The full error text most often encountered is:
ERROR [nucleusNamespace.atg.search.repository.IncrementalLoader]
atg.repository.search.indexing.IndexingException: Cannot complete
indexing update because we do not currently own the configuration,
owned by "null"
Each Oracle ATG Web Commerce instance in which the DAF.Search.Index
module is running performs a periodic check of the searchConfig
items in the IncrementalItemQueue
repository, checking for searchConfig
items that are claimed, but for which the lastActivity
has not been updated in 10 minutes (by default). The error indicates that one ATG instance removed the lock held by the ATG instance doing the indexing.
The most common causes of this error are:
Two ATG instances are running on the same machine with the same DRP port.
Check the
/atg/dynamo/Configuration
component for each instance to make sure that the DRP ports are unique.Another machine clears out the lock that the indexing machine holds in the database table.
Lock managers and related repository properties are not configured correctly, so other instances see stale data.
Make sure that lock managers are configured and shared correctly.
You should also check the other servers in your cluster for warning messages on the other servers in your cluster that begin with the words “Releasing configuration” and name the IndexingOutputConfig
in question, and verify their DRP port and lock management configuration on those servers.
If none of the above problems is the cause, it may be that the application hung for more than ten minutes during indexing. To investigate this, collect and examine thread dumps from the indexing instance to identify where indexing is stuck. If indexing does not appear stuck, you may need to increase the timeout period configured in the timeOutMsecs
property of the /atg/search/repository/ConfigStatePersister
component from the default ten minutes to a longer interval.
Caution: Increase this setting only if you are absolutely certain that no other problem is the cause.