greenplum install
注意事項：
1.hostname 不能大寫##
2.如果segment和mirror副本數比較大，建議調整/etc/ssh/sshd_config的maxconnect數值
3.系統參數需要修改kernel.sem = 500 1024000 200 8192 ##############這個配置是三個節點 6塊磁盤，每個磁盤6個segment 6個mirror 如果segment超過這個比例的話，需要對應調整sem的參數
4.要使用gpadmin這個用戶安裝軟件 vim /etc/sudoers 賦予gpadmin root的權限
# User privilege specification
root ALL=(ALL:ALL) ALL
gpadmin ALL=(ALL:ALL) ALL
安裝步驟
一.master standby work節點系統參數調整
1.1用戶准備
sudo groupadd -g 530 gpadmin
sudo useradd -g 530 -u 530 -m -d /home/gpadmin -s /bin/bash gpadmin
sudo chown -R gpadmin:gpadmin /home/gpadmin/
sudo passwd gpadmin
sudo vim /etc/hosts
# hosts for greenplum
192.168.0.200 mpp01
192.168.2.201 mpp02
192.168.0.202 mpp03
su gpadmin
#免密登陸
ssh-keygen -t rsa
cp ~/.ssh/id_rsa.pub ~/.ssh/authorized_keys
chmod 600 ~/.ssh/authorized_keys
chmod 700 ~/.ssh
ssh-copy-id mpp02
ssh-copy-id mpp03
#####修改ubuntu18.04的主機名
沒有這個文件創建這個文件
/etc/cloud/cloud.cfg
preserve_hostname: true
1.2 系統參數優化
#參考文檔：https://blog.csdn.net/ctypyb2002/article/details/84107389
sudo vim /etc/sysctl.conf
xfs_mount_options = rw,noatime,inode64,allocsize=16m ###如果不修改這個gpcheck會提示異常
kernel.shmmax = 500000000
kernel.shmmni = 4096
kernel.shmall = 4000000000
#kernel.sem = 500 1024000 200 8192 （108個segment及其以上,使用這個sem參數）
kernel.sem = 500 1024000 200 4096 （108個segment以下,使用這個sem參數）
kernel.sysrq = 1
kernel.core_uses_pid = 1
kernel.msgmnb = 1265536
kernel.msgmax = 1265536
kernel.msgmni = 2048
net.ipv4.tcp_syncookies = 1
net.ipv4.ip_forward = 0
net.ipv4.conf.default.accept_source_route = 0
net.ipv4.tcp_tw_recycle = 1
net.ipv4.tcp_max_syn_backlog = 4096
net.ipv4.conf.all.arp_filter = 1
net.ipv4.ip_local_port_range = 1025 65535
net.core.netdev_max_backlog = 10000
net.core.rmem_max = 2097152
net.core.wmem_max = 2097152
vm.overcommit_memory = 2
sudo sysctl -p
1.3 #修改打開文件句柄的數量
vim /etc/security/limits.conf
* soft nofile 65536
* hard nofile 65536
* soft nproc 131072
* hard nproc 131072
1.4 #設置磁盤訪問I/O調度策略（stat磁盤 選擇deadline ）
Linux磁盤I/O調度器對磁盤的訪問支持不同的策略，默認的為CFQ，GP建議設置為deadline 要查看某驅動器的I/O調度策略，可通過如下命令查看，下面示例的為正確的配置：
# cat /sys/block/{devname}/queue/scheduler
noop anticipatory [deadline] cfq
針對我們服務器上的磁盤 修改i/o調度策略
echo deadline > /sys/block/sda/queue/scheduler
echo deadline > /sys/block/sdb/queue/scheduler
echo deadline > /sys/block/sdc/queue/scheduler
echo deadline > /sys/block/sdd/queue/scheduler
echo deadline > /sys/block/sde/queue/scheduler
echo deadline > /sys/block/sdf/queue/scheduler
1.5 #打開磁盤預熱
每個磁盤設備文件需要配置read-ahead(blockdev)值為65536
官方文檔的推薦值為16384，但譯者認為應該為65536更合理，該值設置的是預讀扇區數，實際上預讀的字節數是blockdev設置除以2，而GP缺省的blocksize為32KB，剛好與65536(32768B/32KB)對應。
65536/16384 ？ 待定 我默認使用官網的配置
/sbin/blockdev --getra /dev/sda
/sbin/blockdev --setra 16384 /dev/sda
/sbin/blockdev --getra /dev/sdb
/sbin/blockdev --setra 16384 /dev/sdb
/sbin/blockdev --getra /dev/sdc
/sbin/blockdev --setra 16384 /dev/sdc
/sbin/blockdev --getra /dev/sdd
/sbin/blockdev --setra 16384 /dev/sdd
/sbin/blockdev --getra /dev/sde
/sbin/blockdev --setra 16384 /dev/sde
/sbin/blockdev --getra /dev/sdf
/sbin/blockdev --setra 16384 /dev/sdf
1.6 #修改sshd的配置
vim /etc/ssh/sshd_config
UseDNS no
#PidFile /var/run/sshd.pid
MaxStartups 100:30:1000
#添加ubuntu 18.04 開機自啟的服務
sudo vi /etc/systemd/system/rc-local.service
開機任務(https://www.jianshu.com/p/79d24b4af4e5)
[Unit]
Description=/etc/rc.local Compatibility
ConditionPathExists=/etc/rc.local
[Service]
Type=forking
ExecStart=/etc/rc.local start
TimeoutSec=0
StandardOutput=tty
RemainAfterExit=yes
SysVStartPriority=99
[Install]
WantedBy=multi-user.target
sudo vi /etc/rc.local
#!/bin/sh -e
#
# rc.local
#
# This script is executed at the end of each multiuser runlevel.
# Make sure that the script will "exit 0" on success or any other
# value on error.
#
# In order to enable or disable this script just change the execution
# bits.
#
# By default this script does nothing.
netplan apply
exit 0
sudo chmod +x /etc/rc.local
sudo systemctl enable rc-local
sudo systemctl start rc-local.service
sudo systemctl status rc-local.service
二、安裝greenplum
###################master和standby work節點（mpp集群每個節點都要執行）
2.1 安裝gp-xercesc
git clone https://github.com/greenplum-db/gp-xerces.git
cd gp-xerces
mkdir build
cd build
../configure --prefix=/usr/local && make -j64
sudo make -j64 install
#安裝re2c
tar -zxvf re2c-1.1.1.tar.gz
cd re2c-1
./configure
make -j4
sudo make -j4 install
#安裝ninja
tar -zxvf ninja-1.9.0.tar.gz
cd ninja
./configure.py --bootstrap
sudo cp -rp ninja /usr/bin
2.2 安裝gporca
git clone https://github.com/greenplum-db/gporca.git
cd gporca-3.39.0/
sudo apt install cmake
cmake -GNinja -H. -Bbuild
sudo ninja install -C build
2.3 master節點安裝gpdb（mpp的master節點執行）
git clone https://github.com/greenplum-db/gpdb.git ####選擇合適的gpdb版本https://github.com/greenplum-db/gpdb/releases/ #我選擇的版本是https://github.com/greenplum-db/gpdb/releases/tag/6.0.0-beta.3
cd gpdb
######安裝需要的依賴關系
sudo ./README.ubuntu.bash
#將這個腳本拷貝到mpp其他的節點 安裝基礎包
scp ./README.ubuntu.bash gpadmin@mpp02:/home/gpadmin
scp ./README.ubuntu.bash gpadmin@mpp03:/home/gpadmin
此時系統會默認安裝libxerces-c-dev庫，需要將此庫移除以使其正確編譯。
sudo apt remove libxerces-c-dev ###此庫會有沖突
./configure --enable-orca --with-perl --with-python --with-libxml --prefix=/usr/local/gpdb
make -j64 && sudo make -j64 install
#################
sudo ldconfig #################一定要重載庫文件，否則會報沒有已安裝的兩個插件的庫文件（如libgpos.so）
2.4 standby和work節點安裝greenplum
cd /home/gpadmin
######安裝需要的依賴關系
sudo ./README.ubuntu.bash
壓縮master節點greenplum的安裝目錄，然后把壓縮文件拷貝到standby和wodk的節點相同的目錄，然后解壓
tar -czvf gpdb.tar.gz /usr/local/gpdb
scp -rp gpdb.tar.gz gpadmin@mpp02:/home/gpadmin
scp -rp gpdb.tar.gz gpadmin@mpp03:/home/gpadmin
ssh gadmin@mpp02 "cd /home/gpadmin && tar -zxvf gpdb.tar.gz && mv gpdb /usr/local && sudo ldconfig "
ssh gadmin@mpp03 "cd /home/gpadmin && tar -zxvf gpdb.tar.gz && mv gpdb /usr/local && sudo ldconfig "
2.5 修改所有節點gpadmin的環境變量
#master 節點 和standby節點
sudo mkdir -p /data1/master
sudo chown -R gpadmin:gpadmin /data1/master
vim /home/gpadmin/.profile
source /usr/local/gpdb/greenplum_path.sh
export PGPORT=5432
export PGDATABASE=gp_sydb
TZ='Asia/Shanghai'; export TZ
export MASTER_DATA_DIRECTORY=/data1/master/gpseg-1/ #master節點的數據目錄
#然后每個節點執行
sudo ldconfig && source /home/gpadmin/.profile && source /usr/local/gpdb/greenplum_path.sh
不重啟刷新磁盤
ls /sys/class/scsi_host/
"---" 將后面的值改為空，重新刷新磁盤
sudo echo "- - -" > /sys/class/scsi_host/host0/scan
sudo echo "- - -" > /sys/class/scsi_host/host1/scan
sudo echo "- - -" > /sys/class/scsi_host/host2/scan
sudo echo "- - -" > /sys/class/scsi_host/host3/scan
sudo echo "- - -" > /sys/class/scsi_host/host4/scan
sudo echo "- - -" > /sys/class/scsi_host/host5/scan
sudo echo "- - -" > /sys/class/scsi_host/host6/scan
sudo echo "- - -" > /sys/class/scsi_host/host7/scan
sudo echo "- - -" > /sys/class/scsi_host/host8/scan
sudo echo "- - -" > /sys/class/scsi_host/host9/scan
sudo echo "- - -" > /sys/class/scsi_host/host10/scan
sudo echo "- - -" > /sys/class/scsi_host/host11/scan
sudo echo "- - -" > /sys/class/scsi_host/host12/scan
sudo echo "- - -" > /sys/class/scsi_host/host13/scan
sudo echo "- - -" > /sys/class/scsi_host/host14/scan
sudo echo "- - -" > /sys/class/scsi_host/host15/scan
sudo echo "- - -" > /sys/class/scsi_host/host16/scan
sudo echo "- - -" > /sys/class/scsi_host/host17/scan
sudo echo "- - -" > /sys/class/scsi_host/host18/scan
sudo echo "- - -" > /sys/class/scsi_host/host19/scan
sudo echo "- - -" > /sys/class/scsi_host/host20/scan
sudo echo "- - -" > /sys/class/scsi_host/host21/scan
sudo echo "- - -" > /sys/class/scsi_host/host22/scan
sudo echo "- - -" > /sys/class/scsi_host/host23/scan
sudo echo "- - -" > /sys/class/scsi_host/host24/scan
sudo echo "- - -" > /sys/class/scsi_host/host25/scan
sudo echo "- - -" > /sys/class/scsi_host/host26/scan
sudo echo "- - -" > /sys/class/scsi_host/host27/scan
sudo echo "- - -" > /sys/class/scsi_host/host28/scan
sudo echo "- - -" > /sys/class/scsi_host/host29/scan
sudo echo "- - -" > /sys/class/scsi_host/host31/scan
#########################################
2.6 初始化准備(segment和mirror)
segment 存放數據
mirror是segment的鏡像
#################每個節點執行掛載
mkfs.xfs /dev/sda
mkfs.xfs /dev/sdb
mkfs.xfs /dev/sdc
mkfs.xfs /dev/sdd
mkfs.xfs /dev/sde
mkfs.xfs /dev/sdf
#自動掛載
vim /etc/fstab
/dev/sda /data1 xfs defaults 0 0
/dev/sdb /data2 xfs defaults 0 0
/dev/sdc /data3 xfs defaults 0 0
/dev/sdd /data4 xfs defaults 0 0
/dev/sde /data5 xfs defaults 0 0
/dev/sdf /data6 xfs defaults 0 0
mount -a
##################
mkdir -p /data{1..6}/primary/s{1..6}
mkdir -p /data{1..6}/mirror/s{1..6}
mkdir /data1/master
chown -R gpadmin:gpadmin /data{1..6}
#########################3個盤 3個segment
sudo mkfs.xfs /dev/sdb
sudo vim /etc/fstab
/dev/sdb /data1 xfs defaults 0 0
sudo mkdir /data1 && sudo mount -a
sudo mkdir -p /data1/primary/s1 ###創建segment數據目錄
sudo mkdir -p /data1/mirror/s1 ###創建mirror數據目錄
sudo mkdir /data1/master
sudo chown -R gpadmin:gpadmin /data1
2.7 初始化greenplum 集群
cd /usr/local/gpdb/docs/cli_help/gpconfigs
cp gpinitsystem_config /home/gpadmin/conf/gp_conf
vim gp_conf
ARRAY_NAME="Greenplum Data Platform"
SEG_PREFIX=gpseg
PORT_BASE=6000
ARRAY_NAME="gp_sydb"
declare -a DATA_DIRECTORY=(/data1/primary/s1 /data1/primary/s2 ) ######每個目錄就是一個segment 集群的segment總數是 當前節點segment目錄數量*mpp集群節點數量
MASTER_HOSTNAME=mpp01
MASTER_DIRECTORY=/data1/master
MASTER_DATA_DIRECTORY=/data1/master/gpseg-1
DATABASE_NAME=gp_sydb
#添加seg_hosts
MACHINE_LIST_FILE=/home/gpadmin/conf/seg_hosts
MASTER_PORT=5432
TRUSTED_SHELL=ssh
CHECK_POINT_SEGMENTS=8
ENCODING=UNICODE
#mirror config
MIRROR_PORT_BASE=43000
REPLICATION_PORT_BASE=34000
MIRROR_REPLICATION_PORT_BASE=44000
declare -a MIRROR_DATA_DIRECTORY=(/data1/mirror/s1 /data1/mirror/s2 ) ######每個目錄就是一個mirror 集群的mirror總數是 當前節點mirror目錄數量*mpp集群節點數量
#####確認mpp節點主機名
cat /home/gpadmin/conf/seg_hosts
mpp01
mpp02
mpp03
#####初始化greenplum
每個節點執行
mkdir -p /data1/primary/s1 /data1/primary/s2 && mkdir -p /data1/mirror/s1 /data1/mirror/s2
########declare -a DATA_DIRECTORY=(/data1/primary/s1 /data1/primary/s2 ) ######每個目錄就是一個segment 集群的segment總數是 當前節點segment目錄數量*mpp集群節點數量
########declare -a MIRROR_DATA_DIRECTORY=(/data1/mirror/s1 /data1/mirror/s2 ) ######每個目錄就是一個mirror 集群的mirror總數是 當前節點mirror目錄數量*mpp集群節點數量
rm -rf /data1/master/gpseg-1
rm -rf /data1/primary/s{1..4}/*
初始化之前：
#在各個節點執行，用來檢查gpdb是否正確安裝，自檢相關的配置文件
gpssh-exkeys -f seg_hosts master到各個節點通信(測試一下免密登陸)
gpcheck --local （檢查每個節點的狀態）
gpinitsystem -c gp_conf -h seg_hosts
#遇到的報錯
1.selecting default max_connections ... /usr/local/gpdb/bin/postgres: error while loading shared libraries: libgpopt.so.3: cannot open shared object file: No such file or directory
no data was returned by command ""/usr/local/gpdb/bin/postgres" -V"
The program "postgres" is needed by initdb but was either not found in the same directory as "/usr/local/gpdb/bin/initdb" or failed unexpectedly.
Check your installation; "postgres -V" may have more information.
每個節點重新安裝 插件gporca gp-xercesc 重新加載一下庫文件 sudo ldconfig
#添加配置mirror
2.mirror error
/usr/local/gpdb/bin/gpinitsystem: line 1000: ((: MIRROR_REPLICATION_PORT_OFFSET=-: syntax error: operand expected (error token is "-")
20190517:17:31:29:006271 gpinitsystem:mpp01:gpadmin-[INFO]:-Building group mirror array type , please wait...
./usr/local/gpdb/bin/lib/gp_bash_functions.sh: line 585: 0+: syntax error: operand expected (error token is "+")
20190517:17:31:29:006271 gpinitsystem:mpp01:gpadmin-[INFO]:-Building the Master instance database, please wait...
MIRROR_PORT_BASE=43000
REPLICATION_PORT_BASE=34000
MIRROR_REPLICATION_PORT_BASE=44000
#####################修改添加所有的磁盤做seg和mirror segment和mirror的數量要對應，如果不對應會導致初始化的時候中斷
mkdir
/data1/primary /data1/primary /data1/primary /data1/primary /data1/primary /data1/primary /data2/primary /data2/primary /data2/primary /data2/primary /data2/primary /data2/primary /data3/primary /data3/primary /data3/primary /data3/primary /data3/primary /data3/primary /data4/primary /data4/primary /data4/primary /data4/primary /data4/primary /data4/primary /data5/primary /data5/primary /data5/primary /data5/primary /data5/primary /data5/primary /data6/primary /data6/primary /data6/primary /data6/primary /data6/primary /data6/primary
/data1/mirror /data1/mirror /data1/mirror /data1/mirror /data1/mirror /data1/mirror /data2/mirror /data2/mirror /data2/mirror /data2/mirror /dat
a2/mirror /data2/mirror /data2/mirror /data3/mirror /data3/mirror /data3/mirror /data3/mirror /data3/mirror /data3/mirror /data4/mirror /data4/mirror /data4/mirror /data4/mirror /data4/mirror /data4/mirror /data5/mirror /data5/mirror /data5/mirror /data5/mirror /data5/mirror /data5/mirror /data6/mirror /data6/mirror /data6/mirror /data6/mirror /data6/mirror /data6/mirror /data6/mirror
如果初始化失敗需要刪除mirror和primary里面所有的數據
rm -rf /data{1..6}/primary/s{1..6}/*
rm -rf /data{1..6}/mirror/s{1..6}/*
刪除/data1/master/下所有的數據
rm -rf /data1/master/gpseg-1
########檢查是否存在漏刪的文件：如果存在也會導致初始化報錯，一次初始化檢測，檢查節點的時間比較久，請大家慎重check！
ls -alt /data{1..6}/mirror/s{1..6} && ls -alt /data{1..6}/primary/s{1..6}/
ls -alt /data1/master/
#然后繼續初始化greenplum
gpinitsystem -c gp_conf -h seg_hosts
greenplum 常用的命令說明：
參考文檔：https://www.cnblogs.com/pl-boke/p/9852383.html
########################
添加standby
#gpinitstandby -r -a 刪除故障的standby節點
gpinitstandby -r -a mpp02
##########
20190523:11:00:55:010715 gpstate:mpp01:gpadmin-[WARNING]:-Standby PID = 0 <<<<<<<<
20190523:11:00:55:010715 gpstate:mpp01:gpadmin-[WARNING]:-Standby status = Standby process not running <<<<<<<<
#gpinitstandby -a -s 添加standby的節點
gpinitstandby -a -s mpp02
#添加之前在mpp02 上准備/data1/master/gpseg-1，添加standby master會拷貝master節點的配置到standby master
#查看集群狀態gpstate -f
gpadmin@mpp01:~$ gpstate -f
20190523:11:20:48:011923 gpstate:mpp01:gpadmin-[INFO]:-Starting gpstate with args: -f
20190523:11:20:48:011923 gpstate:mpp01:gpadmin-[INFO]:-local Greenplum Version: 'postgres (Greenplum Database) 6.0.0-beta.1 build dev'
20190523:11:20:48:011923 gpstate:mpp01:gpadmin-[INFO]:-master Greenplum Version: 'PostgreSQL 9.4.20 (Greenplum Database 6.0.0-beta.1 build dev) on x86_64-unknown-linux-gnu, compiled by gcc (Ubuntu 7.4.0-1ubuntu1~18.04) 7.4.0, 64-bit compiled on May 21 2019 14:00:08'
20190523:11:20:48:011923 gpstate:mpp01:gpadmin-[INFO]:-Obtaining Segment details from master...
20190523:11:20:49:011923 gpstate:mpp01:gpadmin-[INFO]:-Standby master details
20190523:11:20:49:011923 gpstate:mpp01:gpadmin-[INFO]:-----------------------
20190523:11:20:49:011923 gpstate:mpp01:gpadmin-[INFO]:- Standby address = mpp02
20190523:11:20:49:011923 gpstate:mpp01:gpadmin-[INFO]:- Standby data directory = /data1/master/gpseg-1
20190523:11:20:49:011923 gpstate:mpp01:gpadmin-[INFO]:- Standby port = 5432
20190523:11:20:49:011923 gpstate:mpp01:gpadmin-[INFO]:- Standby PID = 23755
20190523:11:20:49:011923 gpstate:mpp01:gpadmin-[INFO]:- Standby status = Standby host passive
20190523:11:20:49:011923 gpstate:mpp01:gpadmin-[INFO]:--------------------------------------------------------------
20190523:11:20:49:011923 gpstate:mpp01:gpadmin-[INFO]:--pg_stat_replication
20190523:11:20:49:011923 gpstate:mpp01:gpadmin-[INFO]:--------------------------------------------------------------
20190523:11:20:49:011923 gpstate:mpp01:gpadmin-[INFO]:--WAL Sender State: streaming
20190523:11:20:49:011923 gpstate:mpp01:gpadmin-[INFO]:--Sync state: sync
20190523:11:20:49:011923 gpstate:mpp01:gpadmin-[INFO]:--Sent Location: 0/140000A0
20190523:11:20:49:011923 gpstate:mpp01:gpadmin-[INFO]:--Flush Location: 0/140000A0
20190523:11:20:49:011923 gpstate:mpp01:gpadmin-[INFO]:--Replay Location: 0/140000A0
20190523:11:20:49:011923 gpstate:mpp01:gpadmin-[INFO]:--------------------------------------------------------------
###############################
模擬故障切換
gpstop -a -m 停掉master的postgres
讓mpp02變成主的master
gpadmin@mpp02:/data1/master/gpseg-1$ gpactivatestandby -a -d /data1/master/gpseg-1
sh: 0: getcwd() failed: No such file or directory
20190523:11:27:44:025254 gpactivatestandby:mpp02:gpadmin-[CRITICAL]:-PGPORT environment variable not set.
solution: check /home/gpadmin/.profile MASTER_DATA_DIRECTORY=/data1/master/gpseg-1/
gpactivatestandby -a -d /data1/master/gpseg-1
#主從切換后需要做數據恢復
TIPS：如果集群出現過primary和Mirror節點的切換，則最好再執行下下面的命令：
gprecoverseg -r 執行這步操作的原因：如果主節點down了，mirror節點接管后，會造成部分節點負擔過重
gpadmin@mpp02:~$ gprecoverseg -r
20190523:13:37:12:010532 gprecoverseg:mpp02:gpadmin-[INFO]:-Starting gprecoverseg with args:
20190523:13:37:12:010532 gprecoverseg:mpp02:gpadmin-[INFO]:-local Greenplum Version: 'postgres (Greenplum Database) 6.0.0-beta.1 build dev'
20190523:13:37:12:010532 gprecoverseg:mpp02:gpadmin-[INFO]:-master Greenplum Version: 'PostgreSQL 9.4.20 (Greenplum Database 6.0.0-beta.1 build dev) on x86_64-unknown-linux-gnu, compiled by gcc (Ubuntu 7.4.0-1ubuntu1~18.04) 7.4.0, 64-bit compiled on May 21 2019 14:00:08'
20190523:13:37:12:010532 gprecoverseg:mpp02:gpadmin-[INFO]:-Obtaining Segment details from master...
LOG: failed to acquire resources on one or more segments
DETAIL: FATAL: no pg_hba.conf entry for host "192.168.22.122", user "gpadmin", database "template1"
(seg18 192.168.22.122:43018)
LOG: query plan with multiple segworker groups is not supported
HINT: likely caused by a function that reads or modifies data in a distributed table
20190523:13:37:17:010532 gprecoverseg:mpp02:gpadmin-[CRITICAL]:-gprecoverseg failed. (Reason='FATAL: DTM initialization: failure during startup recovery, retry failed, check segment status (cdbtm.c:1328)
') exiting...
修改/data1/master/gpseg-1/pg_hba.conf 添加訪問ip到pg_hba.conf
host all gpadmin 172.16.1.114/32 trust
host all gpadmin 172.16.1.115/32 trust
host all gpadmin 172.16.1.116/32 trust
host all gpadmin 192.168.22.111/32 trust
host all gpadmin 192.168.22.122/32 trust
host all gpadmin 192.168.22.133/32 trust
#主備切換之后
Total primary segment failures (at master)
[WARNING]:-Total primary segment failures (at master)
[WARNING]:-Total number mirror segments acting as primary segments
#恢復segment mirror
gprecoverseg
#### gpstate -m
20190523:13:44:11:034230 gprecoverseg:mpp02:gpadmin-[INFO]:-Recovery 72 of 72
20190523:13:44:11:034230 gprecoverseg:mpp02:gpadmin-[INFO]:----------------------------------------------------------
20190523:13:44:11:034230 gprecoverseg:mpp02:gpadmin-[INFO]:- Synchronization mode = Incremental
20190523:13:44:11:034230 gprecoverseg:mpp02:gpadmin-[INFO]:- Failed instance host = mpp03
20190523:13:44:11:034230 gprecoverseg:mpp02:gpadmin-[INFO]:- Failed instance address = mpp03
20190523:13:44:11:034230 gprecoverseg:mpp02:gpadmin-[INFO]:- Failed instance directory = /data6/primary/s6/gpseg107
20190523:13:44:11:034230 gprecoverseg:mpp02:gpadmin-[INFO]:- Failed instance port = 6035
20190523:13:44:11:034230 gprecoverseg:mpp02:gpadmin-[INFO]:- Recovery Source instance host = mpp01
20190523:13:44:11:034230 gprecoverseg:mpp02:gpadmin-[INFO]:- Recovery Source instance address = mpp01
20190523:13:44:11:034230 gprecoverseg:mpp02:gpadmin-[INFO]:- Recovery Source instance directory = /data6/mirror/s6/gpseg107
20190523:13:44:11:034230 gprecoverseg:mpp02:gpadmin-[INFO]:- Recovery Source instance port = 43035
20190523:13:44:11:034230 gprecoverseg:mpp02:gpadmin-[INFO]:- Recovery Target = in-place
20190523:13:44:11:034230 gprecoverseg:mpp02:gpadmin-[INFO]:----------------------------------------------------------
Continue with segment recovery procedure Yy|Nn (default=N): 輸入Y
修復集群
gprecoverseg -a
20190523:13:53:49:034230 gprecoverseg:mpp02:gpadmin-[INFO]:-Ensuring that shared memory is cleaned up for stopped segments
20190523:13:53:50:034230 gprecoverseg:mpp02:gpadmin-[INFO]:-Updating configuration with new mirrors
20190523:13:53:50:034230 gprecoverseg:mpp02:gpadmin-[INFO]:-Updating mirrors
20190523:13:53:50:034230 gprecoverseg:mpp02:gpadmin-[INFO]:-Running pg_rewind on required mirrors
檢查集群修復的狀態
gpstate -m
########
20190523:13:53:48:038747 gpstate:mpp02:gpadmin-[INFO]:- mpp03 /data5/mirror/s6/gpseg65 43029 Passive Synchronized
20190523:13:53:48:038747 gpstate:mpp02:gpadmin-[INFO]:- mpp03 /data6/mirror/s1/gpseg66 43030 Passive Synchronized
20190523:13:53:48:038747 gpstate:mpp02:gpadmin-[INFO]:- mpp03 /data6/mirror/s2/gpseg67 43031 Passive Synchronized
20190523:13:53:48:038747 gpstate:mpp02:gpadmin-[INFO]:- mpp03 /data6/mirror/s3/gpseg68 43032 Passive Synchronized
20190523:13:53:48:038747 gpstate:mpp02:gpadmin-[INFO]:- mpp03 /data6/mirror/s4/gpseg69 43033 Passive Synchronized
20190523:13:53:48:038747 gpstate:mpp02:gpadmin-[INFO]:- mpp03 /data6/mirror/s5/gpseg70 43034 Passive Synchronized
20190523:13:53:48:038747 gpstate:mpp02:gpadmin-[INFO]:- mpp03 /data6/mirror/s6/gpseg71 43035 Passive Synchronized
20190523:13:53:48:038747 gpstate:mpp02:gpadmin-[INFO]:- mpp01 /data1/mirror/s1/gpseg72 43000 Acting as Primary Not In Sync
20190523:13:53:48:038747 gpstate:mpp02:gpadmin-[INFO]:- mpp01 /data1/mirror/s2/gpseg73 43001 Acting as Primary Not In Sync
20190523:13:53:48:038747 gpstate:mpp02:gpadmin-[INFO]:- mpp01 /data1/mirror/s3/gpseg74 43002 Acting as Primary Not In Sync
20190523:13:53:48:038747 gpstate:mpp02:gpadmin-[INFO]:- mpp01 /data1/mirror/s4/gpseg75 43003 Acting as Primary Not In Sync
20190523:13:53:48:038747 gpstate:mpp02:gpadmin-[INFO]:- mpp01 /data1/mirror/s5/gpseg76 43004 Acting as Primary Not In Sync
20190523:13:53:48:038747 gpstate:mpp02:gpadmin-[INFO]:- mpp01 /data1/mirror/s6/gpseg77 43005 Acting as Primary Not In Sync
20190523:13:53:48:038747 gpstate:mpp02:gpadmin-[INFO]:- mpp01 /data2/mirror/s1/gpseg78 43006 Acting as Primary Not In Sync
20190523:13:53:48:038747 gpstate:mpp02:gpadmin-[INFO]:- mpp01 /data2/mirror/s2/gpseg79 43007 Acting as Primary Not In Sync
#####
修復成功
20190523:14:00:32:034230 gprecoverseg:mpp02:gpadmin-[INFO]:-For segments updated successfully, streaming will continue in the background.
20190523:14:00:32:034230 gprecoverseg:mpp02:gpadmin-[INFO]:-Use gpstate -s to check the streaming progress.
20190523:14:00:32:034230 gprecoverseg:mpp02:gpadmin-[INFO]:-******************************************************************
修復primary segment 和 mirror segments
gpstate -m 查看greenplum的狀態
gpstop -a -M fast 關閉整個greenplum 集群
以restricted方式啟動數據庫
gpstart -a -R