1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
|
@(#)README 1.6 89/07/07 Legato Systems, Inc.
This directory contains the source for the nhfsstone (pronounced
n-f-s-stone, the "h" is silent) NFS load generating program. This
version of the program can only be compiled on 4.x BSD based UNIX
systems.
nhfsstone is used on an NFS client to generate an artificial load
with a particular mix of NFS operations. It reports the average
response time of the server in milliseconds per call and the load in
calls per second. The program adjusts its calling patterns based on
the client's kernel NFS statistics and the elapsed time. Load can be
generated over a given time or number of NFS calls. See the "nhfsstone.1"
manual page for more details.
The files in this directory are:
DISCLAIMER legal requirements
Makefile Makefile used to build nhfsstone
README This file
nhfsstone.c source file
nhfsstone.1 manual page
nhfsrun shell script to run nhfsstone over multiple loads
nhfsnums shell script to convert nhfsrun output to plot(5)
nhfsgraph shell script to create a graph from nhfsnums output
The file "nhfsstone.1" is a manual page that describes how to use the
nhfsstone program. To look at it type "nroff -man nhfsstone.1".
To build an executable type "make nhfsstone". To install it, become
super-user and then type "make install". This will strip the
executable, set the group to "kmem" and set the setgid bit. If your
site requires different installation of programs that read /dev/kmem
you may have to use different ownership or permissions. Make install
will also set the execute bits on the shell scripts nhfsrun, nhfsnums
and nhfsgraph.
To run an nhfsstone test, create a parent test directory on a filesystem
that is NFS mounted, cd to that directory and type "nhfsstone". This will
do a run with the default settings, load = 30 calls/sec, 5000 calls,
and 7 sub-processes.
If you want to spread the load across several server disks, first
figure out on the server which disk partitions are exported as which
filesystems. If you don't already have more than one of these
filesystems mounted on your test client you can mount them in temporary
locations, like /mnt. Create test directories on these filesystems so
that the load will be distributed according to the simulation that you
want to run (for example, you might put 4 test directories on the
filesystem where the diskless client's root and swap live, and 2 on the
home directories filesystem, and one on the executables filesystem).
Now create a parent test directory cd to it, and make symbolic links
with the names testdir0, testdir1, ... testdir6, that point to the
real test directories. Finally, run nhfsstone from the parent test
directory.
If you are doing the test from a diskless machine, putting half of the
test directories in /tmp or /usr/tmp and running the test from your
home directory will simulate real diskless load patterns fairly well.
To do a run over multiple load levels, edit the shell script "nhfsrun" and
set the shell variables "START", "END", and "INCR" to be the correct
starting and ending loads, and load increment. The script will iterate
from START to END with an increment of INCR, run nhfsstone at each
load level, and put the output in the file "run.out". The output file
name can be changed by editing the nhfsrun script and changing the
"OUTFILE" variable or by passing a file name suffix on the command line:
nhfsrun xysd
This produces the output file "run.xysd".
The script "nhfsnums" takes the output from nhfsrun and converts it
into plot(5) format so that it can be graphed using graph(1) and other
tools. It takes its input either from files given on the command line
or from standard in:
nhfsnums [numsfile] ...
If file names are given, the suffix of each name (the part after the
".") is used as the line label for the set of numbers produced (see
plot(5)).
"nhfsgraph" takes the output from nhfsnums and passes it to graph(1)
with the right arguments to produce PostScript output for a labeled
graph. The nhfsgraph script can be used as a filter:
nhfsnums run.* | nhfsgraph | lpr
This program is provided free of charge to anyone who wants it provided
certain conditions are met (see DISCLAIMER file for more details).
If you would like to receive regular information and bug fixes please
send your name, and both your Email and U.S. mail addresses to:
Legato Systems, Inc.
Nhfsstone
260 Sheridan Avenue
Palo Alto, California 94306
nhfsstone-request@legato.com or uunet!legato.com!nhfsstone-request
and we will add your name to the nhfsstone mailing list. Comments and bug
reports should be sent to:
nhfsstone@legato.com or uunet!legato.com!nhfsstone
|
