It is important to consider the following variables when determining the resources needed on the SQL server that houses the JAMS database.
- Are the jobs evenly distributed throughout the day?
- What is the peak job count in any one hour of the day?
- How much history will you be keeping?
- What is the mix of resubmitting jobs, repeating jobs, scheduled jobs, triggered jobs and manually submitted jobs?
- How many dependencies will each job have?
- How long will you be keeping completed jobs in the monitor?
- What will your schedule advance be?
- What is an acceptable response time?
- Will you be using solid state disks?
- Will you be using RAID disks?
Based on the answers to these questions and the amount of activity per day you plan on having it is possible to make a rough estimate how what resources would be required. There are many variables to be considered when determining the efficiency of the database. We offer the following guidelines at both the minimum and on the high end, for heavier implementations.
- The bare minimum would be a single core processor and at least 4GB of RAM, with at least 10GB of hard disk space plus whatever is required by the OS.
- For installations running over 100,000 jobs per day we recommend Dual Quad-Core CPUs and 64GB of RAM with multiple high speed SCSI disks.
- For database planning, the JAMS test lab runs approximately 1,000 jobs per day using one instance of JAMS. Retaining 30 days of information generates a database approximately 110 Mb in size.
- Monitoring of the system and fine tuning of the server may be necessary in order to provide adequate resources to the SQL back-end.