文档章节

Greenplum 数据库安装部署(生产环境)

闻术苑
 闻术苑
发布于 2017/05/25 07:10
字数 4020
阅读 556
收藏 2

硬件配置:
16 台 IBM X3650,
节点配置:CPU 2 * 8core,内存 128GB,硬盘 16 * 900GB,万兆网卡。
万兆交换机。

安装需求:
1台Master,1台Standby Master,14台Segment计算节点。

安装步骤:

  1. Master节点安装
  2. 创建GP安装配置文件并配置ssh互信
  3. 关闭防火墙及开启自启动
  4. 关闭SELinux
  5. 磁盘调度算法
  6. 磁盘预读取配置
  7. 语言与字符集
  8. Sysctl.conf增加配置
  9. 用户资源限额配置
  10. 时间校对
  11. 设定网卡自启动
  12. 创建用户(可选)
  13. 创建目录并赋权
  14. 各节点GP软件的安装
  15. 初始化GP数据库

1. Master节点安装 

先确认各节点目录结构保持一致:

[root@XXXGPM01 db]# df -h
Filesystem      Size  Used Avail Use% Mounted on
/dev/sda6        20G  440M   18G   3% /
tmpfs            63G   76K   63G   1% /dev/shm
/dev/sda2       485M   39M  421M   9% /boot
/dev/sda1       200M  260K  200M   1% /boot/efi
/dev/sda4       5.5T  3.1G  5.5T   1% /data1
/dev/sdb1       5.8T   34M  5.8T   1% /data2
/dev/sda8       9.7G  150M  9.0G   2% /home
/dev/sda5        49G  3.1G   43G   7% /opt
/dev/sda9       9.7G  151M  9.0G   2% /tmp
/dev/sda7        20G  5.8G   13G  32% /usr
/dev/sda10      9.7G  338M  8.8G   4% /var

主节点/etc/hosts文件,追加内容(万兆网卡的IP地址和主机名):

172.16.99.18 XXXGPM01
172.16.99.19 XXXGPM02
172.16.99.20 XXXGPD01
172.16.99.21 XXXGPD02
172.16.99.22 XXXGPD03
172.16.99.23 XXXGPD04
172.16.99.24 XXXGPD05
172.16.99.25 XXXGPD06
172.16.99.26 XXXGPD07
172.16.99.27 XXXGPD08
172.16.99.28 XXXGPD09
172.16.99.29 XXXGPD10
172.16.99.30 XXXGPD11
172.16.99.31 XXXGPD12
172.16.99.32 XXXGPD13
172.16.99.33 XXXGPD14
  1. 解压安装介质
    GP的安装包:/opt/db.zip
    cd /data1; unzip /opt/db.zip
  2. 安装目录
    mkdir -p /data1/gpinstall

在 XXXGPM1 节点上,将 greenplum-db-4.x.x.x-build-5-RHEL5-x86_64.zip 解开, 以 root 用户执行得到的.bin 文件。按照提示进行安装。

# /bin/bash greenplum-db-4.3.2.0-build-1-RHEL5-x86_64.bin

选择自定义安装目录: /data1/gpinstall/greenplum-db-4.3.2.0

安装完成之后,可以看到如下目录

[root@XXXGPM01 gpinstall]# pwd
/data1/gpinstall
[root@XXXGPM01 gpinstall]# ls -lh
总用量 4.0K
lrwxrwxrwx.  1 root root   22 6月   4 18:51 greenplum-db -> ./greenplum-db-4.3.2.0
drwxr-xr-x. 11 root root 4.0K 6月   4 18:51 greenplum-db-4.3.2.0

2. 创建GP安装配置文件并配置ssh互信 

mkdir -p /data1/gpinstall/config
创建两个配置文件

1. allnodes.txt

XXXGPM01
XXXGPM02
XXXGPD01
XXXGPD02
XXXGPD03
XXXGPD04
XXXGPD05
XXXGPD06
XXXGPD07
XXXGPD08
XXXGPD09
XXXGPD10
XXXGPD11
XXXGPD12
XXXGPD13
XXXGPD14

2. nodes.txt

XXXGPD01
XXXGPD02
XXXGPD03
XXXGPD04
XXXGPD05
XXXGPD06
XXXGPD07
XXXGPD08
XXXGPD09
XXXGPD10
XXXGPD11
XXXGPD12
XXXGPD13
XXXGPD14

3. 配置所有GP节点root用户的ssh互信:

source /data1/gpinstall/greenplum-db/greenplum_path.sh 
gpssh-exkeys -f /data1/gpinstall/config/allnodes.txt 

测试可以正常互信。此时就可以方便的去同步配置所有的参数。

先写脚本同步各节点的/etc/hosts文件:

#!/bin/bash
#Usage: copy files to other hosts in cluster.
#ex: sh bulkcp.sh /etc/hosts /etc/hosts
#Author: AlfredZhao
#Version: 1.0.0
for((i=18;i<=33;i++))
do
 scp $1 172.16.99.$i:$2
 echo "scp $1 172.16.99.$i:$2"
done

gpssh -f /data1/gpinstall/config/allnodes.txt -e '' 回车即可进入交互性界面。

3. 关闭防火墙及开启自启动 

检查是否关闭及开机启动项设置:

service iptables status
service ip6tables status
service libvirtd status

service iptables stop
service ip6tables stop
service libvirtd stop

chkconfig libvirtd off
chkconfig iptables off
chkconfig ip6tables off

4. 关闭SELinux 

检查SELinux的当前状态和配置:

getenforce
more /etc/selinux/config | grep SELINUX=

临时关闭SELinux和永久禁用SELinux:

setenforce 0
sed -i 's/SELINUX=enforcing/SELINUX=disabled/g' /etc/selinux/config

5. 磁盘调度算法 

检查当前的磁盘调度算法:
cat /sys/block/sd*/queue/scheduler

原厂给的设置方法:
修改/boot/grub/menu.lst 找到 kernel /vmlinuz-xxx 这一行,在最后添加 elevator=deadline

检查/验证方法:
系统启动正常后,执行 cat /sys/block/*/queue/scheduler 应能看到: noop anticipatory [deadline] cfq

实际遇到问题:RHEL6.5系统 没有/boot/grub/menu.lst 这个文件。
nl /boot/grub/menu.lst
换一个思路,利用rc.local配置加两行信息如下:

echo deadline > /sys/block/sda/queue/scheduler
echo deadline > /sys/block/sdb/queue/scheduler

这里为了快速同步集群的配置统一,将这两行写入临时配置文件/tmp/gpconfig/1deadline.conf,然后同步到各节点。
mkdir -p /tmp/gpconfig
最后在交互性界面执行命令:
cat /tmp/gpconfig/1deadline.conf >> /etc/rc.d/rc.local

6. 磁盘预读取配置 

检查磁盘预读取设置,应都是 16384:
blockdev --getra /dev/sd* blockdev --getra /dev/dm-*

设置办法:
修改/etc/rc.d/rc.local 增加两行

blockdev --setra 16384 /dev/sd* 
blockdev --setra 16384 /dev/dm-* 

检查办法:
系统重启后,运行
blockdev --getra /dev/sd* blockdev --getra /dev/dm-* 应都是 16384

/etc/rc.d/rc.local
sh bulkcp.sh /tmp/gpconfig/2rclocal.conf /tmp/gpconfig/2rclocal.conf
gpssh -f /data1/gpinstall/config/allnodes.txt -e ''
cat /tmp/gpconfig/2rclocal.conf >> /etc/rc.d/rc.local

7. 语言与字符集 

检查语言和字符集:echo $LANG
en_US.UTF-8
设置办法:系统安装时指定。
检查办法:登录系统,执行命令 locale 结果应该是 en_US.UTF-8

修改:
sed -i 's/zh_CN.UTF-8/en_US.UTF-8/g' /etc/sysconfig/i18n

8. Sysctl.conf增加配置 

配置sysctl.conf
sysctl -p
有时需要 modprobe bridge

设置办法: 修改/etc/sysctl.conf,
增加以下内容:

net.ipv4.ip_forward = 0  
net.ipv4.conf.default.accept_source_route = 0  
kernel.sysrq = 0  
kernel.core_uses_pid = 1  
net.ipv4.tcp_syncookies = 1  
kernel.msgmnb = 65536  
kernel.msgmax = 65536  
kernel.sem = 250 64000 100 2048  
kernel.shmmax = 5000000000  
kernel.shmmni = 40960  
kernel.shmall = 40000000000  
net.ipv4.tcp_tw_recycle=1  
net.ipv4.tcp_max_syn_backlog=4096  
net.core.netdev_max_backlog=10000  
vm.overcommit_memory=2  
net.ipv4.conf.all.arp_filter = 1  
net.ipv4.ip_local_port_range = 1025 65535 

然后执行 sysctl -p。

vi /tmp/gpconfig/4sysctl.conf
sh bulckcp.sh /tmp/gpconfig/4sysctl.conf /tmp/gpconfig/4sysctl.conf
gpssh -f /data1/gpinstall/config/allnodes.txt -e ''
cat /tmp/gpconfig/4sysctl.conf >> /etc/sysctl.conf

检查办法: 用 sysctl ,检查上述各参数是否匹配。

9. 用户资源限额配置 

查看用户资源限额:
ulimit -a

设置办法:
修改 /etc/security/limits.d/90-nproc.conf
Replace with:

* soft nofile 1048576 
* hard nofile 1048576 
* soft nproc 1048576 
* hard nproc 1048576 

同步修改:

vi /tmp/gpconfig/5limits.conf
sh bulkcp.sh /tmp/gpconfig/5limits.conf /tmp/gpconfig/5limits.conf
nl /etc/security/limits.d/90-nproc.conf
cat /tmp/gpconfig/5limits.conf >> /etc/security/limits.d/90-nproc.conf

检查办法: 用任意普通用户登录,执行 ulimit -a,检查上述参数是否匹配。

10. 时间校对 

检查时间:
date

校对时间:
--2015年6月5日9:17
date 060509172015
hwclock -w

11. 设定网卡自启动 

确定需要启用的网卡:eth2
ifconfig eth2 | grep 172
more /etc/sysconfig/network-scripts/ifcfg-eth2 |grep ONBOOT

修改eth2网卡配置,开启自启动
sed -i 's/ONBOOT=no/ONBOOT=yes/g' /etc/sysconfig/network-scripts/ifcfg-eth2

以上步骤修改完毕后统一重启下机器,再次验证以上各个步骤是否修改成功。

12. 创建用户(可选) 

这个也可以在gpseginstall时指定选项创建,这里为了规范业务用户的uid和gid事先先创建好组和用户:
写一个小脚本,分发到各节点,然后在交互式界面下统一执行脚本建立组,用户,密码:

#!/bin/bash
#Usage: create gpadmin
#Author: AlfredZhao
#version: 1.0.0

groupadd gpadmin -g 3030
useradd gpadmin -u 3030 -g 3030
passwd gpadmin <<EOF
> gpadminpwd
> gpadminpwd
> EOF

13. 创建目录并赋权 

Master 目录:

gpssh -h XXXGPM01 -e 'mkdir -p /data1/master' 
gpssh -h XXXGPM02 -e 'mkdir -p /data1/master'

gpssh -h XXXGPM01 -e 'chown gpadmin:gpadmin /data1/master' 
gpssh -h XXXGPM02 -e 'chown gpadmin:gpadmin /data1/master'

数据库数据文件存储目录:
注意这里配置文件变为nodes.txt,主节点并不需要创建这些目录。

gpssh -f /data1/gpinstall/config/nodes.txt -e ''

mkdir -p /data1/primary 
mkdir -p /data1/mirror     
mkdir -p /data2/primary 
mkdir -p /data2/mirror

chown gpadmin:gpadmin /data1/primary 
chown gpadmin:gpadmin /data1/mirror     
chown gpadmin:gpadmin /data2/primary 
chown gpadmin:gpadmin /data2/mirror

14. 各节点GP软件的安装 

gpseginstall -f allnodes.txt -c csv
如果前面没有手工提前建立gpadmin用户,这里可以指定用户密码(-u -p),统一建立gpadmin用户和密码

[root@XXXGPM01 data1]# gpseginstall -f /data1/gpinstall/config/allnodes.txt -c csv
20150605:00:01:07:005656 gpseginstall:XXXGPM01:root-[INFO]:-Installation Info:
link_name greenplum-db
binary_path /data1/gpinstall/greenplum-db-4.3.2.0
binary_dir_location /data1/gpinstall
binary_dir_name greenplum-db-4.3.2.0
20150605:00:01:07:005656 gpseginstall:XXXGPM01:root-[INFO]:-check cluster password access
20150605:00:01:11:005656 gpseginstall:XXXGPM01:root-[INFO]:-de-duplicate hostnames
20150605:00:01:11:005656 gpseginstall:XXXGPM01:root-[INFO]:-master hostname: XXXGPM01
20150605:00:01:13:005656 gpseginstall:XXXGPM01:root-[INFO]:-chown -R gpadmin:gpadmin /data1/gpinstall/greenplum-db
20150605:00:01:13:005656 gpseginstall:XXXGPM01:root-[INFO]:-chown -R gpadmin:gpadmin /data1/gpinstall/greenplum-db-4.3.2.0
20150605:00:01:14:005656 gpseginstall:XXXGPM01:root-[INFO]:-rm -f /data1/gpinstall/greenplum-db-4.3.2.0.tar; rm -f /data1/gpinstall/greenplum-db-4.3.2.0.tar.gz
20150605:00:01:14:005656 gpseginstall:XXXGPM01:root-[INFO]:-cd /data1/gpinstall; tar cf greenplum-db-4.3.2.0.tar greenplum-db-4.3.2.0
20150605:00:01:17:005656 gpseginstall:XXXGPM01:root-[INFO]:-gzip /data1/gpinstall/greenplum-db-4.3.2.0.tar
20150605:00:01:36:005656 gpseginstall:XXXGPM01:root-[INFO]:-remote command: mkdir -p /data1/gpinstall
20150605:00:01:37:005656 gpseginstall:XXXGPM01:root-[INFO]:-remote command: rm -rf /data1/gpinstall/greenplum-db-4.3.2.0
20150605:00:01:39:005656 gpseginstall:XXXGPM01:root-[INFO]:-scp software to remote location
20150605:00:01:41:005656 gpseginstall:XXXGPM01:root-[INFO]:-remote command: gzip -f -d /data1/gpinstall/greenplum-db-4.3.2.0.tar.gz
20150605:00:01:46:005656 gpseginstall:XXXGPM01:root-[INFO]:-md5 check on remote location
20150605:00:01:48:005656 gpseginstall:XXXGPM01:root-[INFO]:-remote command: cd /data1/gpinstall; tar xf greenplum-db-4.3.2.0.tar
20150605:00:01:51:005656 gpseginstall:XXXGPM01:root-[INFO]:-remote command: rm -f /data1/gpinstall/greenplum-db-4.3.2.0.tar
20150605:00:01:52:005656 gpseginstall:XXXGPM01:root-[INFO]:-remote command: cd /data1/gpinstall; rm -f greenplum-db; ln -fs greenplum-db-4.3.2.0 greenplum-db
20150605:00:01:54:005656 gpseginstall:XXXGPM01:root-[INFO]:-remote command: chown -R gpadmin:gpadmin /data1/gpinstall/greenplum-db
20150605:00:01:55:005656 gpseginstall:XXXGPM01:root-[INFO]:-remote command: chown -R gpadmin:gpadmin /data1/gpinstall/greenplum-db-4.3.2.0
20150605:00:01:57:005656 gpseginstall:XXXGPM01:root-[INFO]:-rm -f /data1/gpinstall/greenplum-db-4.3.2.0.tar.gz
20150605:00:01:57:005656 gpseginstall:XXXGPM01:root-[INFO]:-version string on master: gpssh version 4.3.2.0 build 1
20150605:00:01:57:005656 gpseginstall:XXXGPM01:root-[INFO]:-remote command: . /data1/gpinstall/greenplum-db/./greenplum_path.sh; /data1/gpinstall/greenplum-db/./bin/gpssh --version
20150605:00:01:59:005656 gpseginstall:XXXGPM01:root-[INFO]:-remote command: . /data1/gpinstall/greenplum-db-4.3.2.0/greenplum_path.sh; /data1/gpinstall/greenplum-db-4.3.2.0/bin/gpssh --version
20150605:00:02:06:005656 gpseginstall:XXXGPM01:root-[INFO]:-SUCCESS -- Requested commands completed

15. 初始化GP数据库 

以下工作,在Master节点上以gpadmin用户登陆完成。
cd /data1/gpinstall/config
新建gpinitsystem_config配置文件,内容如下:

ARRAY_NAME="XXXGPDB"  
SEG_PREFIX=gpseg  
PORT_BASE=40000  
declare -a DATA_DIRECTORY=(/data1/primary /data1/primary /data1/primary /data1/primary /data1/primary /data1/primary /data2/primary /data2/primary /data2/primary /data2/primary /data2/primary /data2/primary)  
MASTER_HOSTNAME=XXXGPM01 
MASTER_DIRECTORY=/data1/master  
MASTER_PORT=5432  
TRUSTED_SHELL=ssh 
CHECK_POINT_SEGMENTS=256 
ENCODING=UNICODE 
MIRROR_PORT_BASE=50000  
REPLICATION_PORT_BASE=41000  
MIRROR_REPLICATION_PORT_BASE=51000  
declare -a MIRROR_DATA_DIRECTORY=(/data1/mirror /data1/mirror /data1/mirror /data1/mirror /data1/mirror /data1/mirror /data2/mirror /data2/mirror /data2/mirror /data2/mirror /data2/mirror /data2/mirror)   

15.1 配置下gpadmin用户的各节点间互信:

gpssh-exkeys -f /data1/gpinstall/config/allnodes.txt

15.2 初始化数据库:

gpinitsystem -c gpinitsystem_config -h nodes.txt -B 8

15.3 配置环境变量:

source /usr/local/greenplum-db/greenplum_path.sh
export MASTER_DATA_DIRECTORY=/data1/master/gpseg-1

15.4 建立冗余的master节点:

gpinitstandby -s XXXGPM02

15.5 调整数据库参数:

以下数据库参数调整后必须重新启动数据库。

调整方法:执行命令 gpconfig -c 参数名 -v 参数值 -m Master节点值

检查方法:重启数据库后,执行命令 gpconfig -s 参数名

参数名 参数值 master节点值

gpconfig -c shared_buffers -v 128MB -m 128MB
gpconfig -c gp_vmem_protect_limit -v 15360 -m 15360
gpconfig -c max_connections -v 1000 -m 200
gpconfig --skipvalidation -c wal_send_client_timeout -v 60s -m 60s 

[gpadmin@XXXGPM01 greenplum-db]$  gpconfig -c shared_buffers -v 128MB -m 128MB
20150605:14:50:53:017038 gpconfig:XXXGPM01:gpadmin-[INFO]:-completed successfully
[gpadmin@XXXGPM01 greenplum-db]$ gpconfig -c gp_vmem_protect_limit -v 15360 -m 15360
20150605:14:52:51:017179 gpconfig:XXXGPM01:gpadmin-[INFO]:-completed successfully
[gpadmin@XXXGPM01 greenplum-db]$ gpconfig -c max_connections -v 1000 -m 200
20150605:14:53:08:017271 gpconfig:XXXGPM01:gpadmin-[INFO]:-completed successfully
[gpadmin@XXXGPM01 greenplum-db]$ gpconfig --skipvalidation -c wal_send_client_timeout -v 60s -m 60s 
20150605:14:53:23:017363 gpconfig:XXXGPM01:gpadmin-[INFO]:-completed successfully

15.6 关闭数据库

[gpadmin@XXXGPM01 greenplum-db]$ gpstop -a
20150605:14:54:40:017533 gpstop:XXXGPM01:gpadmin-[INFO]:-Starting gpstop with args: -a
20150605:14:54:40:017533 gpstop:XXXGPM01:gpadmin-[INFO]:-Gathering information and validating the environment...
20150605:14:54:40:017533 gpstop:XXXGPM01:gpadmin-[INFO]:-Obtaining Greenplum Master catalog information
20150605:14:54:40:017533 gpstop:XXXGPM01:gpadmin-[INFO]:-Obtaining Segment details from master...
20150605:14:54:42:017533 gpstop:XXXGPM01:gpadmin-[INFO]:-Greenplum Version: 'postgres (Greenplum Database) 4.3.2.0 build 1'
20150605:14:54:42:017533 gpstop:XXXGPM01:gpadmin-[INFO]:-There are 0 connections to the database
20150605:14:54:42:017533 gpstop:XXXGPM01:gpadmin-[INFO]:-Commencing Master instance shutdown with mode='smart'
20150605:14:54:42:017533 gpstop:XXXGPM01:gpadmin-[INFO]:-Master host=XXXGPM01
20150605:14:54:42:017533 gpstop:XXXGPM01:gpadmin-[INFO]:-Commencing Master instance shutdown with mode=smart
20150605:14:54:42:017533 gpstop:XXXGPM01:gpadmin-[INFO]:-Master segment instance directory=/data1/master/gpseg-1
20150605:14:54:43:017533 gpstop:XXXGPM01:gpadmin-[INFO]:-Stopping master standby host XXXGPM02 mode=fast
20150605:14:54:45:017533 gpstop:XXXGPM01:gpadmin-[INFO]:-Successfully shutdown standby process on XXXGPM02
20150605:14:54:45:017533 gpstop:XXXGPM01:gpadmin-[INFO]:-Commencing parallel primary segment instance shutdown, please wait...
........................... 
20150605:14:55:12:017533 gpstop:XXXGPM01:gpadmin-[INFO]:-Commencing parallel mirror segment instance shutdown, please wait...
................ 
20150605:14:55:28:017533 gpstop:XXXGPM01:gpadmin-[INFO]:-----------------------------------------------------
20150605:14:55:28:017533 gpstop:XXXGPM01:gpadmin-[INFO]:-   Segments stopped successfully      = 336
20150605:14:55:28:017533 gpstop:XXXGPM01:gpadmin-[INFO]:-   Segments with errors during stop   = 0
20150605:14:55:28:017533 gpstop:XXXGPM01:gpadmin-[INFO]:-----------------------------------------------------
20150605:14:55:28:017533 gpstop:XXXGPM01:gpadmin-[INFO]:-Successfully shutdown 336 of 336 segment instances 
20150605:14:55:28:017533 gpstop:XXXGPM01:gpadmin-[INFO]:-Database successfully shutdown with no errors reported

15.7 启动数据库:

[gpadmin@XXXGPM01 greenplum-db]$ gpstart -a
20150605:14:56:39:017669 gpstart:XXXGPM01:gpadmin-[INFO]:-Starting gpstart with args: -a
20150605:14:56:39:017669 gpstart:XXXGPM01:gpadmin-[INFO]:-Gathering information and validating the environment...
20150605:14:56:39:017669 gpstart:XXXGPM01:gpadmin-[INFO]:-Greenplum Binary Version: 'postgres (Greenplum Database) 4.3.2.0 build 1'
20150605:14:56:39:017669 gpstart:XXXGPM01:gpadmin-[INFO]:-Greenplum Catalog Version: '201310150'
20150605:14:56:39:017669 gpstart:XXXGPM01:gpadmin-[INFO]:-Starting Master instance in admin mode
20150605:14:56:40:017669 gpstart:XXXGPM01:gpadmin-[INFO]:-Obtaining Greenplum Master catalog information
20150605:14:56:40:017669 gpstart:XXXGPM01:gpadmin-[INFO]:-Obtaining Segment details from master...
20150605:14:56:42:017669 gpstart:XXXGPM01:gpadmin-[INFO]:-Setting new master era
20150605:14:56:42:017669 gpstart:XXXGPM01:gpadmin-[INFO]:-Master Started...
20150605:14:56:42:017669 gpstart:XXXGPM01:gpadmin-[INFO]:-Shutting down master
20150605:14:56:46:017669 gpstart:XXXGPM01:gpadmin-[INFO]:-Commencing parallel primary and mirror segment instance startup, please wait...
.................. 
20150605:14:57:04:017669 gpstart:XXXGPM01:gpadmin-[INFO]:-Process results...
20150605:14:57:04:017669 gpstart:XXXGPM01:gpadmin-[INFO]:-----------------------------------------------------
20150605:14:57:04:017669 gpstart:XXXGPM01:gpadmin-[INFO]:-   Successful segment starts                                            = 336
20150605:14:57:04:017669 gpstart:XXXGPM01:gpadmin-[INFO]:-   Failed segment starts                                                = 0
20150605:14:57:04:017669 gpstart:XXXGPM01:gpadmin-[INFO]:-   Skipped segment starts (segments are marked down in configuration)   = 0
20150605:14:57:04:017669 gpstart:XXXGPM01:gpadmin-[INFO]:-----------------------------------------------------
20150605:14:57:04:017669 gpstart:XXXGPM01:gpadmin-[INFO]:-
20150605:14:57:04:017669 gpstart:XXXGPM01:gpadmin-[INFO]:-Successfully started 336 of 336 segment instances 
20150605:14:57:04:017669 gpstart:XXXGPM01:gpadmin-[INFO]:-----------------------------------------------------
20150605:14:57:04:017669 gpstart:XXXGPM01:gpadmin-[INFO]:-Starting Master instance XXXGPM01 directory /data1/master/gpseg-1 
20150605:14:57:05:017669 gpstart:XXXGPM01:gpadmin-[INFO]:-Command pg_ctl reports Master XXXGPM01 instance active
20150605:14:57:06:017669 gpstart:XXXGPM01:gpadmin-[INFO]:-Starting standby master
20150605:14:57:06:017669 gpstart:XXXGPM01:gpadmin-[INFO]:-Checking if standby master is running on host: XXXGPM02  in directory: /data1/master/gpseg-1
20150605:14:57:09:017669 gpstart:XXXGPM01:gpadmin-[INFO]:-Database successfully started
[gpadmin@XXXGPM01 greenplum-db]$ 

15.8 检查数据库的参数

gpconfig -s shared_buffers
gpconfig -s gp_vmem_protect_limit
gpconfig -s max_connections
gpconfig -s wal_send_client_timeout

[gpadmin@XXXGPM01 greenplum-db]$ gpconfig -s shared_buffers
20150605:11:37:40:011591 gpconfig:XXXGPM01:gpadmin-[ERROR]:-Failed to retrieve GUC information: error 'ERROR:  function gp_toolkit.gp_param_setting(unknown) does not exist
LINE 1: select * from gp_toolkit.gp_param_setting('shared_buffers')
                      ^
HINT:  No function matches the given name and argument types. You may need to add explicit type casts.
' in 'select * from gp_toolkit.gp_param_setting('shared_buffers')'

psql postgres

show shared_buffers;
show gp_vmem_protect_limit;
show max_connections;
show wal_send_client_timeout;

15.9 创建业务数据库XXX

项目内规范:所有后期创建数据库,都必须在psql 登陆到postgres数据库后创建!!

psql postgres
postgres=# create database XXX;
CREATE DATABASE

psql XXX

15.10 调整连接控制参数

修改文件 $MASTER_DATA_DIRECTORY/pg_hba.conf
增加一行:
host all all 0/0 md5
修改standby master上的文件 $MASTER_DATA_DIRECTORY/pg_hba.conf
增加一行:
host all all 0/0 md5

15.11 查看数据库的状态(gpstate)

[gpadmin@XXXGPM01 ~]$ gpstate
20150605:13:48:23:015288 gpstate:XXXGPM01:gpadmin-[INFO]:-Starting gpstate with args: 
20150605:13:48:24:015288 gpstate:XXXGPM01:gpadmin-[INFO]:-local Greenplum Version: 'postgres (Greenplum Database) 4.3.2.0 build 1'
20150605:13:48:24:015288 gpstate:XXXGPM01:gpadmin-[INFO]:-master Greenplum Version: 'PostgreSQL 8.2.15 (Greenplum Database 4.3.2.0 build 1) on x86_64-unknown-linux-gnu, compiled by GCC gcc (GCC) 4.4.2 compiled on Jul 12 2014 17:02:40'
20150605:13:48:24:015288 gpstate:XXXGPM01:gpadmin-[INFO]:-Obtaining Segment details from master...
20150605:13:48:25:015288 gpstate:XXXGPM01:gpadmin-[INFO]:-Gathering data from segments...
................ 
20150605:13:48:41:015288 gpstate:XXXGPM01:gpadmin-[INFO]:-Greenplum instance status summary
20150605:13:48:41:015288 gpstate:XXXGPM01:gpadmin-[INFO]:-----------------------------------------------------
20150605:13:48:41:015288 gpstate:XXXGPM01:gpadmin-[INFO]:-   Master instance                                           = Active
20150605:13:48:41:015288 gpstate:XXXGPM01:gpadmin-[INFO]:-   Master standby                                            = XXXGPM02
20150605:13:48:41:015288 gpstate:XXXGPM01:gpadmin-[INFO]:-   Standby master state                                      = Standby host passive
20150605:13:48:41:015288 gpstate:XXXGPM01:gpadmin-[INFO]:-   Total segment instance count from metadata                = 336
20150605:13:48:41:015288 gpstate:XXXGPM01:gpadmin-[INFO]:-----------------------------------------------------
20150605:13:48:41:015288 gpstate:XXXGPM01:gpadmin-[INFO]:-   Primary Segment Status
20150605:13:48:41:015288 gpstate:XXXGPM01:gpadmin-[INFO]:-----------------------------------------------------
20150605:13:48:41:015288 gpstate:XXXGPM01:gpadmin-[INFO]:-   Total primary segments                                    = 168
20150605:13:48:41:015288 gpstate:XXXGPM01:gpadmin-[INFO]:-   Total primary segment valid (at master)                   = 168
20150605:13:48:41:015288 gpstate:XXXGPM01:gpadmin-[INFO]:-   Total primary segment failures (at master)                = 0
20150605:13:48:41:015288 gpstate:XXXGPM01:gpadmin-[INFO]:-   Total number of postmaster.pid files missing              = 0
20150605:13:48:41:015288 gpstate:XXXGPM01:gpadmin-[INFO]:-   Total number of postmaster.pid files found                = 168
20150605:13:48:41:015288 gpstate:XXXGPM01:gpadmin-[INFO]:-   Total number of postmaster.pid PIDs missing               = 0
20150605:13:48:41:015288 gpstate:XXXGPM01:gpadmin-[INFO]:-   Total number of postmaster.pid PIDs found                 = 168
20150605:13:48:41:015288 gpstate:XXXGPM01:gpadmin-[INFO]:-   Total number of /tmp lock files missing                   = 0
20150605:13:48:41:015288 gpstate:XXXGPM01:gpadmin-[INFO]:-   Total number of /tmp lock files found                     = 168
20150605:13:48:41:015288 gpstate:XXXGPM01:gpadmin-[INFO]:-   Total number postmaster processes missing                 = 0
20150605:13:48:41:015288 gpstate:XXXGPM01:gpadmin-[INFO]:-   Total number postmaster processes found                   = 168
20150605:13:48:41:015288 gpstate:XXXGPM01:gpadmin-[INFO]:-----------------------------------------------------
20150605:13:48:41:015288 gpstate:XXXGPM01:gpadmin-[INFO]:-   Mirror Segment Status
20150605:13:48:41:015288 gpstate:XXXGPM01:gpadmin-[INFO]:-----------------------------------------------------
20150605:13:48:41:015288 gpstate:XXXGPM01:gpadmin-[INFO]:-   Total mirror segments                                     = 168
20150605:13:48:41:015288 gpstate:XXXGPM01:gpadmin-[INFO]:-   Total mirror segment valid (at master)                    = 168
20150605:13:48:41:015288 gpstate:XXXGPM01:gpadmin-[INFO]:-   Total mirror segment failures (at master)                 = 0
20150605:13:48:41:015288 gpstate:XXXGPM01:gpadmin-[INFO]:-   Total number of postmaster.pid files missing              = 0
20150605:13:48:41:015288 gpstate:XXXGPM01:gpadmin-[INFO]:-   Total number of postmaster.pid files found                = 168
20150605:13:48:41:015288 gpstate:XXXGPM01:gpadmin-[INFO]:-   Total number of postmaster.pid PIDs missing               = 0
20150605:13:48:41:015288 gpstate:XXXGPM01:gpadmin-[INFO]:-   Total number of postmaster.pid PIDs found                 = 168
20150605:13:48:41:015288 gpstate:XXXGPM01:gpadmin-[INFO]:-   Total number of /tmp lock files missing                   = 0
20150605:13:48:41:015288 gpstate:XXXGPM01:gpadmin-[INFO]:-   Total number of /tmp lock files found                     = 168
20150605:13:48:41:015288 gpstate:XXXGPM01:gpadmin-[INFO]:-   Total number postmaster processes missing                 = 0
20150605:13:48:41:015288 gpstate:XXXGPM01:gpadmin-[INFO]:-   Total number postmaster processes found                   = 168
20150605:13:48:41:015288 gpstate:XXXGPM01:gpadmin-[INFO]:-   Total number mirror segments acting as primary segments   = 0
20150605:13:48:41:015288 gpstate:XXXGPM01:gpadmin-[INFO]:-   Total number mirror segments acting as mirror segments    = 168
20150605:13:48:41:015288 gpstate:XXXGPM01:gpadmin-[INFO]:-----------------------------------------------------
[gpadmin@XXXGPM01 ~]$ 

本文转载自:http://www.cnblogs.com/jyzhao/p/4555171.html

闻术苑
粉丝 63
博文 167
码字总数 75630
作品 0
济南
产品经理
私信 提问
【干货】在docker中安装配置Greenplum集群的过程

—— 原文发布于本人的微信公众号“大数据与人工智能Lab”(BigdataAILab),欢迎关注。 Greenplum是一个MPP(海量并行处理)计算框架的分布式数据库,其数据库引擎层是基于著名的Postgresq...

雪饼
2018/01/12
2.5K
7
Deepgreen(Greenplum) 多机部署测试 , TPC-H VS citus

标签 PostgreSQL , deepgreen , greenplum , citus , tpch , 多机部署 背景 多机部署deepgreen,与greenplum部署方法类似。 本文仅用于测试。 1 环境 与之前测试citus的环境一致,9台ECS虚拟...

德哥
04/14
0
0
GreenPlum 5.10.0 集群部署

第1部分 初始化系统配置 1.1 部署环境 1.2 设置主机名、同步时间 1.3 添加hosts解析 cat > /etc/hosts << EOF 127.0.0.1 localhost localhost.localdomain localhost4 localhost4.localdomai......

faer615
2018/08/01
0
0
VMWare中CentOS6.8的Greenplum4.3.12安装

经过查看网络视频、查找网上资料。安装成功GP4.3。下面是安装过程的目录 1. 实验环境4 1.1. 硬件环境4 1.2. 虚拟机配置4 2. 系统设置(所有主节点和数据节点)4 2.1. 基础环境设置4 2.1.1. ...

moonstarseu
2017/05/05
0
0
如何将一个大规模并行的PostgreSQL数据库移植到Kubernetes上?

作者简介 Oz Basarir,Pivotal技术产品经理 译者简介 李伦,中兴通讯上海研发中心虚拟化架构师 校对者简介 朱君鹏,华东师范大学博士研究生,个人兴趣主要集中在:新型硬件(GPU、RDMA、FPG...

PostgreSQL中文社区
05/06
0
0

没有更多内容

加载失败,请刷新页面

加载更多

64.监控平台介绍 安装zabbix 忘记admin密码

19.1 Linux监控平台介绍 19.2 zabbix监控介绍 19.3/19.4/19.6 安装zabbix 19.5 忘记Admin密码如何做 19.1 Linux监控平台介绍: 常见开源监控软件 ~1.cacti、nagios、zabbix、smokeping、ope...

oschina130111
今天
9
0
当餐饮遇上大数据,嗯真香!

之前去开了一场会,主题是「餐饮领袖新零售峰会」。认真听完了餐饮前辈和新秀们的分享,觉得获益匪浅,把脑子里的核心纪要整理了一下,今天和大家做一个简单的分享,欢迎感兴趣的小伙伴一起交...

数澜科技
今天
7
0
DNS-over-HTTPS 的下一代是 DNS ON BLOCKCHAIN

本文作者:PETER LAI ,是 Diode 的区块链工程师。在进入软件开发领域之前,他主要是在做工商管理相关工作。Peter Lai 也是一位活跃的开源贡献者。目前,他正在与 Diode 团队一起开发基于区块...

红薯
今天
6
0
CC攻击带来的危害我们该如何防御?

随着网络的发展带给我们很多的便利,但是同时也带给我们一些网站安全问题,网络攻击就是常见的网站安全问题。其中作为站长最常见的就是CC攻击,CC攻击是网络攻击方式的一种,是一种比较常见的...

云漫网络Ruan
今天
11
0
实验分析性专业硕士提纲撰写要点

为什么您需要研究论文的提纲? 首先当您进行研究时,您需要聚集许多信息和想法,研究论文提纲可以较好地组织你的想法, 了解您研究资料的流畅度和程度。确保你写作时不会错过任何重要资料以此...

论文辅导员
今天
8
0

没有更多内容

加载失败,请刷新页面

加载更多

返回顶部
顶部