gluster 文件系统

molecar

lustre是可以动态加节点的

nntp

原帖由 molecar 于 2008-1-16 17:14 发表
lustre是可以动态加节点的

好久不见你了....

troyme

How do I add a new node to an already running cluster of GlusterFS
Yes, you can add more bricks in your volume specification file and restart GlusterFS (re-mount). Its schedulers (alu) are designed to balance the file system data as you grow.

For releases after 1.3.0-pre5

Just add the extra node in unify's subvolumes list, and restart the GlusterFS, the directory structure is automatically replicated in the new server The much desired self-heal property of unify solves the burden of manually maintaining equal directory structure in all the servers before mount.

from http://www.gluster.org/docs/inde ... luster_of_GlusterFS

这个不知道算不算动态加，看起来好像可以在线加节点
最后几天正在测试，测试完放假，哈哈哈哈

[ 本帖最后由 troyme 于 2008-2-2 10:46 编辑 ]

yuhuohu

关注一下

nntp

"and restart the GlusterFS"

molecar

原帖由 nntp 于 2008-1-16 19:29 发表



好久不见你了....

哈哈，老大用什么即时通工具啊？我天天都在关注，发贴少而已！我可是非常崇拜你的。哈哈

[ 本帖最后由 molecar 于 2008-2-4 11:39 编辑 ]

nntp

原帖由 molecar 于 2008-2-4 11:33 发表

哈哈，老大用什么即时通工具啊？我天天都在关注，发贴少而已！我可是非常崇拜你的。哈哈

我只能用Pidgin.里面加的是msn.

troyme

原帖由 nntp 于 2008-2-3 04:26 发表
"and restart the GlusterFS"

一个remount过程，客户端的一个remount过程，server端没有变动，基本上和在线动态没有很大的应用层的区别吧，虽然底层是不一样的

root_man

原帖由 troyme 于 2008-2-2 10:44 发表
最后几天正在测试，测试完放假，哈哈哈哈

关注测试结果

1> client端
wget http://dev.centos.org/centos/4.4 ... .6.9-42.EL.i686.rpm
wget http://linux.web.psi.ch/dist/sci ... 2.5.3-1.SL.i686.rpm
http://ftp.physics.auth.gr/pub/m ... 2.5.3-1.SL.i686.rpm
rpm -ivh kernel-devel-2.6.9-42.EL.i686.rpm
cp /boot/config-2.6.9-42.EL /usr/src/kernel/2.6.9-42/.config
cd /usr/src/kernel/2.6.9-42/
make prepare
(注:一定要找对与自己内核匹配的kernel-devel,否则后面会出现一系列问题.如:fuse mount failed: invalid argument.)
yum install bison flex

tar -zxvf fuse-VERSION.tar.gz
cd fuse-VERSION
./configure --perfix=/usr/local/ --enable-kernel-module --with-kernel=/usr/src/kernels/2.6.9-42.EL-i686/
make
make install

tar -xvzf glusterfs-VERSION.tar.gz
cd gluster-VERSION
export CFLAGS="-O3"
./configure --sysconfdir=/etc --disable-server
make
make install

配置 /etc/glusterfs/glusterfs-client.vol
### File: /etc/glusterfs-client.vol - GlusterFS Client Volume Specification
volume client
type protocol/client
option transport-type tcp/client   
option remote-host 192.168.1.*     
option remote-port 6996            
option transport-timeout 1                                            
option remote-subvolume brick   
end-volume
volume readahead
type performance/read-ahead
option page-size 32MB
option page-count 16
subvolumes client
end-volume
volume iocache
type performance/io-cache
# option cache-size 64MB
option page-size 32MB
option page-count 16
subvolumes readahead
end-volume
volume writeback
type performance/write-behind
option aggregate-size 512MB
subvolumes iocache
end-volume
volume booster
type performance/booster
# option transport-type tcp
subvolumes writeback
end-volume
volume iothreads
type performance/io-threads
option thread-count 16
option cache-size 512MB
subvolumes booster
end-volume
并拷贝该文件到server端的/etc/gluster目录下

启动
glusterfs -LDEBUG -f /etc/glusterfs/glusterfs-client.vol /mnt/test -l /var/log/glusterfs.log

2> server端
yum install bison flex
tar xvzf glusterfs-VERSION.tar.gz
cd gluster-VERSION
export CFLAGS="-O3"
./configure --sysconfdir=/etc --disable-fuse-client
make
make install

配置 /etc/glusterfs/glusterfs-server.vol
### Export volume "brick" with the contents of "/home/export" directory.
volume brick
type storage/posix                   # POSIX FS translator
option directory /tmp/export       # Export this directory
end-volume
volume server
type protocol/server
option transport-type tcp/server     # For TCP/IP transport
option bind-address 192.168.1.*     # Default is to listen on all interfaces
option listen-port 6996              # Default is 6996
# option client-volume-filename /etc/glusterfs/glusterfs-client.vol
subvolumes brick
option auth.ip.brick.allow 192.168.* # Allow access to "brick" volume
end-volume

启动
glusterfsd -LDEBUG -l /var/log/glusterfs.log -f /etc/glusterfs/glusterfs-server.vol


单server 单client , 我测试结果glusterfs读写性能是本地读写性能的1/10 - 1/7
不尽人意

[ 本帖最后由 root_man 于 2008-2-27 12:27 编辑 ]

root_man