KADC(8) Maintenance Commands and Procedures KADC(8)

NAME


kadc - data collector for kar

SYNOPSIS


/usr/lib/ka/kadc

DESCRIPTION


The kstat activity reporter, kar, extends sar (the system activity
reporter) by saving all kstats regularly, rather than a predefined
subset. This allows a much wider range of statistics to be monitored,
without having to choose which statistics to be monitored in advance.

The kadc script should be run regularly from the sys user's crontab.
Every 5 minutes is suggested. Less frequently means that data may not
be collected with sufficient granularity, while more frequently may
generate too much data to be stored.

The saved data may be analysed with the kar(1) utility.

CRONTAB ENTRIES


To collect activity data every 5 minutes, the sys user's crontab
could contain an entry like:

0,5,10,15,20,25,30,35,40,45,50,55 * * * * /usr/lib/ka/kadc


The collection process requires no special privileges, so there is no
need t run the collection process as any particular user. The user
performing the collection should be the owner of the /var/adm/ka
directory.

You should also create a system for automatically archiving or
removing old data. A simple kaclean script is provided, which cleans
up records over 30 days old, and which can be added to the sys
crontab as follows:

1 1 * * * /usr/lib/ka/kaclean


FILES


The data archives are zip files named by date, normally stored in the
/var/adm/ka directory. The zip archives contain individual files for
each reporting interval, the file being a dump of the entire
kstat(3KSTAT) hierarchy in JSON format.

SEE ALSO


kar(1), kstat(8), kstat(3KSTAT)

March 7, 2025 KADC(8)

tribblix@gmail.com :: GitHub :: Privacy