|
|
coherent
sa Command sa
Process accounting
ssaa [-aabbccjjllmmnnrrssttuu][-vv _N][_f_i_l_e]
One of the accounting mechanisms available on the COHERENT system
is process accounting (also called shell accounting), which
records the commands executed by each user. The command accton
enables or disables shell accounting.
The command sa scans the accounting information in file and
prints a summary. If file is omitted, it reads the file
/usr/adm/acct by default. For each command executed, sa prints
the number of calls made, the total CPU time (user and system),
and the total real time. The output is ordered by decreasing CPU
time.
sa recognizes the following options:
aa Place commands executed only once and command names with un-
printable characters in the category ``***other''.
bb Sort by average CPU time per call.
cc Also print CPU time as a percentage of all CPU time used.
jj Print average times per call rather than totals.
ll Separate user and system time.
mm Accumulate information for each user rather for each command.
nn Sort by number of calls.
rr Reverse the order of the sort.
ss After scanning, condense the accounting file and merge it
into the summary files.
tt Also print the CPU time as a percentage of real time.
uu Print the user and command for each accounting record; this
option overrides all others.
vv _N For commands called no more than N times, where N is a digit,
sa asks whether to place the command in the category
``**junk**''.
sa uses the summary files /usr/adm/savacct and /usr/adm/usracct
to lessen disk usage.
***** Files *****
/uussrr/aaddmm/aacccctt -- Default account data
/uussrr/aaddmm/ssaavvaacccctt -- Summary
COHERENT Lexicon Page 1
sa Command sa
/uussrr/aaddmm/uussrraacccctt -- Summary
***** See Also *****
ac, acct(), acct.h, accton, commands
***** Notes *****
The file /usr/adm/acct can become very large; therefore, you
should truncate it periodically. Special care should be taken if
process accounting is enabled on a COHERENT system with limited
disk space.
COHERENT Lexicon Page 2
This archive runs on limited infrastructure. Preserving old code on modern bandwidth. Automated agents are requested to crawl responsibly.