Other Kernel Configuration Operations

Other sections below describe some special kernel configuration operations and special uses of the kernel configuration commands.

The usage of some kernel resources can be monitored, with alarms delivered when usage rises above a set threshold. These alarms can be configured and reviewed using the kcalarm command or the HP SMH tool. The resource usages can be reviewed using the kcusage command or HP SMH. For more information, see “Monitoring Kernel Resource Usage”.

Administrators of older versions of HP-UX may be accustomed to using text files (“system files” or “dfiles”) to specify kernel configurations and make changes to them. The format of these files has been enhanced to accommodate new kernel configuration innovations, while retaining the usefulness of a text file for configuration operations. They are particularly useful when using the same configuration on multiple systems, since they can be easily moved between systems. The use of system files is described in “Managing Configurations with System Files”. The system file formats from previous releases of HP-UX are still accepted.

All kernel configuration changes made using the kernel configuration commands are logged to the file /var/adm/kc.log. Details about this log file can be found in “The Kernel Configuration Log File”, and the kconfig(5) and kclog(1M) manpages.

The primary kernel configuration commands support a specialized output format that is designed for use by scripts and applications that need to parse the output of the commands. Such scripts and applications must use this specialized output format since HP does not guarantee release-to-release compatibility for any other output format of these commands. More detail is available in “Parsing Command Output” and the kconfig(5) manpage.

It is possible to have an undesirable, or even unbootable, kernel configuration because of mistaken configuration changes, hardware failures, or software defects. Mechanisms exist both to prevent such problems and to help recover from them. For more details see “Recovering from Errors”.