How big does IM's Data Manager database need to be?

How big does IM's Data Manager database need to be?

book

Article ID: KB0089148

calendar_today

Updated On:

Products Versions
TIBCO IntegrationManager -
Not Applicable -

Description

Resolution:
The two tables of concern are the job and dupkey tables. Other Data Manager tables typically do not grow very much.

The job table stores checkpoints, so its size depends on how much data is put on the job. Checkpoints are removed when the job ends. So the job table must be big enough to hold checkpoints for the maximum number of jobs that could be in the engine at any given time. This equals the number of running jobs plus flow controlled jobs (if JC is checkpointed), which in turn depends on the rate of incoming messages.

If duplicate detection is used, IM stores a row per dupkey in the dupkey table. The size of each row is pretty small, but the row typically stays in the table for a long period (user configurable) and hence the table needs to be big enough to retain this information. So the amount of time a row stays in the table combined with the rate at which rows are inserted into the table (which depends on how quickly messages arrive) will determine the size of this table.

Issue/Introduction

How big does IM's Data Manager database need to be?