January 2011 Agent Update

On January 13, Scribe released an update to the Scribe Online Agent to address several software issues and enhance Agent functionality. If you have a Scribe Online Agent installed and running, the Agent will update itself automatically with this latest release. For information about Scribe Online Agent updates, see Updating an Agent (http://help.scribesoft.com/scribe/en/index.htm#sol/agent/agentupdate.htm) in the Scribe Online Help Center.


This release of the Agent adds the following features:



  • Database Views: If you build a Map in Scribe Online SYS and select Microsoft SQLServer as your source connection, you can now select database views as source entities (in addition to tables). Note that database views will display in the list of target entities, but cannot be selected.
  • Custom SQL Datatypes: Scribe Online now supports custom Microsoft SQL Server data types. Columns that are defined with custom data types in a SQL Server table will no longer appear as fields with a datatype of " Unsupported" . Instead, these fields now display as having a data type of " Object." You can now link to or from fields with custom data types, although you may need to convert the data using a data conversion function such as TOSTRING. For more information, see About Data Conversion (http://help.scribesoft.com/scribe/en/index.htm#sol/formulas/dataconversi...).


Additionally, this latest Agent update addresses the following defects:



  • Scribe Online SYS attempts to process the last record each time it runs when using the LastRunDate feature. When configuring a Map to only process records that have changed since the last time that Map was run, Scribe Online saves the latest value of the last modified timestamp of the source entity, and uses that value on subsequent runs to only select records that have been modified since the latest last modified timestamp value seen so far. In this way, only new or recently modified records should be processed. Scribe Online was persisting the last modified timestamp down to the second -- it was not persisting millisecond-level information. If the source system supports sub-second timestamps, Scribe Online was thereby selecting that last record multiple times. Scribe Online now correctly persists the last modified timestamp with millisecond-level precision.
  • Updated Agents are not able to connect to Microsoft SQL Azure (Customer Ticket #40130). Updated Agents received an error that an index array was out of bounds when attempting to establish a connection to Microsoft SQL Azure. Newly installed Agents worked correctly. Microsoft recently released updated versions of their code libraries which are utilized by the Agent. However, the latest versions of those code libraries were not installed as part of the Agent update process. With this new Agent release, all Agents will have the latest Microsoft libraries and are able to connect to Microsoft SQL Azure.
  • Microsoft Dynamics CRM schema updates are not automatically detected. Microsoft Dynamics CRM has the ability to indicate when schema-level information has changed. In the August 2011 release of Scribe Online, Scribe added the capability for Scribe Online RS to recognize when a schema change occurred, and update the schema of the target database automatically. With the December 2011 release, the feature was mistakenly removed, requiring you to manually choose to refresh metadata from the Connections page. With this new Agent release, the capability to automatically detect schema changes in Microsoft Dynamics CRM and update the target schema automatically is restored.
  • Refresh metadata occasionally does not update Scribe Online. This issue was caused by the Agent caching certain information. Depending on the specifics of the situation, deleting the cached copy of the metadata stored on the agent machine will resolve the issue. In other cases, credential information was not updating after the connection was edited. Stopping and restarting the Agent addressed this situation. With this latest Agent release, metadata updates should reliably update in Scribe Online without the need to manually delete the metadata cache file or to stop and start the Agent.
  • When a Scribe Online SYS Map performs an update or delete operation that would result in multiple records being affected, Scribe Online gives a confusing error message. Match fields are intended to uniquely identify a record in the target entity. If an update or delete operation would impact more than one record, the match field is not enough to uniquely identify this record. Such an occurrence generally indicates a design problem with the Map and as such Scribe Online correctly fails the record. However, the error message presented in this situation indicated that an index was out of bounds, which was not meaningful to the user who built the map. With this Agent release, this situation will result in a clearer message in the Execution History explaining the reason the record failed.




Posted by Scribe Support on Feb 3 2012 2:55PM { U359 , F35 , T2416 , C7845 }