SQLite performance and tuning

General zFS performance queries

zFS performance is dependent on many factors. To help you to optimize performance, zFS provides performance information to help determine bottlenecks. You can enter the following system commands to get information about the current operation of zFS:
  • F ZFS,QUERY,SETTINGS
  • F ZFS,QUERY,ALL
To query and reset performance counters, enter the following z/OS® UNIX® System Services command:
zfsadm

Resource Management Facility (RMF™) support for zFS

RMF™ support is available for zFS. When you are considering zFS performance, investigate the zFS components that are involved in I/O processing to or from a zFS file system. In a shared file system environment it is better for performance if you can mount a file system as read-only rather than as read-write. If a file system is mounted as read-write, but is accessed primarily from a single system, such as SYS1, it can improve performance if that file system is z/OS® UNIX® owned on system SYS1.

You can also optimize zFS performance by tailoring the size of its caches to reduce I/O rates and path length. It is also important to monitor DASD performance to ensure that no disc volumes or channels are required to perform beyond the intended capacity.

Monitoring and tuning cache size to improve zFS performance

You can improve the performance of zFS by controlling the size of the caches that hold file system and log data. Monitor the following caches so that you can control them effectively to reduce I/O rates:
  • The user file cache is used for all user files and performs I/O for all user files greater than 7 KB.
  • The metadata cache performs I/O for all user files that are smaller than 7 KB.
  • The log file cache stores file record transactions that describe changes to the file system.