SQL/MX Release 2.0 Best Practices

Database Sizing Considerations 19
Network Data
For each legacy system involved:
• What is the network connection?
• Are multiple links to the legacy systems needed or desired?
• What is the preferred transfer mechanism—for example, file transfer protocol (FTP) or network data
management protocol (NDMP)?
• Does the transfer mechanism support the anticipated volumes and cycles of updates? (This
consideration is not really a sizing parameter, but rather a “sanity check” regarding the preferred
mechanism.)
Operational Data
For each legacy system involved:
• What is the system platform, operating system, and database or file system?
• What are the system-maintenance windows?
For the Database System
• What is the system availability (for example, 9 x 5, 15 x 7, or something else)?
• How much data will be archived, and what is the archive cycle?
Work and Swap Space Considerations
• Establish effective, consistent environmental settings for users of both MXCS clients and ODBC clients,
including the clients using third-party tools.
• Designate specific disk volumes for workspace requirements—specifically, the sort workspace and
temporary workspace for query-process components, when possible.
• Distribute this information to interactive users of MXCS clients, or have the information loaded into
their HP Open System Services (OSS) segments at logon.
• Add this information to the ODBC/MX configuration.
• Monitor workspace usage over time to determine if and when additional space is required. To get
helpful information about space usage, you can use the HP Event Management Service (EMS) logs,
Availability Stats and Performance (ASAP), Disk Space Analysis Program (DSAP), and the HP Measure
product.
About SQL/MX Scratch Disks
NonStop SQL/MX selects a disk to be used for a scratch file from the pool of available disks. The pool
initially consists of the set of all suitable disks. Disks such as optical disks, phantom disks, and SMS virtual
disks are not considered suitable. The disks specified by the SCRATCH_DISKS_EXCLUDED control are
removed. If the SCRATCH_DISKS control is specified, the disks that are not specified in the SCRATCH_DISKS
control are removed from the pool. From this disk pool, a disk is selected based on this criteria:
• The amount of used space on the disk. (rank * 30)
• The number of scratch files on the disk. (rank * 70)
• The number of fragments on the disk. (rank * 20)
• The biggest available fragment on the disk. (inverted rank * 80)
• Is the disk a preferred disk? (10000)
• Is the disk the primary disk of the CPU of this process? (100000)
The value in parentheses indicates the weighting of that criterion. The rank is the ordinal rank of that disk
among all the disks in the pool based on the criterion. The inverted rank is the inverted ordinal rank. In the
case of the biggest available fragment criterion, if the pool contains 20 disks, the disk with the biggest
available fragment would have an inverted rank of 20. The weights are summed for all the disks in the
pool, and the disk with the biggest weight is selected. As can be seen, the primary disk of the current CPU
is given a large weight.