This page collects general things to consider for getting Data Archiving into operation.
Concepts for Data Archiving are described in the section About Data Archiving.
The data archive process is controlled by two System Parameters. Setting Data Archive process startup interval and Data Archive process startup, starts the data archiving process.
Parameter | Description |
---|---|
Data Archive process startup interval (seconds) | defines how often the data archive process wakes up and looks for data archive orders to execute. |
Data Archive process startup | determines whether the data archive process should be started or not. |
The data archive process gets start date and start time when you save the
parameter values. Using dbms_job.next_date(job, next_date)
can
change the next executing date for the data archive process.
When started it reads Data Archive process startup interval to determine how often it should wake up and puts an entry in Oracle’s job scheduler.
When stopped it removes the entry from Oracle’s job scheduler.
Using the Administration/Monitoring/Database Processes in Solution manager, you can monitor the data archive process. If anything goes wrong in the data archive process that can’t be handled by normal error handling, the error is written in Oracle’s alert log file.
The data archive process can be monitored using Runtime Monitoring Server
Processes form. The data archive process name is
Data_Archive_Util_API.Process_Archive_(0)
. It is also important to look
in Oracle’s own alert.log file for fatal errors that can’t be handled
inside the data archive process.
Below follows a description of how to install a running data archive object:
When archiving a data archive object the first time on a production database, where it exists a lot of objects to archive, it is wise to start with calculating the time it would take to execute an exact amount of objects. Then you can calculate an approximation on how long the whole operation will take. This can be accomplished by doing the following:
”SELECT count(*) FROM master_table
” + the master table
where-clause and eventually execution parameter values.
You can see the SQL statement generated in Data archive object form with
method Check Master Statement.
’AND rownum <= x
’, where x is a number you choose
depending on how many objects you want execute and see how long it takes to
execute them. Execute the data archive order and look in the data archive
log how long it took to execute the data archive object.
It is a good idea to test the data archive execution in an test environment first, or run the data archive execution for only one data archive object to verify that it does the right thing.
Remember that executing the data archive object with data archive type set to Move, Copy or Remove will take some more time depending on network speed (Oracle tables in another Oracle instance) and how much data there is in the data archive object tables.
If the calculated time for the first execution of the data archive objects is longer than the scheduled time for batch jobs, it is a good idea to restrict the master where-clause so the data archive process finishes in time. Perhaps the data archive object must be run with the restricted where-clause for a while before the data archive process has catch up with the production of new objects to archive.
Data archive objects, data archive orders and data archive order executions has a parameter called active connected to them. If the data archive process should be able to process a data archive execution all the three objects active parameters must be set to TRUE (checked) otherwise the data archive process don’t process the data archive order execution.
The data archive order execution is not executed if any of the parameters is set to FALSE (unchecked).