[toc]
sge高性能集群的搭建与使用
集群环境的准备
Node1(master) |
CentOS7.4 |
iptables/selinux(off) |
IP:10.180.66.11 |
hostname:node1 |
ali yum源 |
Node2(slave) |
CentOS7.4 |
iptables/selinux(off) |
IP:10.180.66.12 |
hostname:node2 |
ali yum源 |
Node3(slave) |
CentOS7.4 |
iptables/selinux(off) |
IP:10.180.66.13 |
hostname:node3 |
ali yum源 |
Node4(slave) |
CentOS7.4 |
iptables/selinux(off) |
IP:10.180.66.14 |
hostname:node4 |
ali yum源 |
Node5(slave) |
CentOS7.4 |
iptables/selinux(off) |
IP:10.180.66.15 |
hostname:node5 |
ali yum源 |
master 节点安装
安装相关依赖包
新建sge管理员用户
| # groupadd -g 490 sgeadmin |
| # useradd -u 495 -g 490 -r -m -d /home/sgeadmin -s /bin/bash -c "SGE Admin" sgeadmin |
| # sed -i '/^%wheel/a\%sgeadmin ALL=(ALL) NOPASSWD: ALL' /etc/sudoers |
安装sge
sge 链接 密码:c7hy
qmaster 安装自动回答脚本,依赖软件包expect
, 所有节点都需要安装
| |
| |
| |
| user="sgeadmin" |
| /usr/bin/expect <<-EOF |
| spawn ./install_qmaster |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send "y |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send "n |
| " |
| |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect eof |
| EOF |
| |
修改主节点环境变量
添加节点
master 服务器搭建 nfs 服务
所有节点都需要安装
master节点操作
| |
| /data/sge 10.180.66.0/24(rw,sync) |
| |
slave 节点挂载
(node2,node3,node4,node5)执行
slave 服务器安装sgeexecd
(node2,node3,node4,node5) 执行
生效环境变量
进行安装,所有节点都执行此脚本
| |
| |
| |
| user="sgeadmin" |
| /usr/bin/expect <<-EOF |
| spawn ./install_execd |
| |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect "*>>" |
| send " |
| " |
| expect eof |
| EOF |
| |
完成集群搭建
| |
| HOSTNAME ARCH NCPU LOAD MEMTOT MEMUSE SWAPTO SWAPUS |
| ------------------------------------------------------------------------------- |
| global - - - - - - - |
| node1 linux-x64 1 0.01 968.3M 193.0M 2.0G 64.0K |
| node2 linux-x64 1 0.01 976.3M 151.0M 2.0G 0.0 |
| node3 linux-x64 1 0.02 978.3M 152.2M 2.0G 84.0K |
| node4 linux-x64 1 0.02 976.3M 155.4M 2.0G 0.0 |
| node5 linux-x64 1 0.01 978.3M 148.5M 2.0G 84.0K |
sge集群的使用