Subject: bin/2107: sa's ``memory integral'' broken?
To: None <gnats-bugs@NetBSD.ORG>
From: Michael Graff <explorer@flame.org>
List: netbsd-bugs
Date: 02/21/1996 12:16:34
>Number:         2107
>Category:       bin
>Synopsis:       sa's ``memory integral'' broken?
>Confidential:   no
>Severity:       non-critical
>Priority:       low
>Responsible:    bin-bug-people (Utility Bug People)
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Wed Feb 21 13:50:07 1996
>Last-Modified:
>Originator:     Michael Graff
>Organization:
flame.org:  yes, we do know everything
>Release:        20-Feb-1996 <NetBSD-current source date>
>Environment:
	
System: NetBSD packrat.flame.org 1.1A NetBSD 1.1A (PACKRAT) #0: Sat Feb 17 18:06:28 CST 1996 explorer@packrat.flame.org:/usr/src/sys/arch/i386/compile/PACKRAT i386


>Description:
	sa -m produces the following output:

root       4475602     13509.78cpu     26804306tio           0k*sec
daemon         243        77.29cpu        16388tio           0k*sec
postgres      2846        32.30cpu        72942tio           0k*sec
explorer    319858  20433708.81cpu      8694079tio           0k*sec
basilisk       946         8.80cpu        18110tio           0k*sec
flipk          483         1.45cpu         7072tio           0k*sec
gendalia       132         5.89cpu         7821tio           0k*sec
spirit          22         0.09cpu          146tio           0k*sec
feather         79         0.11cpu          449tio           0k*sec
graphix        178         0.41cpu         1316tio           0k*sec
spberry      22467       275.73cpu       429479tio           0k*sec
nsheeley      2344        16.20cpu        56819tio           0k*sec
grpeters        26         0.01cpu          102tio           0k*sec
ccsanady      1691         5.01cpu        19718tio           0k*sec
32001           31         0.06cpu          271tio           0k*sec
ftp          20781       479.42cpu       519009tio           0k*sec
nobody         978        26.77cpu       277204tio           0k*sec

Some of these stats are at the very least questionable, but the 0k*sec
is quite clearly broken.
	
>How-To-Repeat:
sa -m with accounting on.
	
>Fix:
None known yet -- I was going to look into this later next month, but
wanted to remind myself about this, as well as get it in the pr
database in case someone else wants to look at it.  It's not high on
my personal list  ;)
	
>Audit-Trail:
>Unformatted: