This instance is used to deploy the experiment environment. It uses the Citrix virtualization environment and allocates three RHEL6.4 hosts.
This instance is used to deploy the experiment environment. It uses the Citrix virtualization environment and allocates three RHEL6.4 hosts.
1. Prepare three hosts
This instance is used to deploy the experiment environment. It uses the Citrix virtualization environment and allocates three RHEL6.4 hosts.
After the template is created for the Master, an additional 20 GB disk/dev/xvdb is added, and two NICs eth1 and eth2 are added.
After the Standby template is created, an additional 20 GB disk/dev/xvdb is added, and two NICs eth1 and eth2 are added.
After the Segment01 template is created, an additional 50 GB disk/dev/xvdb is added, and two NICs eth1 and eth2 are added.
Network Planning
Eth0 (external IP) eth1eth2
Master192.168.9.123172.16.10.101172.16.11.101
Standby192.168.9.122.162.16.10.102172.16.11.102
Segment01194259.125 (optional) 172.16.10.1172.16.11.1
The experiment environment resources are limited. Currently, three nodes are configured. Segment02 and Segment03. may be added as needed in the future...
Modify host name
Set the three host names of Master, Standby, and Segment01 to mdw, smdw, and sdw1 respectively.
Host Name modification method:
Hostname host name vi/etc/sysconfig/network modify hostname
Options: configure the script. It is optional to facilitate synchronization between nodes.
Export NODE_LIST = 'mdw SMDW sdw'
Vi/etc/hosts temporary Configuration
192.168.9.123 mdw192.168.9.124 smdw192.168.9.125 sdw1
Configure the first node to log on to itself and other machines without a password
Ssh-keygen-t rsassh-copy-id-I/root /. ssh/id_rsa.pub root@192.168.9.123ssh-copy-id-I/root /. ssh/id_rsa.pub root@192.168.9.124ssh-copy-id-I/root /. ssh/id_rsa.pub root@192.168.9.125cluster _ run_all_nodes "hostname; date"
Disk Planning
Gp is recommended to use the xfs file system. dependencies must be installed on all nodes.
# Rpm-ivh xfsprogs-3.1.1-10.el6.x86_64.rpm
Create/data folders for all nodes to mount the xfs File System
Mkdir/data
Mkfs. xfs/dev/xvdb
[Root @ smdb Packages], agsize = 1310720 blks =, imaxpct = 25 = blksinternal log, version = 2 = blks, none, rtextents = 0
Add the following line to vi/etc/fstab:
/Dev/xvdb/data xfs rw, noatime, inode64, allocsize = 16m1 12. disable iptables and containers "hostname; service iptables stop" cluster_run_all_nodes "hostname; chkconfig iptables off" kernel "hostname; chkconfig ip6tables off" kernel "hostname; chkconfig lib1_d off" kernel "hostname; setenforce 0 "cluster_run_all_nodes" hostname; sestatus "vi/etc/selinux/configcluster_copy_all_nodes/etc/selinux/config/etc/selinux/
Note: All nodes must be set in a unified manner. I have configured trust first and synchronized using scripts. If no configuration is available, each node must be set in sequence.
3. Set Recommended System Parameters
Vi/etc/sysctl. conf
Kernel. shmmax = 5001000000kernel. shmmni = 4096kernel. shmall = 4000000000kernelkernel. sysrq = 1kernel. core_uses_pid = 1kernel. msgmnb = 65536kernel. msgmax = 65536net. ipv4.tcp _ syncookies = 1net. ipv4.ip _ forward = 0net = 0net. ipv4.tcp _ tw_recycle = 1net. ipv4.tcp _ max_syn_backlog = 4096net = 1net = 1net. core. netdev_max_backlog = VMS. overcommit_memory = 2kernel. msgmni = 2048net
Vi/etc/security/limits. conf
* Soft nofile 65536 * hard nofile 65536 * soft nproc 131072 * hard nproc 131072
Synchronize to each node:
Cluster_copy_all_nodes/etc/sysctl. conf/etc/sysctl. confcluster_copy_all_nodes/etc/security/limits. conf/etc/security/limits. conf
Disk pre-read parameters and deadline Algorithm
Add
Blockdev -- setra 16385/dev/xvdbecho deadline>/sys/block/xvdb/queue/schedulercluster_copy_all_nodes/etc/rc. d/rc. local/etc/rc. d/rc. local
Note: Check whether blockdev -- getra/dev/xvdb takes effect after restart.
Verify the character set of all nodes
Cluster_run_all_nodes "hostname; echo $ LANG"
Restart all nodes to verify that the modification takes effect:
Blockdev -- getra/dev/xvdb more/sys/block/xvdb/queue/schedulercluster_run_all_nodes "hostname; service iptables status" 4. install mkdir-p/data/softon the Master to upload greenplum-db-4.3.4.2-build-1-rhel5-x86_64.zip to the Masterunzip greenplum-db-4.3.4.2-build-1-RHEL5-x86_64.zip/bin/bash greenplum-db-4.3.4.2-build-1-RHEL5-x86_64.bin5. install and configure Greenplum on all nodes
Configure/etc/hosts
192.168.9.123 mdw172.16.10.101 mdw-1172.16.11.101 mdw-2192.168.9.124 smdw172.16.10.102 smdw-1172.16.11.102 smdw-2192.168.9.125 sdw1172.16.10.1 sdw1-1172.16.11.1 sdw1-2
Synchronize/etc/hosts configurations
Cluster_copy_all_nodes/etc/hosts
Configure mutual trust required by gp