[Gluster-users] 1client+2server performance problem
eagleeyes
eagleeyes at 126.com
Tue Feb 17 06:14:21 UTC 2009
Hello everybody:
I have some problem with performance of 1client+2server application.
The system is AS5.2, 2.6.18-92.el5PAE kernel.
GFS is 1.3.9 built on Feb 11 2009 16:07:40 ,Repository revision: glusterfs--mainline--2.5--patch-770
FUSE is fuse-2.7.3glfs10
======================================================
GFS server has 4G memory , no raid ,PC server
172.20.22.0
/dev/sda10 665G 198M 631G 1% /opt
/dev/sda8 24G 373M 23G 2% /var
/dev/sda7 26G 2.6G 22G 11% /usr
/dev/sda9 116G 191M 109G 1% /usr/local
/dev/sda6 22G 173M 21G 1% /tmp
/dev/sda5 28G 173M 26G 1% /home
/dev/sda1 114M 13M 96M 12% /boot
/dev/sdb1 917G 1.8G 869G 1% /disk1
/dev/sdc1 917G 1.7G 869G 1% /disk2
tmpfs 2.0G 0 2.0G 0% /dev/shm
172.20.22.9
/dev/hda9 510G 199M 483G 1% /opt
/dev/hda5 45G 383M 42G 1% /var
/dev/hda3 25G 2.5G 21G 11% /usr
/dev/hda6 282G 194M 268G 1% /usr/local
/dev/hda2 20G 173M 19G 1% /home
/dev/hda1 99M 13M 81M 14% /boot
tmpfs 2.0G 0 2.0G 0% /dev/shm
/dev/hdc1 917G 1.8G 869G 1% /disk1
/dev/hdd1 917G 1.7G 869G 1% /disk2
======================================================
GFS client has 4G memory , virtual terminal
Filesystem Size Used Avail Use% Mounted on
/dev/sda3 141G 5.6G 128G 5% /
/dev/sda1 190M 12M 169M 7% /boot
tmpfs 948M 0 948M 0% /dev/shm
glusterfs 1.8T 3.4G 1.7T 1% /data1
======================================================
server configure
volume disk1
type storage/posix # POSIX FS translator
option directory /disk1/data # Export this directory
end-volume
volume disk2
type storage/posix # POSIX FS translator
option directory /disk2/data # Export this directory
end-volume
volume ns
type storage/posix # POSIX FS translator
option directory /disk1/export-ns # Export this directory
end-volume
volume iot1
type performance/io-threads
subvolumes disk1
option thread-count 16 # default value is 1
option cache-size 4MB # default is 64MB (This is per thread, so configure it according to your RAM size and thread-count.
end-volume
volume iot2
type performance/io-threads
subvolumes disk2
option thread-count 16 # default value is 1
option cache-size 4MB # default is 64MB (This is per thread, so configure it according to your RAM size and thread-count.
end-volume
volume server
type protocol/server
option transport-type tcp/server # For TCP/IP transport
subvolumes disk1 disk2 iot1 iot2 ns
option auth.ip.disk1.allow * # Allow access to "disk1" volume
option auth.ip.disk2.allow * # Allow access to "disk2" volume
option auth.ip.ns.allow * # Allow access to "ns" volume
option auth.ip.iot1.allow * # Allow access to "disk2" volume
option auth.ip.iot2.allow * # Allow access to "disk2" volume
end-volume
=====================================================
client configure
volume client1
type protocol/client
option transport-type tcp/client # for TCP/IP transport
option remote-host 172.20.22.0 # IP address of the remote brick
option remote-subvolume iot1 # name of the remote volume
option transport-timeout 20
end-volume
volume client2
type protocol/client
option transport-type tcp/client # for TCP/IP transport
option remote-host 172.20.22.0 # IP address of the remote brick
option remote-subvolume iot2
option transport-timeout 20
end-volume
volume client3
type protocol/client
option transport-type tcp/client # for TCP/IP transport
option remote-host 172.20.22.9 # IP address of the remote brick
option remote-subvolume iot1
option transport-timeout 20
end-volume
volume client4
type protocol/client
option transport-type tcp/client # for TCP/IP transport
option remote-host 172.20.22.9 # IP address of the remote brick
option remote-subvolume iot2
option transport-timeout 20
end-volume
volume ns1
type protocol/client
option transport-type tcp/client # for TCP/IP transport
option remote-host 172.20.22.0 # IP address of the remote brick
option remote-subvolume ns
option transport-timeout 20
end-volume
volume ns2
type protocol/client
option transport-type tcp/client # for TCP/IP transport
option remote-host 172.20.22.9 # IP address of the remote brick
option remote-subvolume ns
option transport-timeout 20
end-volume
volume afr1
type cluster/afr
subvolumes client1 client3
end-volume
volume afr2
type cluster/afr
subvolumes client2 client4
end-volume
volume afr-ns
type cluster/afr
subvolumes ns1 ns2
end-volume
volume unify
type cluster/unify
option namespace afr-ns
option scheduler rr
subvolumes afr1 afr2
end-volume
volume iot
type performance/io-threads
option thread-count 32
option cache-size 4MB
subvolumes unify
end-volume
=======================================================
When i writed some data into gluster directory /data1 by using dozen dd commands ,
i use "time ll -h " in /data1 with result :
real 0m8.388s
user 0m0.000s
sys 0m0.000s
How could i improve my performance ? what should i do ? waiting for your return ,thanks a lot
Could some tell me what your GFS configure?
2009-02-17
eagleeyes
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://supercolony.gluster.org/pipermail/gluster-users/attachments/20090217/b08c2f39/attachment.html>
More information about the Gluster-users
mailing list