CDH高可用部署

前期准备

配置hosts

192.168.245.105 scm-node1
192.168.245.106 scm-node2
192.168.245.107 scm-node3

设置hostname

在192.168.245.105上执行

sudo hostnamectl --static --transient set-hostname scm-node1

在192.168.245.106上执行

sudo hostnamectl --static --transient set-hostname scm-node2

在192.168.245.107上执行

sudo hostnamectl --static --transient set-hostname scm-node3

关闭防火墙

CentOS-6

在scm-node2、scm-node3两台主机上执行以下命令：

sudo chkconfig iptables off

sudo service iptables stop

CentOS-7

在scm-node2、scm-node3两台主机上执行以下命令：

sudo chkconfig firewalld off

sudo service firewalld stop

关闭SELinux

在scm-node2、scm-node3两台主机上执行以下命令：

sudo sed -i ‘/SELINUX/s/enforcing/disabled/’ /etc/selinux/config

安装软件

在scm-node2、scm-node3两台主机上已安装CDH、MySQL。

其他配置

已设置NTP时钟同步和双机互信。

安装nfs

在线安装nfs

在scm-node1主机上执行以下命令：

sudo yum -y install nfs-utils rpcbind

离线安装nfs

CentOS-6安装包

keyutils-1.4-5.el6.x86_64.rpm
libevent-1.4.13-4.el6.x86_64.rpm
libgssglue-0.1-11.el6.x86_64.rpm
libtirpc-0.2.1-15.el6.x86_64.rpm
nfs-utils-1.2.3-78.el6_10.1.x86_64.rpm
nfs-utils-lib-1.1.5-13.el6.x86_64.rpm
python-argparse-1.2.1-2.1.el6.noarch.rpm
rpcbind-0.2.0-16.el6.x86_64.rpm

CentOS-7安装包

gssproxy-0.7.0-17.el7.x86_64.rpm
keyutils-1.5.8-3.el7.x86_64.rpm
libbasicobjects-0.1.1-29.el7.x86_64.rpm
libevent-2.0.21-4.el7.x86_64.rpm
libini_config-1.3.1-29.el7.x86_64.rpm
libcollection-0.7.0-29.el7.x86_64.rpm
libpath_utils-0.2.1-29.el7.x86_64.rpm
libnfsidmap-0.25-19.el7.x86_64.rpm
libref_array-0.1.5-29.el7.x86_64.rpm
libtirpc-0.2.4-0.10.el7.x86_64.rpm
libverto-libevent-0.2.5-4.el7.x86_64.rpm
nfs-utils-1.3.0-0.54.el7.x86_64.rpm
quota-nls-4.01-17.el7.noarch.rpm
rpcbind-0.2.0-44.el7.x86_64.rpm
quota-4.01-17.el7.x86_64.rpm
tcp_wrappers-7.6-77.el7.x86_64.rpm

安装所有RPM包

在scm-node1主机上进入安装包所在目录，然后执行以下命令：

sudo rpm -ivh *.rpm

参数说明

ro：共享目录只读；
rw：共享目录可读可写；
all_squash：所有访问用户都映射为匿名用户或用户组；
no_all_squash（默认）：访问用户先与本机用户匹配，匹配失败后再映射为匿名用户或用户组；
root_squash（默认）：将来访的root用户映射为匿名用户或用户组；
no_root_squash：来访的root用户保持root帐号权限；
anonuid=<UID>：指定匿名访问用户的本地用户UID，默认为nfsnobody（65534）；
anongid=<GID>：指定匿名访问用户的本地用户组GID，默认为nfsnobody（65534）；
secure（默认）：限制客户端只能从小于1024的tcp/ip端口连接服务器；
insecure：允许客户端从大于1024的tcp/ip端口连接服务器；
sync：将数据同步写入内存缓冲区与磁盘中，效率低，但可以保证数据的一致性；
async：将数据先保存在内存缓冲区中，必要时才写入磁盘；
wdelay（默认）：检查是否有相关的写操作，如果有则将这些写操作一起执行，这样可以提高效率；
no_wdelay：若有写操作则立即执行，应与sync配合使用；
subtree_check（默认）
：若输出目录是一个子目录，则nfs服务器将检查其父目录的权限；
no_subtree_check
：即使输出目录是一个子目录，nfs服务器也不检查其父目录的权限，这样可以提高效率；

启动nfs

sudo service rpcbind start

sudo service nfs start

sudo chkconfig rpcbind on

sudo chkconfig nfs on

安装corosync+pacemaker

在线安装corosync+pacemaker

在scm-node2、scm-node3两台主机上执行以下命令：

sudo yum install -y corosync pacemaker

离线安装corosync+pacemaker

CentOS-6安装包

corosync-1.4.7-6.el6.x86_64.rpm
corosynclib-1.4.7-6.el6.x86_64.rpm
ConsoleKit-0.4.1-6.el6.x86_64.rpm
ConsoleKit-libs-0.4.1-6.el6.x86_64.rpm
avahi-libs-0.6.25-17.el6.x86_64.rpm
cifs-utils-4.8.1-20.el6.x86_64.rpm
clusterlib-3.0.12.1-84.el6.x86_64.rpm
cman-3.0.12.1-84.el6.x86_64.rpm
cvs-1.11.23-16.el6.x86_64.rpm
cyrus-sasl-md5-2.1.23-15.el6_6.2.x86_64.rpm
dbus-1.2.24-9.el6.x86_64.rpm
dmidecode-2.12-7.el6.x86_64.rpm
eggdbus-0.6-3.el6.x86_64.rpm
fence-agents-4.0.15-13.el6_9.2.x86_64.rpm
fence-virt-0.2.3-24.el6.x86_64.rpm
gettext-0.17-18.el6.x86_64.rpm
gnutls-2.12.23-22.el6.x86_64.rpm
gnutls-utils-2.12.23-22.el6.x86_64.rpm
hal-0.5.14-14.el6.x86_64.rpm
hal-info-20090716-5.el6.noarch.rpm
hal-libs-0.5.14-14.el6.x86_64.rpm
hdparm-9.43-4.el6.x86_64.rpm
ipmitool-1.8.15-2.el6.x86_64.rpm
libgomp-4.4.7-23.el6.x86_64.rpm
libqb-0.17.1-2.el6.x86_64.rpm
libtalloc-2.1.5-1.el6_7.x86_64.rpm
libtdb-1.3.8-3.el6_8.2.x86_64.rpm
libtevent-0.9.26-2.el6_7.x86_64.rpm
libvirt-client-0.10.2-64.el6.x86_64.rpm
libxslt-1.1.26-2.el6_3.1.x86_64.rpm
libibverbs-1.1.8-4.el6.x86_64.rpm
libnl-1.1.4-2.el6.x86_64.rpm
librdmacm-1.0.21-0.el6.x86_64.rpm
lm_sensors-libs-3.1.1-17.el6.x86_64.rpm
modcluster-0.16.2-35.el6.x86_64.rpm
nc-1.84-24.el6.x86_64.rpm
net-snmp-utils-5.5-60.el6.x86_64.rpm
net-snmp-libs-5.5-60.el6.x86_64.rpm
numactl-2.0.9-2.el6.x86_64.rpm
oddjob-0.30-6.el6.x86_64.rpm
openais-1.1.1-7.el6.x86_64.rpm
openaislib-1.1.1-7.el6.x86_64.rpm
pacemaker-1.1.18-3.el6.x86_64.rpm
pacemaker-cli-1.1.18-3.el6.x86_64.rpm
pacemaker-cluster-libs-1.1.18-3.el6.x86_64.rpm
pacemaker-libs-1.1.18-3.el6.x86_64.rpm
parted-2.1-29.el6.x86_64.rpm
pciutils-3.1.10-4.el6.x86_64.rpm
perl-Net-Telnet-3.03-11.el6.noarch.rpm
perl-TimeDate-1.16-13.el6.noarch.rpm
pexpect-2.3-6.el6.noarch.rpm
pm-utils-1.2.5-11.el6.x86_64.rpm
polkit-0.96-11.el6.x86_64.rpm
pyOpenSSL-0.13.1-2.el6.x86_64.rpm
python-suds-0.4.1-3.el6.noarch.rpm
quota-3.17-23.el6.x86_64.rpm
rdma-6.9_4.1-3.el6.noarch.rpm
resource-agents-3.9.5-46.el6.x86_64.rpm
ricci-0.16.2-87.el6.x86_64.rpm
samba-common-3.6.23-51.el6.x86_64.rpm
samba-winbind-3.6.23-51.el6.x86_64.rpm
samba-winbind-clients-3.6.23-51.el6.x86_64.rpm
sg3_utils-1.28-13.el6.x86_64.rpm
sg3_utils-libs-1.28-13.el6.x86_64.rpm
tcp_wrappers-7.6-58.el6.x86_64.rpm
telnet-0.17-48.el6.x86_64.rpm
yajl-1.0.7-3.el6.x86_64.rpm

CentOS-7安装包

corosync-2.4.3-2.el7_5.1.x86_64.rpm
corosynclib-2.4.3-2.el7_5.1.x86_64.rpm
bc-1.06.95-13.el7.x86_64.rpm
cifs-utils-6.2-10.el7.x86_64.rpm
cups-libs-1.6.3-35.el7.x86_64.rpm
libldb-1.2.2-1.el7.x86_64.rpm
libtalloc-2.1.10-1.el7.x86_64.rpm
libtevent-0.9.33-2.el7.x86_64.rpm
libtdb-1.3.15-1.el7.x86_64.rpm
libwbclient-4.7.1-9.el7_5.x86_64.rpm
libcgroup-0.41-15.el7.x86_64.rpm
libxslt-1.1.28-5.el7.x86_64.rpm
libqb-1.0.1-6.el7.x86_64.rpm
pacemaker-cluster-libs-1.1.18-11.el7_5.3.x86_64.rpm
perl-TimeDate-2.30-2.el7.noarch.rpm
pacemaker-1.1.18-11.el7_5.3.x86_64.rpm
pacemaker-cli-1.1.18-11.el7_5.3.x86_64.rpm
psmisc-22.20-15.el7.x86_64.rpm
samba-common-4.7.1-9.el7_5.noarch.rpm
resource-agents-3.9.5-124.el7.x86_64.rpm
samba-common-libs-4.7.1-9.el7_5.x86_64.rpm
pacemaker-libs-1.1.18-11.el7_5.3.x86_64.rpm
samba-client-libs-4.7.1-9.el7_5.x86_64.rpm

安装所有RPM包

在scm-node2、scm-node3两台主机进入安装包所在目录，然后执行以下命令：

sudo rpm -ivh *.rpm

安装crmsh

下载crmsh

CentOS-6下载地址

http://download.opensuse.org/repositories/network:/ha-clustering:/Stable/CentOS_CentOS-6/noarch/

CentOS-7下载地址

http://download.opensuse.org/repositories/network:/ha-clustering:/Stable/CentOS_CentOS-7/noarch/

离线安装crmsh

CentOS-6安装包

crmsh-3.0.0-6.1.noarch.rpm
crmsh-scripts-3.0.0-6.1.noarch.rpm
python-parallax-1.0.1-28.1.noarch.rpm
python-lxml-2.2.3-1.1.el6.x86_64.rpm
python-six-1.9.0-2.el6.noarch.rpm
python-dateutil-1.4.1-7.el6.noarch.rpm
redhat-rpm-config-9.0.3-51.el6.centos.noarch.rpm

CentOS-7安装包

libxslt-1.1.28-5.el7.x86_64.rpm
python-dateutil-1.5-7.el7.noarch.rpm
python-lxml-3.2.1-4.el7.x86_64.rpm
python-parallax-1.0.1-29.1.noarch.rpm
crmsh-scripts-3.0.0-6.2.noarch.rpm
crmsh-3.0.0-6.2.noarch.rpm

安装所有RPM包

在scm-node2、scm-node3两台主机进入安装包所在目录，然后执行以下命令：

sudo rpm -ivh *.rpm

配置corosync+pacemaker集群

配置corosync

配置version1.x(CentOS-6)

在scm-node2主机上执行以下命令：

sudo vi /etc/corosync/corosync.conf

/etc/corosync/corosync.conf内容：

compatibility: whitetank
totem {
version: 2
secauth: off
interface {
member {
memberaddr: scm-node2
}
member {
memberaddr: scm-node3
}
ringnumber: 0
bindnetaddr: scm-node2
mcastport: 5405
}
transport: udpu
}
logging {
fileline: off
to_logfile: yes
to_syslog: yes
logfile: /var/log/cluster/corosync.log
debug: off
timestamp: on
logger_subsys {
subsys: AMF
debug: off
}
}
service {
name: pacemaker
ver: 0
use_mgmtd: yes
}

在scm-node3主机上执行以下命令：

sudo vi /etc/corosync/corosync.conf

/etc/corosync/corosync.conf内容：

compatibility: whitetank
totem {
version: 2
secauth: off
interface {
member {
memberaddr: scm-node2
}
member {
memberaddr: scm-node3
}
ringnumber: 0
bindnetaddr: scm-node3
mcastport: 5405
}
transport: udpu
}
logging {
fileline: off
to_logfile: yes
to_syslog: yes
logfile: /var/log/cluster/corosync.log
debug: off
timestamp: on
logger_subsys {
subsys: AMF
debug: off
}
}
service {
name: pacemaker
ver: 0
use_mgmtd: yes
}

配置version2.x(CentOS-7)

在scm-node2、scm-node3两台主机上执行以下命令：

sudo vi /etc/corosync/corosync.conf

/etc/corosync/corosync.conf内容：

totem {
version: 2
secauth: off
cluster_name: cmf
transport: udpu
}
nodelist {
node {
ring0_addr: scm-node2
nodeid: 1
}
node {
ring0_addr: scm-node3
nodeid: 2
}
}
logging {
fileline: off
to_logfile: yes
to_syslog: yes
logfile: /var/log/cluster/corosync.log
debug: off
timestamp: on
logger_subsys {
subsys: AMF
debug: off
}
}
quorum {
provider: corosync_votequorum
two_node: 1
}

启动集群

启动version1.x(CentOS-6)

启动服务：

sudo service corosync start

开机启动：

sudo chkconfig corosync on

启动version2.x(CentOS-7)

启动服务：

sudo service pacemaker start

sudo service corosync start

开机启动：

sudo chkconfig pacemaker on

sudo chkconfig corosync on

配置集群

禁用仲裁检查

在scm-node2主机上执行以下命令：

sudo crm configure property no-quorum-policy=ignore

在两个节点中，当节点达不到法定票数时（节点数不是奇数），即两个节点一个坏了，没法投票，正常的节点达不到法定票数，此时如果是默认参数，即正常的机器不能工作，所以需要该为ignore，使正常机器接管。

禁用stonith

在scm-node2主机上执行以下命令：

sudo crm configure property stonith-enabled=false

因为我们这里没有stonith设备所有要禁用。

修改默认粘性值

在scm-node2主机上执行以下命令：

sudo crm configure rsc_defaults resource-stickiness= 100

一些环境中会要求尽量避免资源在节点之间移动。移动资源通常一位置一段时间内无法提供服务，某些负载的服务，比如Oracle数据库，这个时间可能会很长。为了达到这个效果，pacemaker有一个叫做资源粘性值的概念，它能够控制一个服务（资源）有多想待在它正在运行的节点上。你可以把它认为是无法提供服务的“代价”。pacemaker为了达到最优分部各个资源的目的，默认设置这个值为0.我们可以为每个资源定义不同的粘性值，但一般来说更改默认粘性值就够了。

检查配置是否正确

在scm-node2、scm-node3两台主机上执行以下命令：

sudo crm_verify -L -V

假若没有输出任何则配置正确。

常用操作

查看集群状态

在scm-node2主机上执行以下命令：

sudo crm status

查看资源配置

在scm-node2主机上执行以下命令：

sudo crm configure show

查看资源代理

在scm-node2主机上执行以下命令：

sudo crm ra classes

移动资源位置

在scm-node2主机上执行以下命令：

sudo crm resource move cloudera-scm-server scm-node3

定义克隆集群

定义2个克隆资源：

sudo crm configure clone mysql-cluster mysql clone-max=2 clone-node-max=2 notify=true

停止克隆：

sudo crm resource stop mysql-cluster

删除克隆：

sudo crm configure delete mysql-cluster

注意：

1.mysql-cluster是自定义集群名

2.必须先定义原生(primitive)mysql资源，详情见6.1.2

定义资源分组

定义分组，可以保证mysql和cloudera-scm-server在同一节点上：

sudo crm configure group server-group mysql cloudera-scm-server

停止分组：

sudo crm resource stop server-group

删除分组：

sudo crm configure delete server-group

注意：proxy-group是自定义组名。

定义资源约束

定义排列约束，mysql和cloudera-scm-server必须在同一节点上：

sudo crm configure colocation mysql-with-cloudera-scm-server inf: mysql
cloudera-scm-server

定义顺序约束，先启动mysql之后才启动cloudera-scm-server：

sudo crm configure order mysql_before_cloudera-scm-server mandatory: mysql cloudera-scm-server

定义位置约束，资源vip固定在scm-node2：

sudo crm configure location vip_pref_node2 vip inf: scm-node2

删除约束：

sudo crm configure delete mysql-with-cloudera-scm-server

sudo crm configure delete mysql_before_cloudera-scm-server

sudo crm configure delete vip_pref_node2

注意：mysql-with-cloudera-scm-server、mysql_before_cloudera-scm-server、vip_pref_node2是自定义约束名

删除集群资源

sudo crm resource stop vip

sudo crm configure delete vip

注意：必须先停止资源，才能删除。

配置CDH HA集群

配置nfs

配置mysql

在scm-node1主机上创建文件夹：

sudo mkdir -p /media/mysql

在scm-node1主机上给文件夹授权：

sudo chmod 666 /media/mysql

在scm-node1主机上配置exports：

sudo vi /etc/exports

/etc/exports添加内容：

/media/mysql scm-node*(rw,async,no_root_squash,no_subtree_check)

注意：scm-node*表示允许主机名以scm-node开头的主机访问/media/mysql

配置生效命令：

sudo exportfs -r

在scm-node2、scm-node3两台主机上挂载目录：

sudo vi /etc/fstab

在/etc/fstab文件最后添加内容：

scm-node1:/media/mysql /var/lib/mysql nfs auto,noatime,nolock,intr,tcp,actimeo=1800 0 0

配置cloudera-scm-server

在scm-node1主机上创建文件夹：

sudo mkdir -p /media/cloudera-scm-server

在scm-node1主机上给文件夹授权：

sudo chmod 666 /media/ cloudera-scm-server

配置exports：

sudo vi /etc/exports

/etc/exports添加内容：

/media/ cloudera-scm-server scm-node*(rw,async,no_root_squash,no_subtree_check)

配置生效命令：

sudo exportfs -r

在scm-node2、scm-node3两台主机上文件夹授权：

sudo chown -R cloudera-scm:cloudera-scm /var/lib/cloudera-scm-server

在scm-node2、scm-node3两台主机上挂载目录：

sudo vi /etc/fstab

在/etc/fstab最后添加内容：

scm-node1:/media/cloudera-scm-server /var/lib/cloudera-scm-server nfs auto,noatime,nolock,intr,tcp,actimeo=1800 0 0

配置cloudera-scm-agent

在scm-node1主机上创建文件夹：

sudo mkdir -p /media/cloudera-scm-agent

sudo mkdir -p /media/cloudera-host-monitor

sudo mkdir -p /media/cloudera-scm-eventserver

sudo mkdir -p /media/cloudera-service-monitor

在scm-node1主机上给文件夹授权：

sudo chmod 666 /media/ cloudera-scm-agent

sudo chmod 666 /media/ cloudera-host-monitor

sudo chmod 666 /media/ cloudera-scm-eventserver

sudo chmod 666 /media/ cloudera-service-monitor

配置exports：

sudo vi /etc/exports

/etc/exports添加内容：

/media/ cloudera-scm-agent scm-node*(rw,async,no_root_squash,no_subtree_check)

/media/ cloudera-host-monitor scm-node*(rw,async,no_root_squash,no_subtree_check)

/media/ cloudera-scm-eventserver scm-node*(rw,async,no_root_squash,no_subtree_check)

/media/ cloudera-service-monitor scm-node*(rw,async,no_root_squash,no_subtree_check)

配置生效命令：

sudo exportfs -r

在scm-node2、scm-node3两台主机上文件夹授权：

sudo chown -R cloudera-scm:cloudera-scm /var/lib/cloudera-scm-agent

sudo chown -R cloudera-scm:cloudera-scm /var/lib/cloudera-host-monitor

sudo chown -R cloudera-scm:cloudera-scm /var/lib/cloudera-scm-eventserver

sudo chown -R cloudera-scm:cloudera-scm /var/lib/cloudera-service-monitor

在scm-node2、scm-node3两台主机上挂载目录：

sudo vi /etc/fstab

在/etc/fstab最后添加内容：

scm-node1:/media/cloudera-scm-agent /var/lib/cloudera-scm-agent nfs
auto,noatime,nolock,intr,tcp,actimeo=1800 0 0

scm-node1:/media/cloudera-host-monitor /var/lib/cloudera- host-monitor nfs
auto,noatime,nolock,intr,tcp,actimeo=1800 0 0

scm-node1:/media/cloudera-scm-eventserver /var/lib/cloudera-scm-eventserver
nfs auto,noatime,nolock,intr,tcp,actimeo=1800 0 0

scm-node1:/media/cloudera-service-monitor /var/lib/cloudera- service-monitor
nfs auto,noatime,nolock,intr,tcp,actimeo=1800 0 0

在CDH管理界面执行以下操作：

1.停止Cloudera Management Service，然后删除

2.关闭HTTP Referer Check

添加资源

添加VIP资源

在scm-node2主机上执行以下命令：

sudo crm configure primitive vip ocf?IPaddr2 params
ip=‘192.168.245.165’ op monitor interval=5s timeout=20s on-fail=restart

注意：vip是自定义资源名，ip必须和当前主机在同一网段。

添加mysql资源

在scm-node2、scm-node3主机上执行以下命令(CentOS-6)：

sudo chkconfig mysql off

在scm-node2、scm-node3主机上执行以下命令(CentOS-7)：

sudo chkconfig mysqld off

在scm-node2主机上执行以下命令(CentOS-6)：

sudo crm configure primitive mysql lsb:mysql op monitor interval=20s
timeout=100s on-fail=restart

在scm-node2主机上执行以下命令(CentOS-7)：

sudo crm configure primitive mysql systemd:mysqld op monitor interval=20s
timeout=100s on-fail=restart

添加cloudera-scm-server资源

在scm-node2、scm-node3主机上执行以下命令(CentOS-6)：

sudo chkconfig cloudera-scm-server off

在scm-node2、scm-node3主机上执行以下命令(CentOS-7)：

sudo chkconfig cloudera-scm-server off

在scm-node2、scm-node3主机上修改db.propertie：

sudo vi /etc/cloudera-scm-server/db.properties

/etc/cloudera-scm-server/db.properties变更内容：

com.cloudera.cmf.db.host=192.168.245.165

在scm-node2主机上添加cloudera-scm-server资源：

sudo crm configure primitive cloudera-scm-server lsb:cloudera-scm-server op
monitor interval=20s timeout=40s on-fail=restart

添加cloudera-scm-agent资源

在scm-node2、scm-node3主机上执行以下命令(CentOS-6)：

sudo chkconfig cloudera-scm-agent off

在scm-node2、scm-node3主机上执行以下命令(CentOS-7)：

sudo chkconfig cloudera-scm-agent off

在scm-node2、scm-node3主机上创建文件夹：

sudo mkdir -p /usr/lib/ocf/resource.d/cm

在scm-node2、scm-node3主机上创建文件：

sudo vi /usr/lib/ocf/resource.d/cm/cloudera-scm-agent

/usr/lib/ocf/resource.d/cm/cloudera-scm-agent内容(CentOS-6)：

#!/bin/sh
#######################################################################
# CM Agent OCF script
#######################################################################
#######################################################################
# Initialization:
: ${__OCF_ACTION=$1}
OCF_SUCCESS=0
OCF_ERROR=1
OCF_STOPPED=7
#######################################################################
meta_data() {
cat <<END
<?xml version=“1.0”?>
<!DOCTYPE resource-agent SYSTEM “ra-api-1.dtd”>
<resource-agent name=“Cloudera Manager Agent” version=“1.0”>
<version>1.0</version>
<longdesc lang=“en”>
This OCF agent handles simple monitoring, start, stop of the Cloudera
Manager Agent, intended for use with Pacemaker/corosync for failover.
</longdesc>
<shortdesc lang=“en”>Cloudera Manager Agent OCF script</shortdesc>
<parameters />
<actions>
<action name=“start” timeout=“20” />
<action name=“stop” timeout=“20” />
<action name=“monitor” timeout=“20” interval=“10” depth=“0”/>
<action name=“meta-data” timeout=“5” />
</actions>
</resource-agent>
END
}
#######################################################################
agent_usage() {
cat <<END
usage: $0 {start|stop|monitor|meta-data}
Cloudera Manager Agent HA OCF script - used for managing Cloudera Manager
Agent and managed processes lifecycle for use with Pacemaker.
END
}
agent_start() {
service cloudera-scm-agent start
if [ $? = 0 ]; then
return $OCF_SUCCESS
fi
return $OCF_ERROR
}
agent_stop() {
service cloudera-scm-agent hard_stop_confirmed
if [ $? = 0 ]; then
return $OCF_SUCCESS
fi
return $OCF_ERROR
}
agent_monitor() {
# Monitor _MUST!_ differentiate correctly between running
# (SUCCESS), failed (ERROR) or _cleanly_ stopped (NOT RUNNING).
# That is THREE states, not just yes/no.
service cloudera-scm-agent status
if [ $? = 0 ]; then
return $OCF_SUCCESS
fi
return $OCF_STOPPED
}
case $__OCF_ACTION in
meta-data) meta_data
exit $OCF_SUCCESS
;;
start) agent_start;;
stop) agent_stop;;
monitor) agent_monitor;;
usage|help) agent_usage
exit $OCF_SUCCESS
;;
*) agent_usage
exit $OCF_ERR_UNIMPLEMENTED
;;
esac
rc=$?
exit $rc
/usr/lib/ocf/resource.d/cm/cloudera-scm-agent内容(CentOS-7)：
#!/bin/sh
#######################################################################
# CM Agent OCF script
#######################################################################
#######################################################################
# Initialization:
: ${__OCF_ACTION=$1}
OCF_SUCCESS=0
OCF_ERROR=1
OCF_STOPPED=7
#######################################################################
meta_data() {
cat <<END
<?xml version=“1.0”?>
<!DOCTYPE resource-agent SYSTEM “ra-api-1.dtd”>
<resource-agent name=“Cloudera Manager Agent” version=“1.0”>
<version>1.0</version>
<longdesc lang=“en”>
This OCF agent handles simple monitoring, start, stop of the Cloudera
Manager Agent, intended for use with Pacemaker/corosync for failover.
</longdesc>
<shortdesc lang=“en”>Cloudera Manager Agent OCF script</shortdesc>
<parameters />
<actions>
<action name=“start” timeout=“20” />
<action name=“stop” timeout=“20” />
<action name=“monitor” timeout=“20” interval=“10” depth=“0”/>
<action name=“meta-data” timeout=“5” />
</actions>
</resource-agent>
END
}
#######################################################################
agent_usage() {
cat <<END
usage: $0 {start|stop|monitor|meta-data}
Cloudera Manager Agent HA OCF script - used for managing Cloudera Manager
Agent and managed processes lifecycle for use with Pacemaker.
END
}
agent_start() {
service cloudera-scm-agent start
if [ $? = 0 ]; then
return $OCF_SUCCESS
fi
return $OCF_ERROR
}
agent_stop() {
service cloudera-scm-agent next_stop_hard
service cloudera-scm-agent stop
if [ $? = 0 ]; then
return $OCF_SUCCESS
fi
return $OCF_ERROR
}
agent_monitor() {
# Monitor _MUST!_ differentiate correctly between running
# (SUCCESS), failed (ERROR) or _cleanly_ stopped (NOT RUNNING).
# That is THREE states, not just yes/no.
service cloudera-scm-agent status
if [ $? = 0 ]; then
return $OCF_SUCCESS
fi
return $OCF_STOPPED
}
case $__OCF_ACTION in
meta-data) meta_data
exit $OCF_SUCCESS
;;
start) agent_start;;
stop) agent_stop;;
monitor) agent_monitor;;
usage|help) agent_usage
exit $OCF_SUCCESS
;;
*) agent_usage
exit $OCF_ERR_UNIMPLEMENTED
;;
esac
rc=$?
exit $rc

在scm-node2、scm-node3主机上给文件授权：

sudo chmod 770 /usr/lib/ocf/resource.d/cm/cloudera-scm-agent

在scm-node2、scm-node3主机上修改config.ini：

sudo vi /etc/cloudera-scm-agent/config.ini

/etc/cloudera-scm-agent/config.ini变更内容：

server_host=192.168.245.165
lib_dir=/var/lib/cloudera-scm-agent

在scm-node2主机上添加cloudera-scm-agent资源：

sudo crm configure primitive cloudera-scm-agent ocf:cm:cloudera-scm-agent op
monitor interval=20s timeout=40s on-fail=restart

注意：添加cloudera-scm-agent资源主要是为了Cloudera Management
Service故障转移，如果不需要可以不加

定义约束

在scm-node2主机上定义CDH分组：

sudo crm configure group cdh-group vip mysql cloudera-scm-server
cloudera-scm-agent

在scm-node2主机上定义启动顺序约束：

sudo crm configure order vip_before_mysql mandatory: vip mysql

sudo crm configure order mysql_before_cloudera-scm-server mandatory: mysql
cloudera-scm-server

sudo crm configure order cloudera-scm-server_before_cloudera-scm-agent
mandatory: cloudera-scm-server cloudera-scm-agent