In general LDAP is an atomic/single unit transaction. However, if Transaction Failover is configure Penrose implement an intelligent connection pooling mechanism so that if Penrose detect a transaction time out or fails on a node, it will retry to a different node. Penrose will attempts to repeat the transaction on each available node until all nodes have been exhausted. Only then is a failure returned to the client by Penrose. ========================================================= Issue dump from jira $VAR1 = { 'priority' => '3', 'customFieldValues' => [], 'project' => 'PENROSE', 'status' => '1', 'components' => [ { 'name' => 'Connectors', 'id' => '10059' }, { 'name' => 'Engine', 'id' => '10009' } ], 'reporter' => 'jimyang', 'key' => 'PENROSE-285', 'assignee' => 'jimyang', 'summary' => 'Transaction Failover', 'id' => '10978', 'updated' => '2008-03-06 09:51:29.0', 'votes' => '0', 'fixVersions' => [ { 'releaseDate' => '2008-09-18 00:00:00.0', 'sequence' => '31', 'name' => 'Penrose-2.1', 'released' => 'false', 'id' => '10130', 'archived' => 'false' } ], 'affectsVersions' => [], 'description' => 'In general LDAP is an atomic/single unit transaction. However, if Transaction Failover is configure Penrose implement an intelligent connection pooling mechanism so that if Penrose detect a transaction time out or fails on a node, it will retry to a different node. Penrose will attempts to repeat the transaction on each available node until all nodes have been exhausted. Only then is a failure returned to the client by Penrose. ', 'created' => '2008-03-06 09:51:29.0', 'type' => '2' }; =========================================================