Bookshelf Home | Contents | Index | PDF |
Global Deployment Guide > Deploying with Global Time Zone > Converting Historical Data to UTC > Preparing Your Data for Conversion to UTCBefore you convert your data to UTC, you need to prepare it. This task is a step in Process of Converting Historical Data to UTC.
NOTE: Log space is controlled through transaction logs. On DB2 UDB, this is called transaction log. On Oracle, this is called rollback segment. On MS SQL Server, this is called log file. Set Time Zones for UsersFor each of your users, specify the time zone in the Contacts screen. This data is stored in the S_CONTACT table. You must also specify a value for the Default Time Zone system preference. For more information, see Setting UTC System Preferences. Save a Report of User Time ZonesPrior to running the UTC conversion utility, save a report of your user time zones as a record of the input data used during the conversion. Edit the driver_utc.ucf File to Specify More Input FilesThe file driver_utc.ucf identifies the input files for UTC conversion. If you create additional input files, you need to add them to driver_utc.ucf. A default set of input files is provided. You may need to create additional files to specify more columns to convert. An example from the driver_utc.ucf file appears as follows: File Name = $SiebelRoot\bin\utcupgd Command Line = /u $UserName /p $Password /c "$ODBCDataSource" /d $DatabaseOwner /n "$RepositoryName" /g $Language /x $DatabasePlatform /j $SiebelRoot/bin/s_camp_con_01.inp /l $SiebelRoot/log/s_camp_con_01.log /s $SiebelRoot/bin/server_time.inp Number of 10 Second Wait Loops = 2000 To edit driver_utc.ucf file parameters to accommodate additional input files
Modify Default Input Files As NeededModify the default input files for the UTC conversion utility as appropriate for your deployment, or create additional input files. An example of customizing input files to partition data into separate input files is provided later in this section. See also About the UTC Conversion Utility. Partitioning Prerequisite for OracleBy default, some tables are partitioned by making use of month (in the Modify Input Files for Partitioned TablesAs the UTC conversion utility processes input files, it typically processes each table as a whole. However, tables with very large record counts may encounter errors due to constraints on log space at the database level. The UTC conversion utility prevents errors that could occur due to insufficient log space at the database level, by using multiple input files to partition large tables into subsets of records for processing. The utility updates each record set individually to convert all rows in a partitioned table. The UTC conversion utility uses partition keys to control how a table is divided into record sets. For example, large tables may be divided based on the calendar month in which each record was created, resulting in twelve approximately equal-sized partitions. Partition keys are supplied for tables that are typically very large and that generally use a lot of log space if updated as a single input file. The tables that are delivered with partition keys are: The default value for each key is customizable. You determine the way that your tables are partitioned, and you can partition your own tables that you know to have large record counts by adding or modifying the input files. NOTE: If you require a different partitioning method, or if you want reduced partitioning in order to optimize performance, contact Siebel Technical Support or Expert Services for assistance. If you create additional partitioned files, you may decrease performance. The input file includes a The following example is from the file s_evt_act_00.inp. This particular file is used to define one partition of the activities table that includes all records created in the month of January (month=1) or February (month=2). [S_EVT_ACT] NOTE: If the Allocate Maximum Log SpacePrior to running the UTC conversion utility, set the log space parameters on the database server to the maximum. The utility requires a large amount of log space in order to run properly. For Oracle, allocate a single large rollback segment and take other rollback segments offline to make sure that large transactions succeed. |
Global Deployment Guide |