server\logs\threads.log (MMA: File > Open Threads Log). MetaMiner writes thread dumps there when an action runs longer than expected.
  3. If BusinessObjects is healthy again and the pull is still stuck, restart the MetaMiner service. See Start and Stop the Server.

What to send support. A log bundle including threads.log, created before you restart the service.

Pulls are slow

Symptom. Changes take a long time to appear, or each cycle takes longer than the pull period.

Cause. Large repositories, many Web Intelligence documents to deep-process, slow audit or CMS databases, or too few threads or CPUs for the workload.

Fix.

  1. Configure the CMS database connection for each environment; repository reads are faster than SDK queries.
  2. In MMA Preferences > Settings > Pull, review Threads per datatype and Threads per Webi.
  3. In MMA Preferences > Settings > System, review Maximum Allowed CPUs and the memory allocation. See Performance and Memory.
  4. Use pull exclusion windows to keep pulls out of the BusinessObjects maintenance window.

What to send support. A log bundle, the number of environments, and the approximate document and universe counts.

Web Intelligence Processing Failure email

Symptom. You receive an email with the subject "... Webi Processing Failure".

Cause. The MetaMiner server cannot reach one or more Web Intelligence Processing Servers.

Fix. See Web Intelligence Processing Failures.

What to send support. The email and a log bundle.

Stale or missing objects

Deleted objects are still shown in the MMC

Symptom. An object deleted in BusinessObjects still appears in the MMC.

Cause. MetaMiner detects deletions during each pull cycle by comparing the objects in the CMS with the objects in the MetaMiner Database (MMDB). The object stays visible until a cycle completes for that environment and the MMC receives the update. A failing or stalled pull delays deletions.

Fix.

  1. Press F5 in the MMC to refresh the data from the server.
  2. Check Data Refresh Statistics for a recent successful pull. If there is none, work through the entries above.
  3. Use Force On Demand Pull... on the environment row.

What to send support. The object's CUID and type, the time it was deleted, and a log bundle.

Excessive deletions email

Symptom. You receive a notification that many objects were deleted in one pull.

Cause. One pull detected more deletions than the excessive deletions notification threshold. This can be a real cleanup in BusinessObjects, or a sign that MetaMiner read an incomplete object list.

Fix. Confirm with your BusinessObjects administrators whether content was removed. Adjust the threshold in MMA Preferences > Settings > Notifications if large cleanups are normal for you. See Notifications.

What to send support. If the deletions were not expected: the email and a log bundle.

Some documents or universes have no detail (no queries, objects or SQL)

Symptom. A document or universe appears in the CMS Explorer but its deep metadata is empty.

Cause. Deep processing of that item failed or timed out, the item could not be opened in BusinessObjects, or the license edition does not include deep processing for that content type.

Fix. Open the item in BusinessObjects to confirm it opens and refreshes. Check server.log and universe-errors.log for its name or CUID. See Document Processing.

What to send support. The CUID, the matching log lines, and a log bundle.

Where to look in the logs

All server logs are in server\logs\ under the installation folder. See Log Files.

Log Look here for
server.log Pull cycles, errors per environment, deep processing results
slc.log Service start, stop and reload events
threads.log Thread dumps written when actions take too long
diagnostic.log Diagnostic messages and notification activity
universe-errors.log Universe processing errors
mmdb.log MetaMiner Database messages and schema updates
admin.log Environment validation run from the MMA

Next steps