From: Edward Sanborn (
Date: Wed Apr 10 1996 - 15:35:45 CDT


  Here is where you can pick up vsar:

  Here's the docs(thanks to

 * Copyright (c) 1992, 1995 by Sun Microsystems, Inc.

Command line options:

-i <interval> Sample collection interval. (All output values are
                always per second irrespective of the sample interval)

-l Lock down vsar pages in memory.

-d <dumpfile> Dump output file name. Default is "vsar.out"

-D <interval> Dump output to file every "interval" seconds.

-r Reset counters after dump

-p <N> Display top N processes by time consumed in last sample.

-k Display kmem information on a per pool basis.

Keystrokes when "vsar" is active:

n/N Go to next screen
p/P Go to previous screen
^D Scroll up half screen
^U Scroll down half screen
j/J Scroll up one line
k/K Scroll down one line

^L Redraw the screen

m/M turn on/off per pool kmem allocation information

w/W turn on/off detailed active processes information
                Number of procs displayed is NCPU * 2 if none provided
                in the command line with -p option.

d/D Dump the current screen image to file <dumpfile>.0000
                (next d command will dump it to file 0001, 0002 and so
r/R Reset all counters (Need to do this if you add/remove

! Escape to user shell. vsar keeps running in the
                (will have to do a ^L to redraw the screen once you come
                back into vsar)

Display information:

Top left corner Sample collection start time or last reset time.
Center Vsar version # Hostname & uptime
(N.NN) Length of the last sample in seconds
Top Right corner Current time

All delta values are per second deltas irrespective of the interval

The second row under the label is the average value. This average is
computed between the time on the top left and right corners of the

CPU, Net and Disk values are per column.

Label Description

CPU Logical cpu number
ID Physical CPU id
%usr % time spent in the last sample in user mode
%avg % average time spent in the user mode (since collecting
%sys % time spent in the last sample in system mode
%avg % average time spent in the system mode (since .. )
%wio % time spent in the last sample in idle waiting for IO
%avg % average time spent in idle waiting for IO (since .. )
%idle % time spent in the last sample in idle
%avg % average time spent in idle (since .. )
MadpE Adaptive mutex enters for this cpu
avg Average adaptive mutex enters for this cpu
Average System wide average for all categories

Proc information (-p N or 'w') key.
(NOTE: With the 'w' the # of processes displayed is NCPU *2.)

Pid unique process id of the process
User login name for that uid
S state of the process
                        S - on processor
                        S - sleeping
                        T - being traced
                        Z - defunct (zombie)
                        R - on runq
                        I - being created
CLS Class
                        TS - Time Share
                        RT - Real Time
                        SYS - System Process

Sz(M) size of the process in Mb
Rsz(M) resident set size of the process in Mb
Time total cpu time consumed in minutes and seconds
%CPU percentage time spent executing in the last sample
Syscall last system call executed by this process
TTY controlling tty device under /dev directory
Process name of this process

(Entries are reverse sorted by the %cpu taken in the last sample
and for among all the processes that have not run in the last sample
they are sorted based on the cpu time they have consumed)

Physmem Physical memory available to the kernel at boot time in
freemem Free memory in MB.

SwapDevM Swap device space Max in MB
SwapDevF Swap device space Free in MB

SwapResv Swap Space reserved (MB)
SwapAvail Swap Space available (MB)
SwapAlloc Swap Space allocated (MB)
SwapFree Swap Space free (MB)

                                        (These values are system wide)

traps cpu_sysinfo.trap (traps)
intr cpu_sysinfo.intr (interrupts)
scall cpu_sysinfo.syscall (system calls)

fork cpu_sysinfo.sysvfork+cpu_sysinfo.sysfork (forks +

exec cpu_sysinfo.sysexec (exec system calls)
read cpu_sysinfo.sysread (read system calls)
write cpu_sysinfo.syswrite (srite system calls)
bread cpu_sysinfo.bread (reads from the disk into buffer
bwrit cpu_sysinfo.bwrite (writes to the disk from buffer
lread cpu_sysinfo.lread (reads from buffer cache)
lwrit cpu_sysinfo.lwrite (writes to buffer cache)
pread cpu_sysinfo.phread (raw device reads)
pwrit cpu_sysinfo.phwrite (raw device writes)

asflt cpu_vminfo.as_fault (as faults)
htflt cpu_vminfo.hat_fault (hat faults)
mjflt cpu_vminfo.maj_fault (major faults)
cwflt cpu_vminfo.cow_fault (copy-on-write faults)
prflt cpu_vminfo.prot_fault (protection faults)
ksflt cpu_vminfo.kernel_asflt (kernel as fault)
pgin cpu_vminfo.pgin (page in's)
pgout cpu_vminfo.pgout (page out's)
pgpgi cpu_vminfo.pgpgin (pages paged in)
pgpgo cpu_vminfo.pgpgout (pages paged out)
pswpi cpu_vminfo.swapin (swap in's)
pswpo cpu_vminfo.swapout (swap out's)
zfod cpu_vminfo.zfod (zero fill on demand)

swtch cpu_sysinfo.pswitch (context switches)
runq sysinfo.runque (entries in run q)
swapq sysinfo.swpque (entries in swap q)
iowq sysinfo.waiting (threads waiting for IO to
namei cpu_sysinfo.namei (name to inode translations)
iget cpu_sysinfo.ufsiget (ufs inode get operations)
ipg cpu_sysinfo.ufsipage (inodes found with pages)
inopg cpu_sysinfo.ufsinopage (inodes found without pages)
dblk cpu_sysinfo.ufsdirblk (directory block access)
idlet cpu_sysinfo.idlethread (idle thread scheduled to run)
inctx cpu_sysinfo.inv_swtch (involuntary context switch)
migr cpu_sysinfo.cpumigrate (thread migrated to another cpu)
intrb cpu_sysinfo.intrblk (intrrupt thread blocked)

MadpE cpu_sysinfo.mutex_adenters (adaptive mutex enters)
rwrdf cpu_sysinfo.rw_rdfails (reader-writer locks read fails)
rwwrf cpu_sysinfo.rw_wrfails (reader-writer locks write
dfree cpu_vminfo.dfree (pages freed by daemon (swapped
dscan cpu_vminfo.scan (pages scanned by daemon to be
rcvi cpu_sysinfo.rcvint (serial port rcv interrupts)
xmti cpu_sysinfo.xmtint (serial port transmit
mdmi cpu_sysinfo.mdmint (modem interrupts ???)
rawch cpu_sysinfo.rawch (raw characters processed)
canch cpu_sysinfo.canch (canonical characters processed)
outch cpu_sysinfo.outch (characters output)
msgs cpu_sysinfo.msg (IPC:message operations)
sems cpu_sysinfo.sema (IPC:semaphore operations)

KemS/Max(K) Total Memory (Small pool) being used/Total allocated
KemL/Max(K) Total Memory (Large pool) being used/Total allocated
KemO/Max(K) Total Memory (Oversize pool) being used/Total allocated
FailS Total Memory allocation failures (Small pool)
FailL Total Memory allocation failures (Large pool)
FailO Total Memory allocation failures (Oversize pool)
nfile Total number of open files in the system
proc/max Total active processes/Max limit in system
povf Total Number of proc table overflows

For >= 2.4 systems:

Perm/Kmem Permanently allocated/Kmem allocated in MB
calloc kmem cache allocs per second
halloc kmem hash allocs per second
palloc kmem permanent allocs per second
slabc kmem slab creates per second
total total individual chunks of kmem allocated
max max individual chunks of kmem allocated
afail total kmem allocations failures since boot
load1 load average for the last 1 minute
load5 load average for the last 5 minutes
load15 load average for the last 15 minutes

inos/max Inodes in use/Max inodes configured in the system
hits Inodes found in the cache
miss Inodes not found in the cache
%hit Percentage hit rate
%miss Percentage miss rate
alloc Memory allocated for new inodes
free Memory freed from old inodes
freef Inode freed to the front of the freelist
freeb Inode freed to the back of the freelist

For 2.4 systems (running -k option or with 'm' turned on):

Size Size of the kmem cache pool
alloc Allocations from that pool in the last sample
afail failures in allocation from this pool
bftot total number of allocated entries from this pool
bftot total memory allocated in this pool (Mb)
bfmax max number of allocated entries from this pool (Mb)
bfavl available buffers in this pool (Mb)
bfrep reapable buffers in this pool (Mb)
name name of this kmem cache pool

Net Network device name
ipkts Number of input packets
avg Average number of input packets
opkts Number of output packets
avg Average number of output packets
colli Number of collisions
avg Average number of collisions
ierr Number of input packet error in last sample
oerr Number of output packet erros in last sample
frm Number of frame error in last sample
crc Number of CRC errors in last sample
uflo Number of underflow errors in last sample
oflo Number of overflow errors in last sample

Disk Disk number
%busy % time busy in the last sample
%avg % average busy
avque Average queue length
await Average wait times (milli seconds)
aserv Average service times (milli seconds)
nrd Number of reads from the disk
avg Average number of reads
nwr Number of write to the disk
avg Average number of writes

rd(k) Bytes read (KB)
wr(k) Bytes written (KB)

Avg. Average for %busy, %avg, avque, await and aserv
                Total for nrd, avg, nwr, avg, rd(k) and wr(k)

Edward A. Sanborn         Senior Network Engineer
Datacube, Inc.	               (508) 777-4200 x.3067 
300 Rosewood Drive
Danvers, MA. 01923        ~~~~~~~~~~~~~~~~~~~~~

This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:10:57 CDT