Killer SCSI 1 TB fileserver (Part 2)
Many of your questions may have already been answered in earlier discussions or in the FAQ. The search results page will indicate current discussions as well as past list serves, articles, and papers.
Bill Broadley bill at math.ucdavis.eduMon Jun 3 16:30:49 PDT 2002
- Previous message: MPICH memory growth problem.
- Next message: Power controlers... on another note.
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
A couple of you asked me to keep you up to date with my fileserver
experience.
I built a new fileserver:
6 GB ram
Dual p4 xeon (with hyperthreading, yes linux sees 4 cpu's)
Intel e7500 chipset based board
AIC-7899A * 2 (U160) channels on the motherboard
1 TB disk (3 channels of U160, 5 73 GB disk per channel).
I managed to almost TRIPLE my read performance of the old config, 300 MB/sec
sustained for over 12 GB! I don't think the cache helped much if at
all considering I never saw the buffers/cache exceed 150 MB or so.
5 disk raid-0:
2xP4,12G,,,88470,88,44419,34,,,144447,38,746.6,2,16,1863,98,+++++,+++,+++++,+++,1936,99,+++++,+++,5191,99
Version @version@ ------Sequential Output------ --Sequential Input- --Random-
-Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP
2xP4 12G 88470 88 44419 34 144447 38 746.6 2 ------Sequential Create------ --------Random Create--------
-Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
files:max:min /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
2xP4 16 1863 98 +++++ +++ +++++ +++ 1936 99 +++++ +++ 5191 99
10 disk raid-0:
2xP4,12G,,,96023,99,81299,73,,,311175,89,944.8,4,16,1883,99,+++++,+++,32624,99,1916,99,+++++,+++,5088,100
total used free shared buffers cached
Mem: 6199836 231724 5968112 0 30996 40908
-/+ buffers/cache: 159820 6040016
Swap: 6291408 0 6291408
Version @version@ ------Sequential Output------ --Sequential Input- --Random-
-Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP
2xP4 12G 96023 99 81299 73 311175 89 944.8 4 ------Sequential Create------ --------Random Create--------
-Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
files:max:min /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
2xP4 16 1883 99 +++++ +++ 32624 99 1916 99 +++++ +++ 5088 100
Immediately after the run:
total used free shared buffers cached
Mem: 6199836 204124 5995712 0 29240 35904
-/+ buffers/cache: 138980 6060856
Swap: 6291408 0 6291408
Any suggestions on how to improve writespeed? Even with double the number
of disks I only saw minimal improvement (86.4 MB/sec to 93.77 MB/sec).
Oh the partition was formated as Ext3. Has anyone else managed significantly
higher write bandwidth with hardware or software raid?
--
Bill Broadley
Mathematics/Institute of Theoretical Dynamics
UC Davis
- Previous message: MPICH memory growth problem.
- Next message: Power controlers... on another note.
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the Beowulf mailing list
