本篇采用命令行方式在三台虚机上部署SDB集群。

SequoiaDB v3.2.3及以上版本支持sac一键部署集群
此次部署集群使用centos7.x版本，每台虚机内存分配2g或以上

一. 安装前准备工作

1. 修改主机名

# 按照顺序，修改三台主机名分别为sdb01，sdb02，sdb03
vi /etc/hosts

2. 关闭防火墙

# 查看防火墙状态
systemctl status firewalld
# 永久关闭防火墙
systemctl  disable firewalld
# 确认关闭
systemctl status firewalld

3. 修改系统参数（可省略）

#########调整调整 ulimit############################
vim /etc/security/limits.conf
## for sequoiadb
sdbadmin  -  core   0
sdbadmin  -  data   unlimited
sdbadmin  -  fsize  unlimited
sdbadmin  -  rss    unlimited
sdbadmin  -  as     unlimited
sdbadmin  -  stack  1024
sdbadmin  -  nproc  unlimited
sdbadmin  -  nofile  300000
## end for sequoiadb



#检查/etc/security/limits.conf
cat /etc/security/limits.conf

vim /etc/security/limits.d/90-nproc.conf
将*          soft    nproc     1024修改为:# *          soft    nproc     1024
#新增
sdbadmin  -  nproc  unlimited

#检查 /etc/security/limits.d/90-nproc.conf
cat  /etc/security/limits.d/90-nproc.conf

4. 调整内核参数（可省略）

########调整内核参数################################
vim /etc/sysctl.conf
## begin for sequoiadb
vm.swappiness = 0
vm.dirty_ratio = 100
vm.dirty_background_ratio = 80
vm.dirty_expire_centisecs = 3000 
vm.dirty_writeback_centisecs = 500
vm.vfs_cache_pressure = 200
vm.min_free_kbytes = 10240
vm.overcommit_memory = 2
vm.overcommit_ratio = 85
fs.file-max = 10000000
## end for sequoiadb


#执行如下命令，使配置生效：
/sbin/sysctl -p

#检查配置是否生效
cat /proc/sys/vm/swappiness
cat /proc/sys/vm/dirty_ratio
cat /proc/sys/vm/dirty_background_ratio
cat /proc/sys/vm/dirty_expire_centisecs
cat /proc/sys/vm/vfs_cache_pressure
cat /proc/sys/vm/min_free_kbytes
cat /proc/sys/vm/overcommit_memory
cat /proc/sys/vm/overcommit_ratio

5. 关闭selinux

# 检查selinux状态
[root@localhost ~]# getenforce
Enforcing

# 永久关闭：
[root@localhost ~]# vim /etc/sysconfig/selinux
# SELINUX=enforcing 改为 SELINUX=disabled
# 然后需要重启系统:  init 6

5. 关闭numa，和transparent_hugepage

# 修改配置文件
# GRUB_CMDLINE_LINUX="crashkernel=auto rd.lvm.lv=centos/root rd.lvm.lv=centos/swap rhgb quiet numa=off transparent_hugepage=never",确保name值为off和transparent_hugepage=never
vi /etc/default/grub

# 重启服务器
reboot

# 查看numa状态
dmesg | grep -i numa
# 显示结果
[    0.000000] Command line: BOOT_IMAGE=/vmlinuz-3.10.0-327.el7.x86_64 root=/dev/mapper/centos-root ro crashkernel=auto rd.lvm.lv=centos/root rd.lvm.lv=centos/swap rhgb quiet numa=off
[    0.000000] NUMA turned off
[    0.000000] Kernel command line: BOOT_IMAGE=/vmlinuz-3.10.0-327.el7.x86_64 root=/dev/mapper/centos-root ro crashkernel=auto rd.lvm.lv=centos/root rd.lvm.lv=centos/swap rhgb quietnuma=off

二. 开始安装

1. 上传安装包，解压
2. 运行sequoiadb安装文件
3. 这里om服务只在sdb01上安装，其他两台不安装
4. 一路默认即可完成安装

三. 集群环境配置

1. 登录sdb01，切换sdbadmin用户，检查sdbcm运行状态

# 查看sdbcm运行状态
service status sdbcm
# 若停止，则开启
service start sdbcm

2. 进入sdb shell，创建临时协调节点（使用完毕之后需要删除）

sdb
>oma = new Oma("localhsot",11790);
# 创建临时协调节点
>oma.createCoord(18800, "/opt/sequoiadb/database/coord/18800")
# 启动临时协调节点
> oma.startNode(18800)

3. 配置和启动编目节点

# 连接临时协调节点
> var db = new Sdb("localhost",18800)
# 创建一个编目节点组
> db.createCataRG("sdb01", 11800, "/opt/sequoiadb/database/cata/11800")
# 创建另外两个编目节点
> var cataRG = db.getRG("SYSCatalogGroup"); 
> var node1 = cataRG.createNode("sdb02", 11800,"/opt/sequoiadb/database/cata/11800") 
> var node2 = cataRG.createNode("sdb03", 11800,"/opt/sequoiadb/database/cata/11800")
> node1.start()
> node2.start()

4. 配置和启动数据节点组

# 创建数据组
> db.createRG("datagroup01")
> dg = db.getRG("datagroup01")
# 增加节点
> dg.createNode("sdb01",11900,"/opt/sequoiadb/database/data/11900")
> dg.createNode("sdb02",11900,"/opt/sequoiadb/database/data/11900")
> dg.createNode("sdb03",11900,"/opt/sequoiadb/database/data/11900")
# 开启数据节点
> dg.start()

5. 配置和启动协调节点组

# 创建协调节点组
> var rg = db.createCoordRG()
> rg.createNode("sdb01",11810,"/opt/sequoiadb/database/coord/11810")
> rg.createNode("sdb02",11810,"/opt/sequoiadb/database/coord/11810")
> rg.createNode("sdb03",11810,"/opt/sequoiadb/database/coord/11810")
# 启动协调节点组
> rg.start()

6. 创建完毕之后，关闭临时协调节点并删除

> oma.removeCoord(18800)

7. 修改sequoiadb的limits.conf文件

# 使用sdbadmin用户登录
vim /opt/sequoiadb/conf/limits.conf
# 修改其中的 open_files=60000 为 open_files=300000
# 重启sdb集群节点
sdbstop
sdbstart

四 安装mysql

1. 运行sequoiadb安装tar中包含的mysql安装包，一路默认即可完成安装。
2. 配置mysql

1）增加mysql实例

# 打开mysql安装位置
cd /opt/sequoiasql/mysql/

# 运行脚本 增加实例 -p可指定端口号，默认3306
bin/sdb_sql_ctl addinst sdb01inst -D database/3306

# 检查实例配置，sdb信息是否正确
vim database/3306/auto.cnf

#修改配置信息之后重启实例
bin/sdb_sql_ctl restart sdb01inst 

#确认实例
bin/sdb_sql_ctl list

2） 修改mysql实例的最大连接数

# 使用sdbadmin用户修改mysql的最大连接数：
vi /opt/sequoiasql/mysql/database/3306/auto.cnf
max_connections=3000

# 使用root用户，在服务配置文件中增加两个配置参数：
vi  /usr/lib/systemd/system/sequoiasql-mysql.service
LimitNOFILE=infinity
LimitMEMLOCK=infinity

# 使用root用户，重启sequoiasql-mysql服务：
systemctl restart sequoiasql-mysql

3. 审计日志插件安装

mysql之间同步ddl需要配置审计日志

1）在mysql中创建sdbadmin用户

> CREATE USER 'sdbadmin'@'%' IDENTIFIED BY 'sdbadmin';
> GRANT all on *.* TO 'sdbadmin'@'%' with grant option;

2）将审计插件放到lib/plugin目录下

cp /opt/sequoiasql/mysql/tools/lib/server_audit.so /opt/sequoiasql/mysql/lib/plugin

3）配置mysql实例的配置文件，/opt/sequoiasql/mysql/database/3306/auto.cnf，在mysqld下添加以下内容

日志目录auditlog需要手动创建

# 加载审计插件
plugin-load=server_audit=server_audit.so
# 审计记录的审计，建议只记录需要同步的DCL和DDL操作
server_audit_events=CONNECT,QUERY_DDL,QUERY_DCL
# 开启审计
server_audit_logging=ON
# 审计日志文件名  
server_audit_file_path=/opt/sequoiasql/mysql/data/auditlog/server_audit.log
# 强制切分审计日志文件
server_audit_file_rotate_now=OFF
# 审计日志文件大小10MB，超过该大小进行切割，单位为byte
server_audit_file_rotate_size=10485760
# 审计日志保留个数，超过后会丢弃最旧的
server_audit_file_rotations=999
# 输出类型为文件
server_audit_output_type=file
# 限制每行查询日志的大小为100kb，若表比较复杂，对应的操作语句比较长，建议增大该值
server_audit_query_log_limit=102400

4）编辑config文件，配置需要同步的mysql 的ip和端口

5）启动同步脚本

python /opt/sequoiasql/mysql/tools/metaSync/meta_sync.py &

6）设置定时任务

crontab -e
# 每隔一秒执行一次
*/1 * * * * python /opt/sequoiasql/mysql/tools/metaSync/meta_sync.py &

7）重启mysql实例

sdb_sql_ctl restart sdb01inst

8）其他两台机器步骤相同

9）建库验证

五 安装pg

1. 运行sequoiadb压缩包中的pg包。
2. 一路默认即可完成安装。
3. 创建实例，配置pg

# 创建pg实例

#1. 进入安装目录
cd /opt/sequoiasql/postgresql/

#2. 创建实例
bin/sdb_sql_ctl addinst sdb01pginst -D database/5432

#3. 运行实例
bin/sdb_sql_ctl start sdb01pginst

4. 修改系统参数，不重启数据库，否则可能会报错server closed the connection unexpectedly This probably means the server…

vim /etc/systemd/logind.conf

# 将 RemoveIPC=yes 改为no

6. 配置pg实例的配置参数文件postgresql.conf

cd /opt/sequoiasql/postgresql/database/5432/
sed -i "s/#log_destination = 'stderr'/log_destination = 'stderr'/g" postgresql.conf
sed -i "s/#logging_collector = off/logging_collector = on/g" postgresql.conf
sed -i "s/#log_directory = 'pg_log'/log_directory = 'pg_log'/g" postgresql.conf
sed -i "s/#log_connections = off/log_connections = on/g" postgresql.conf
sed -i "s/#log_disconnections = off/log_disconnections = on/g" postgresql.conf
sed -i "s/#log_line_prefix = ''/log_line_prefix = '%m %p %r %x'/g" postgresql.conf
sed -i "s/#log_directory = pg_log/log_directory = pg_log/g" postgresql.conf
sed -i "s/#log_filename = /log_filename = /g" postgresql.conf
sed -i "s/#exit_on_error = off/exit_on_error = on/g" postgresql.conf
sed -i "s/#listen_addresses = 'localhost'/listen_addresses = '*'/g" postgresql.conf
sed -i "s/#standard_conforming_strings = on/standard_conforming_strings = off/g" postgresql.conf
sed -i "s/#log_min_duration_statement = -1/log_min_duration_statement = 100/g" postgresql.conf

5. 进入pg shell，创建pg表

# 创建数据库，进入pg shell界面
createdb sdbadmin01
psql sdbadmin01

# 加载sdb驱动
sdbadmin01=# create extension sdb_fdw;
# 配置与sdb连接的参数
sdbadmin01=# create server sdb_server foreign data wrapper sdb_fdw options(address '127.0.0.1', service '11810', user 'sdbUserName', password 'sdbPassword', preferedinstance 'A', transaction 'off');

# 关联sdb中的集合空间和集合（集合空间和集合需要在sdb中提前建好）
sdbadmin01=# create foreign table test(id int ,name varchar) server sdb_server options(collectionspace: "sdbadmin01",collection:"test",decimal:"on)
# 插入一条数据，在sdb shell中查看对应的集合中是否有这条数据
sdbadmin01=# insert into test values(234 ,"yangss")

六 安装spark

1. 环境要求，前期准备

spark需要jdk环境，分别在三台虚拟机配好jdk1.8的环境变量。

配置免秘钥

# 密钥生成*3
ssh-keygen -t rsa
# 拷贝公钥*3
ssh-copy-id node1

# 分发sdb01上的公钥文件
scp .ssh/authorized_keys sdb02:$PWD/.ssh
scp .ssh/authorized_keys sdb03:$PWD/.ssh

2.准备spark安装包，修改配置文件。（三台机器操作相同）

# 进入spark安装目录下的conf目录
cd /opt/spark-2.1.3-bin-hadoop2.7/conf
# 重命名文件
mv spark-env.sh.template spark-env.sh
# 修改，添加以下参数
cat > spark-env.sh <<EOF
#!/bin/bash/env bash
# 建议添加，不加也行
export JAVA_HOME=/opt/jdk1.8.0_201
# master节点
export SPARK_MASTER_HOST=sdb01
# master管理端口
export SPARK_MASTER_PORT=7077

EOF

# 修改salves文件，配置主机名，sdb01，sdb02，sdb03
vim salves

#
# Licensed to the Apache Software Foundation (ASF) under one or more
# contributor license agreements.  See the NOTICE file distributed with
# this work for additional information regarding copyright ownership.
# The ASF licenses this file to You under the Apache License, Version 2.0
# (the "License"); you may not use this file except in compliance with
# the License.  You may obtain a copy of the License at
#
#    http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
#

# A Spark Worker will be started on each of the machines listed below.
sdb01
sdb02
sdb03

3. 启动

start-all.sh