kaldi NFS/GlusterFS

NFS

详细介绍参考《鸟哥的LINUX私房菜-服务器架设篇》第十三章
主要流程

server端配置

1. 安装两个软件：rpcbind和nfs-utils
2. 配置/etc/exports

<DIR> *(rw,no_root_squash)

1. 启动nfs

/etc/init.d/rpcbind start
/etc/init.d/nfs start
/etc/init.d/nfslock start
chkconfig rpcbind on
chkconfig nfs on
chkconfig nfslock on

1. 如果更新了/etc/exports文件，需要重新挂载

exportfs -arv

client端配置

1. 安装两个软件：rpcbind和nfs-utils
2. 启动相关服务

/etc/init.d/rpcbind start
/etc/init.d/nfslock start

1. 查看服务器提供了哪些共享目录

showmount -e hostname

1. 建立挂载点并且挂载

mkdir -p <DIR>
mount -t nfs hostname:<DIR> <DIR>
umount <DIR> #卸载

glusterfs

官方文档：http://gluster.readthedocs.io/en/latest/

Brick：对应一台服务器上面的存储空间
Volume：多个brick组成一个volume，对应于多台服务器组成的存储空间
Trusted Storage Pool：多台服务器组成的服务器集群

CENTOS安装

https://wiki.centos.org/SpecialInterestGroup/Storage/gluster-Quickstart
http://gluster.readthedocs.io/en/latest/Quick-Start-Guide/Quickstart/
安装到以下机器：hpc191、hpc192

#1.install on hpc191/hpc192
yum install centos-release-gluster
yum --enablerepo=centos-gluster*-test install glusterfs-server
service glusterd start
service glusterd status

#2.Configure the trusted pool
gluster peer probe hpc192 #on hpc191
gluster peer probe hpc191 #on hpc192

#3.Set up a GlusterFS volume
mkdir -p /euler/glusterfs/gv1 #on hpc191/hpc192
gluster volume create gv1 hpc191:<DIR>/gv1 hpc192:<DIR>/gv1 #on hpc191
gluster volume start gv1 #on hpc191
gluster volume info #on hpc191

#4.mount
mount -t glusterfs hpc191:/gv1 /glfs #on hpc191
mount -t glusterfs hpc192:/gv1 /glfs #on hpc192

扩展节点

新增加hpc193

#1.在client安装
yum install centos-release-gluster
yum --enablerepo=centos-gluster*-test install glusterfs-server
service glusterd start
service glusterd status

#2.Configure the trusted pool,在hpc191或者hpc192执行
gluster peer probe hpc193

#3.add brick to volume
gluster volume add-brick gv1 hpc193:<DIR>/gv1
gluster volume status

#4.mount 
mount -t glusterfs hpc193:/gv1 /glfs

#5.rebalance：add/delete brick 以后都需要rebalance
gluster volume rebalance gv1 fix-layout start #第一种方式：不迁移已有的数据
gluster volume rebalance gv1 start            #第二种方式：迁移已有的数据

卸载

#hpc191
gluster volume stop gv1 
gluster volume delete gv1
gluster peer detach hpc192
gluster peer detach hpc193

#hpc191/hpc192/hpc193
service glusferfsd stop

其他操作

service启动

/etc/init.d/glusterd start
/etc/init.d/glusterd stop
/etc/init.d/glusterd status

pool操作

#add server to pool
gluster peer probe server
#status of pool
gluster peer status
#remove from pool
gluster peer detach server

volume操作

创建设置

类型包括：
Distributed
Striped
Distributed Striped
Distributed Replicated
Distributed Striped Replicated
Striped Replicated
Dispersed
Distributed Dispersed

其他操作

操作包括：
Tuning Volume Options
Configuring Transport Types for a Volume
Expanding Volumes
Shrinking Volumes
Replacing Bricks
Migrating Volumes
Rebalancing Volumes
Stopping Volumes
Deleting Volumes
Triggering Self-Heal on Replicate
Non Uniform File Allocation(NUFA)

log路径