[ https://issues.jboss.org/browse/ISPN-4196?page=com.atlassian.jira.plugin.... ] Martin Gencur commented on ISPN-4196: ------------------------------------- After some debugging, the problems seems to be in the QueryInterceptor which handles clearing the index. The 30-second waiting period is caused by ComponentRegistry calling stop() operation on TransactionTable. This stop() method waits for localTransactions variable to be empty (not containing any LocalTransactions). This is fine for default cache on which the clear() was initially invoked. However, during the clear() operation, another "_cluster_registry_cache_" is used by QueryInterceptor (it creates and registers a new local transaction for the _cluster_registry_cache_) and at the time of calling stop() method on its TransactionTable, the transaction is still there. This transaction should have been removed from the localTransactions variable but was not. I'll investigate further what can be done about this. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira