Join us on IRC server: irc.freenode.net , channel: #openfiler
You are not logged in.
#1 2010-03-11 12:19:42
- s.borrelli
- Member
Some performance issue
Hi
i have installed an OF platform with 4 sata disk, 2 Gb RAM and intel Atom CPU (i know...this isn't a very good choise, but i whant a "green nas" 
)
i have some question:
1) what kind of performance can i expect from OF?
a first test with iozone give me stranger results:
Code:
[root@openfiler current]# ./iozone -Rab Primotest.wks -i0 -i1 -n 256M -g 4G -f /mnt/volgroup/voltest/TEST/testfile
Iozone: Performance Test of File I/O
Version $Revision: 3.338 $
Compiled for 32 bit mode.
Build: linux
Contributors:William Norcott, Don Capps, Isom Crawford, Kirby Collins
Al Slater, Scott Rhine, Mike Wisner, Ken Goss
Steve Landherr, Brad Smith, Mark Kelly, Dr. Alain CYR,
Randy Dunlap, Mark Montague, Dan Million, Gavin Brebner,
Jean-Marc Zucconi, Jeff Blomberg, Benny Halevy, Dave Boone,
Erik Habbinga, Kris Strecker, Walter Wong, Joshua Root,
Fabrice Bacchella, Zhenghua Xue.
Run began: Thu Mar 11 10:25:48 2010
Excel chart generation enabled
Auto Mode
Using minimum file size of 262144 kilobytes.
Using maximum file size of 4194304 kilobytes.
Command line used: ./iozone -Rab Primotest.wks -i0 -i1 -n 256M -g 4G -f /mnt/volgroup/voltest/TEST/testfile
Output is in Kbytes/sec
Time Resolution = 0.000001 seconds.
Processor cache size set to 1024 Kbytes.
Processor cache line size set to 32 bytes.
File stride size set to 17 * record size.
random random bkwd record stride
KB reclen write rewrite read reread read write read rewrite read fwrite frewrite fread freread
262144 64 29792 52145 1234781 1135830
262144 128 31416 64464 1262571 1162016
262144 256 30030 82072 1194131 1104559
262144 512 27447 83622 987054 928154
262144 1024 24723 60620 843736 798486
262144 2048 58305 53431 806500 770347
262144 4096 37258 81501 790979 770653
262144 8192 29009 49228 800569 769956
262144 16384 21972 55887 810344 772365
524288 64 19083 50481 1237000 1141292
524288 128 19564 48707 1264816 1168616
524288 256 14373 52272 1199230 1101864
524288 512 18997 42696 960924 928663
524288 1024 15692 55035 842206 794724
524288 2048 21805 54909 803273 768738
524288 4096 24207 53060 809766 768717
524288 8192 16431 61469 806408 765324
524288 16384 25766 53501 810010 767218
1048576 64 20521 58565 1270964 1165036
1048576 128 21359 50806 1302501 1192973
1048576 256 13880 48541 1232912 1134001
1048576 512 11621 51595 998883 925207
1048576 1024 15850 50012 861368 807170
1048576 2048 19617 53358 817957 784525
1048576 4096 15118 47124 826207 787879
1048576 8192 14633 45592 829214 790273
1048576 16384 13730 50605 826361 784401
2097152 64 10536 46363 200533 272378
2097152 128 15143 40989 207863 267581
2097152 256 13902 46111 210105 267970
2097152 512 11207 47486 216751 263317
2097152 1024 11693 41909 213706 281585
2097152 2048 12995 41685 217177 284694
2097152 4096 11919 54845 221200 282354
2097152 8192 13291 45025 206670 281172
2097152 16384 15879 52257 207916 280951
4194304 64 12498 44525 243365 278766
4194304 128 11138 45482 240288 284684
4194304 256 12404 46446 238102 283503
4194304 512 11937 43954 239111 293646
4194304 1024 10534 45241 234806 280102
4194304 2048 11419 47404 244926 289171
4194304 4096 11229 42869 241098 277394
4194304 8192 11728 45001 226858 276796
4194304 16384 12741 45472 248702 284926seems that, with a 4 Gb file, the write speed is about ~12 Mb/s and the read speed is about 280 Mb/s
i think that the read speed is too hight, and the write speed is too low. what do you think about this?
the second question is:
i have 4 sata disk: 3 WD10EADS and 1 WD10EARS
this is the iostat output during the iozone test:
Code:
avg-cpu: %user %nice %system %iowait %steal %idle
0.25 0.00 11.63 21.47 0.00 66.65
Device: rrqm/s wrqm/s r/s w/s rsec/s wsec/s rkB/s wkB/s avgrq-sz avgqu-sz await svctm %util
sda 0.00 2.08 0.00 4.86 0.00 55.50 0.00 27.75 11.43 0.02 3.51 0.65 0.32
sda1 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
sda2 0.00 0.10 0.00 0.69 0.00 6.34 0.00 3.17 9.14 0.00 1.14 1.14 0.08
sda3 0.00 1.98 0.00 4.16 0.00 49.16 0.00 24.58 11.81 0.02 3.90 0.67 0.28
sdb 32.01 3450.94 8.92 171.66 328.25 28983.94 164.12 14491.97 162.33 0.55 3.03 1.25 22.52
sdb1 32.01 3450.94 8.92 171.66 328.25 28983.94 164.12 14491.97 162.33 0.55 3.03 1.25 22.52
sdc 27.25 3455.30 5.35 173.64 261.65 29035.48 130.82 14517.74 163.68 0.72 4.00 1.30 23.35
sdc1 27.25 3455.30 5.35 173.64 261.65 29035.48 130.82 14517.74 163.68 0.72 4.00 1.30 23.35
sdd 33.20 3475.52 5.75 151.73 306.05 29027.55 153.02 14513.78 186.27 8.42 51.30 6.01 94.63
sdd1 33.20 3475.52 5.75 151.73 306.05 29027.55 153.02 14513.78 186.27 8.42 51.30 6.01 94.63
sde 24.78 3447.87 5.45 175.42 241.82 28988.70 120.91 14494.35 161.61 0.62 3.46 1.23 22.28
sde1 24.78 3447.87 5.45 175.42 241.82 28988.70 120.91 14494.35 161.61 0.62 3.46 1.23 22.28
md0 0.00 0.00 0.00 3456.59 0.00 87011.30 0.00 43505.65 25.17 0.00 0.00 0.00 0.00
dm-0 0.00 0.00 0.00 3456.59 0.00 87011.30 0.00 43505.65 25.17 217.37 60.15 0.29 99.07the WD10EARS is my sdd disk.....is there an issue on that? it seems to be to much busy. (94.63%)
dm-0 hight % of usage is an issue caused by sdd disk??
thanks
Offline
#2 2010-03-11 20:04:54
- s.borrelli
- Member
Re: Some performance issue
another interesting info:
Code:
[root@openfiler Salva]# hdparm -i /dev/sdd
/dev/sdd:
Model=WDC WD10EARS-00Y5B1 , FwRev=80.00A80, SerialNo= WD-WCAV55903990
Config={ HardSect NotMFM HdSw>15uSec SpinMotCtl Fixed DTR>5Mbs FmtGapReq }
RawCHS=16383/16/63, TrkSize=0, SectSize=0, ECCbytes=50
BuffType=unknown, BuffSize=0kB, MaxMultSect=16, MultSect=?16?
CurCHS=16383/16/63, CurSects=16514064, LBA=yes, LBAsects=268435455
IORDY=on/off, tPIO={min:120,w/IORDY:120}, tDMA={min:120,rec:120}
PIO modes: pio0 pio3 pio4
DMA modes: mdma0 mdma1 mdma2
UDMA modes: udma0 udma1 udma2
AdvancedPM=no WriteCache=enabled
Drive conforms to: device does not report version:
* signifies the current active modeBuffSize=0kB???? Zero kB??? hei....my WD10EARS have a very very big cache :-(
the same command on another (WD10EADS) disk:
Code:
[root@openfiler Salva]# hdparm -i /dev/sdc
/dev/sdc:
Model=WDC WD10EADS-65M2B0 , FwRev=01.00A01, SerialNo= WD-WCAV54059470
Config={ HardSect NotMFM HdSw>15uSec SpinMotCtl Fixed DTR>5Mbs FmtGapReq }
RawCHS=16383/16/63, TrkSize=0, SectSize=0, ECCbytes=50
BuffType=unknown, BuffSize=32767kB, MaxMultSect=16, MultSect=?16?
CurCHS=16383/16/63, CurSects=16514064, LBA=yes, LBAsects=268435455
IORDY=on/off, tPIO={min:120,w/IORDY:120}, tDMA={min:120,rec:120}
PIO modes: pio0 pio3 pio4
DMA modes: mdma0 mdma1 mdma2
UDMA modes: udma0 udma1 udma2
AdvancedPM=no WriteCache=enabled
Drive conforms to: device does not report version:
* signifies the current active modereport the right cache size...
Offline
#4 2010-03-11 20:42:39
- s.borrelli
- Member
Re: Some performance issue
andersdr wrote:
try dd test for seq read/write and bonnie++
can you give me the exact command with parameters that i will use?
Offline
#5 2010-03-16 15:38:13
- s.borrelli
- Member
Re: Some performance issue
ok, WD10EARS replaced by a WD10EADS....this is the new iozone output
Code:
random random bkwd record stride
KB reclen write rewrite read reread read write read rewrite read fwrite frewrite fread freread
262144 64 130664 160939 1235141 1139384
262144 128 131666 187192 1258963 1158432
262144 256 129244 179202 1194936 1075900
262144 512 66697 74700 981510 925650
262144 1024 138350 143161 814129 794671
262144 2048 114263 146780 808821 769244
262144 4096 149344 91380 808046 770379
262144 8192 65776 112025 808914 753400
262144 16384 140792 125435 812131 774135
524288 64 90476 81295 1231220 1126209
524288 128 78084 66599 1258441 1149184
524288 256 66024 76176 1208657 1118757
524288 512 56267 87072 984476 924002
524288 1024 58561 67888 837027 795227
524288 2048 69316 74221 809286 770839
524288 4096 80211 64919 798582 767537
524288 8192 48796 90058 809793 765779
524288 16384 61135 85583 810002 771492
1048576 64 64261 75487 1266861 1165419
1048576 128 58153 75081 1294549 1178445
1048576 256 56461 69420 1232794 1133439
1048576 512 50330 59487 1002607 948004
1048576 1024 56115 70466 854146 813828
1048576 2048 48969 65379 815789 783416
1048576 4096 44137 69221 820979 782621
1048576 8192 58679 61086 822184 783421
1048576 16384 63750 63579 826416 784198
2097152 64 59042 76924 279468 290656
2097152 128 57311 71583 286764 295285
2097152 256 57336 76460 259821 282065
2097152 512 51648 71829 290003 299303
2097152 1024 51747 63522 252335 285126
2097152 2048 54126 67803 240897 290148
2097152 4096 42157 59271 258860 298107
2097152 8192 47628 66395 286172 292099
2097152 16384 53281 70289 274011 295006
4194304 64 47438 60868 280975 276182
4194304 128 52459 71069 286219 284707
4194304 256 47639 68217 288364 285744
4194304 512 48401 62828 270048 277593
4194304 1024 45559 63681 265595 295021
4194304 2048 40202 58757 266909 283336
4194304 4096 48004 59404 287681 284000
4194304 8192 42313 59988 257993 289445
4194304 16384 50810 67232 273385 286893and this is the new iostat output
Code:
Device: rrqm/s wrqm/s r/s w/s rsec/s wsec/s rkB/s wkB/s avgrq-sz avgqu-sz await svctm %util sda 0.00 1.27 0.00 5.29 0.00 52.50 0.00 26.25 9.93 0.00 0.81 0.22 0.12 sda1 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 sda2 0.00 0.39 0.00 1.27 0.00 13.32 0.00 6.66 10.46 0.00 0.00 0.00 0.00 sda3 0.00 0.88 0.00 4.02 0.00 39.18 0.00 19.59 9.76 0.00 1.07 0.29 0.12 sdb 81.88 6124.98 65.13 550.05 1152.60 53598.43 576.30 26799.22 89.00 5.36 8.49 0.91 55.91 sdb1 81.88 6124.98 65.13 550.05 1152.60 53598.43 576.30 26799.22 89.00 5.36 8.49 0.91 55.91 sdc 84.52 6125.95 68.07 555.04 1234.87 53466.80 617.43 26733.40 87.79 3.94 6.22 0.89 55.44 sdc1 84.52 6125.95 68.07 555.04 1234.87 53466.80 617.43 26733.40 87.79 3.94 6.22 0.89 55.44 sdd 87.37 6123.31 66.70 562.68 1239.57 53508.33 619.78 26754.16 86.99 4.65 7.39 0.90 56.65 sdd1 87.37 6123.31 66.70 562.68 1239.57 53508.33 619.78 26754.16 86.99 4.65 7.39 0.90 56.65 sde 92.36 6092.26 70.03 578.75 1303.04 53390.01 651.52 26695.00 84.30 5.05 7.61 0.92 59.78 sde1 92.36 6092.26 70.03 578.75 1303.04 53390.01 651.52 26695.00 84.30 5.05 7.61 0.92 59.78 md0 0.00 0.00 0.00 4267.19 0.00 159399.02 0.00 79699.51 37.35 0.00 0.00 0.00 0.00 dm-0 0.00 0.00 0.00 4267.19 0.00 159399.02 0.00 79699.51 37.35 107.26 24.94 0.19 82.12
as you can see % disk utili is moore balanced
now i need to study why WD10EARS is very very bad with openfiler.
bye
Offline