Support maintenance processes
In order to support regular maintenance processes, a MAINTENANCE section has to be defined. The section defines the periodicity of maintenance processes and the processes to be executed.
A protocol for maintenance processes running during a maintenance is written to %TRACE%/maintenance.log.
The example below shows the configuration for a daily server maintenance during night starting at 2:30. Two additional sections required by maintenance processes had been added.
[MAINTENANCE]
CYCLE_TIME=24
KILL_TIME=60
START=2:30
; kill clients still running. Sending a message to active clients and wait 1 minute
RUN01=KillAll Wait=60
; process creates a backup for the database
RUN02=DBBackup DB=VerA_Dat; BF=%RDB_ROOT%\Backup\VerA.ozi;
; the procedure appends a time stamp to the file name
RUN03=Run C=%RDB_ROOT%\DateCopy.cmd; P=%RDB_ROOT%\Backup\VerA.ozi
; checking database consistency
RUN04=DBCheck DB=VerAMaint; P=-C:IXG -R;
; create new replication version when transaction number exceeds 90% of transaction limit
RUN05=DBReplication Database=VerAMaint; Operation=enable; TransactionLimit=10000; Condition=90;
; compress database
RUN06=DBPack DB=VerA_Dat;
; for test, only in order to restor the original state at the end
RUN07=DBRestore DB=VerA_Dat; BF=%RDB_ROOT%\Backup\VerA.ozi;
[VerAMaint]
DICTIONARY=%RDB_ROOT%\VerA.dev
DATABASE=%RDB_ROOT%\VerA.dat
SHARE=NO
ACCESS_MODE=Write
ENABLE_CONTEXT=NO
;KEY_FILE=e:\temp\db.key
Subtopics
- CHECK_TIME - Time waiting for next time check
- CYCLE_TIME - Cycle time for running maintenance processes
- START - Time for starting first maintenance
- KILL_TIME - Time for waiting
- RUNnn - Process definitions