z/OS File System Performance Tool
The File System Performance Tool (FPT), an IBM internal tool, consists of a POSIX application that can simulate a variety of environments though a selection of parameters to access a set of database files using several predefined read/write patterns and ratios.FTP supports the following four file system activities.
- Block random (br)
To use this test mode, the user must specify the size of the block. The minimum value of a block is 1 byte. The maximum value can be the size of the file, however, this is not recommended because of stability issues. Once the block size is established, a random file is picked from the DB set, a random location is picked in that file and a read or a write operation is performed (read or write is determined by a random number generator with a read/write ratio provided by the user). This is done for the entire duration of the test. At the beginning of each iteration, all of the points in the DB set have an equal chance of being selected.
- Block sequential (bs)
This test mode also requires the user to provide the block size. All of the same rules apply as in the block random test mode. Again, for stability reasons, it is not recommended to make the block size equal to the file size. Once the block size is established, a random file is picked from the DB set, a random location is picked in that file (just like in block random). After that, all of the reads are sequential. If an end of file is reached, the next file in the DB set is opened. If the last file has been reached, the first DB file is opened. At the beginning of the test all of the points in the DB set have an equal chance of being selected.
- Character random (cr)
This test mode is identical to the block random except the block size is predefined to be 1 byte.
- Character sequential (cs)
This test mode is identical to block sequential except the block size is predefined to be 1 byte.
- remove existing files? (yes or no)
- number of files to create (0 to 10,000)
- file size for created files (k=kB, m=mB, g=gB)
- number of paths (0 to number of files)
- number of processes (integer specifying the number of simulated users)
- test mode (br=block random, bs=block sequential, cs=character sequential, cr=character random)
- block size (k=kB, m=mB)
- RW ratio (0 to 100)
- run length (s seconds, m minutes, h hours)
- warm-up time (seconds)