Maintaining the Search Log, Common Errors Related to Log Size, and Resetting the Log Directory (Doc ID 1041081.1)

Last updated on APRIL 17, 2017

Applies to:

Oracle Knowledge - Version 8.1.3.3 and later
Information in this document applies to any platform.
The log refered to here is the search binary log. The instance standard out log rotates and should maintain itself.

Symptoms

After 8.5.1.5 the search log will be automatically purged when a new log is written if the limit conditions are reached.  In addition the The 'Log Purge' job ('Log history purge' task) is working from release 8.5.1.5 and on.  Before this the products 'Log history purge' task in the System Manager and automatic purge has not been working.  The task will hang (run forever without completion) and even should you manually stop it, it may left behind a lock on the logfile file that will lead to ConcurrentUpdateException error in the subsequent content crawl (Full or Incremental) job.  Then, user will need to restart the indexer instance in order to recover (release the left-behind lock) and be able to run the content crawl (Full or Incremental) job again without exception.  It was not recommended for use before 8.5.1.5.

In addition the cleaning up of the log derby database is fixed in 8.6.
The Bug 16284520 - RFA: Log Purge task does not cleanup the growing log/catalog/default/dbms/*

As a result in older versions the log files have to be deleted regularly by a different system or manual process.

If not during crawls log errors can happen and the crawls can even fail.

Log errors - these errors can be a symptom that your indexer binary log directory is getting too large.

<INQUIRA_INSTALL>\instances\<app>\<env>\content\data\log\binary\default

 

[5288989 Local Workclient-default-taskrun(279301)] Event(Code=LOG_STORE_EXCEPTION, id=#205B81I279301P279297) occurred at 9/26/12 10:40 PM: java.net.ConnectExceptionDO_NOT_ADDcom.inquira.prep
[5288989 Local Workclient-default-taskrun (279297)] java.net.ConnectException: Connection refused
java.net.ConnectException: Connection refused
at java.net.PlainSocketImpl.socketConnect(Native Method)
at java.net.PlainSocketImpl.doConnect(PlainSocketImpl.java:333)

Eventually if ignored the log directory dbms (set of tasks that have been performed by the indexer) will fail to load giving these messages.

These would be at startup of the indexer.  This will cause jobs to run irratically and most likely fail at some point.

Creating new log catalog schema.
Exception in thread "LogCatalogBuilder" com.inquira.infra.InquiraRuntimeException: [LogCatalogBuilder @ Fri Sep 28 03:04:00 EDT 2012] (LOG_CATALOG_DB_ERROR)
at com.inquira.log.LogStoreImpl.getDatasource(LogStoreImpl.java:819)
at com.inquira.log.store.access.CatalogBuildingContext.<init>(CatalogBuildingContext.java:97)
at com.inquira.log.store.LogSet.getContext(LogSet.java:563)
at com.inquira.log.store.LogSet.deleteCatalog(LogSet.java:635)
at com.inquira.log.LogStoreImpl.deleteCatalog(LogStoreImpl.java:247)
at com.inquira.scheduler.SchedulerServiceImpl$LogCatalogBuilderThread.run(SchedulerServiceImpl.java:1592)
Caused by: SQL Exception: Failed to start database '/inqapps/inquira8/instances/cisco/development/content/data/log/catalog/default/dbms', see the next exception for details.
at org.apache.derby.impl.jdbc.Util.newEmbedSQLException(Unknown Source)
at org.apache.derby.impl.jdbc.Util.newEmbedSQLException(Unknown Source)

You may also experience your disk space getting full as the files under [Indexer Instance Folder]\[environment]\content\data\log\catalog\default\dbms\ can grow really large

Cause

Sign In with your My Oracle Support account

Don't have a My Oracle Support account? Click to get started

My Oracle Support provides customers with access to over a
Million Knowledge Articles and hundreds of Community platforms