| 1 |
On Sun, 30 Nov 2008 14:27:22 +0200 |
| 2 |
Alan McKinnon <alan.mckinnon@×××××.com> wrote: |
| 3 |
|
| 4 |
[-snip-] |
| 5 |
|
| 6 |
> |
| 7 |
> You are the one making them claims! You do the test! |
| 8 |
|
| 9 |
|
| 10 |
Me? No, my dear. It is you who claims that fragmentation only matters |
| 11 |
on MS file systems while Linux FS remain unaffected w/o providing |
| 12 |
any proof. The common sense says otherwise: fragmentation leads to |
| 13 |
seeks, seeks hurt performance. |
| 14 |
|
| 15 |
Anyways, I'm going to do the tests because I want to see the numbers. |
| 16 |
Should the outcome prove you were right I'd be the first to |
| 17 |
congratulate and thank you for opening my eyes. |
| 18 |
|
| 19 |
Since I'm going to use my workstation at work to do the job at night |
| 20 |
I'll limit the repetitions to 400 which would take about 13 hours with |
| 21 |
two files, 2GB each. The machine has 1GB RAM installed and I'll run the |
| 22 |
test in Gentoo's "boot" rc level. I've closed everything except sshd |
| 23 |
and screen (see "ps" outpput below) and have taken care of the cache |
| 24 |
by clearing it before each run via /proc (see the script). The file |
| 25 |
system is ext3 and during the test it will be used in about 30% of its |
| 26 |
capacity: |
| 27 |
|
| 28 |
localhost test # df -h | grep home |
| 29 |
/dev/sdc1 15G 718M 14G 5% /home |
| 30 |
|
| 31 |
localhost test # mount -v | grep home |
| 32 |
/dev/sdc1 on /home/ type ext3 |
| 33 |
(rw,noatime,nodiratime,data=journal,commit=1) |
| 34 |
|
| 35 |
|
| 36 |
localhost test # rc-status |
| 37 |
Runlevel: boot |
| 38 |
bootmisc [ started ] |
| 39 |
checkfs [ started ] |
| 40 |
checkroot [ started ] |
| 41 |
clock [ started ] |
| 42 |
hostname [ started ] |
| 43 |
iptables [ started ] |
| 44 |
localmount [ started ] |
| 45 |
modules [ started ] |
| 46 |
net.eth0 [ started ] |
| 47 |
net.lo [ started ] |
| 48 |
sshd [ started ] |
| 49 |
udev-postmount [ started ] |
| 50 |
|
| 51 |
|
| 52 |
localhost test # ps af |
| 53 |
PID TTY STAT TIME COMMAND |
| 54 |
12191 pts/0 Ss 0:00 -bash |
| 55 |
12227 pts/0 S 0:00 \_ su - |
| 56 |
12231 pts/0 S 0:00 \_ -su |
| 57 |
12266 pts/0 S+ 0:00 \_ screen -x |
| 58 |
8575 pts/3 Ss+ 0:00 -/bin/bash |
| 59 |
4967 pts/2 Ss 0:00 -/bin/bash |
| 60 |
12830 pts/2 R+ 0:00 \_ ps af |
| 61 |
4895 pts/1 Ss+ 0:03 -/bin/bash |
| 62 |
4755 tty6 Ss+ 0:00 /sbin/agetty 38400 tty6 linux |
| 63 |
4754 tty5 Ss+ 0:00 /sbin/agetty 38400 tty5 linux |
| 64 |
4753 tty4 Ss+ 0:00 /sbin/agetty 38400 tty4 linux |
| 65 |
4752 tty3 Ss+ 0:00 /sbin/agetty 38400 tty3 linux |
| 66 |
4751 tty2 Ss+ 0:00 /sbin/agetty 38400 tty2 linux |
| 67 |
4750 tty1 Ss+ 0:00 /sbin/agetty 38400 tty1 linux |
| 68 |
|
| 69 |
|
| 70 |
The script: |
| 71 |
|
| 72 |
#!/bin/bash |
| 73 |
|
| 74 |
cat /usr/portage/distfiles/* > test1 |
| 75 |
cp test1 test2 |
| 76 |
|
| 77 |
filefrag test* |
| 78 |
|
| 79 |
sync |
| 80 |
|
| 81 |
for (( i=0 ; i<=400 ; i++ )) |
| 82 |
do |
| 83 |
|
| 84 |
sync |
| 85 |
echo 3 > /proc/sys/vm/drop_caches |
| 86 |
|
| 87 |
echo -n "try_no=$i " |
| 88 |
|
| 89 |
/usr/bin/time -f "\ |
| 90 |
command=%C|\ |
| 91 |
real_t=%e|\ |
| 92 |
kernel_t=%S|\ |
| 93 |
user_t=%U|\ |
| 94 |
major_faults=%F|\ |
| 95 |
minor_faults=%R|\ |
| 96 |
context_sw=%c|\ |
| 97 |
io_waits=%w|\ |
| 98 |
fs_reads=%I|\ |
| 99 |
fs_writes=%O" cp test1 /dev/null |
| 100 |
|
| 101 |
sync |
| 102 |
echo 3 > /proc/sys/vm/drop_caches |
| 103 |
|
| 104 |
echo -n "try_no=$i " |
| 105 |
|
| 106 |
/usr/bin/time -f "\ |
| 107 |
command=%C|\ |
| 108 |
real_t=%e|\ |
| 109 |
kernel_t=%S|\ |
| 110 |
user_t=%U|\ |
| 111 |
major_faults=%F|\ |
| 112 |
minor_faults=%R|\ |
| 113 |
context_sw=%c|\ |
| 114 |
io_waits=%w|\ |
| 115 |
fs_reads=%I|\ |
| 116 |
fs_writes=%O" cp test2 /dev/null |
| 117 |
|
| 118 |
|
| 119 |
done |
| 120 |
|
| 121 |
###########EOF |
| 122 |
|
| 123 |
|
| 124 |
|
| 125 |
I hope all this will be enough for you to accept the results. |
| 126 |
|
| 127 |
|
| 128 |
|
| 129 |
-- |
| 130 |
Best regards, |
| 131 |
Daniel |
Archives