Galera集群负载均衡器部署

Question

我们有多个网站连接到一个运行MariaDB - 10.3.11的服务器上，由于负载过重，我们决定实现Galera Cluster replication，并且在数据库中更新记录时得到了预期的结果。

我们需要在这些Galera集群节点之间分配流量(而不是单个服务器连接)，因此我们需要部署Galera Cluster Load Balancer。我们在一个新服务器上安装了它，配置在/etc/default/glbd中，如下所示：(在Ubuntu 16.04上运行的所有服务器)。

# This is a configuration file for glbd service script
#
# On Red Hat and derivatives it should be placed in /etc/sysconfig/glbd
#
# On Debian  and derivatives it should be placed in /etc/default/glbd
#
# All settings besides LISTEN_ADDR are optional.
#

# Address to listen for client connections at. Mandatory parameter.
# To bind to all interfaces only port should be specified.
LISTEN_ADDR="3306"

# Address for controlling connection. Mandatory part is port.
# If not specified control socket will not be opened
CONTROL_ADDR="4444"

# Control FIFO name. It is opened always. glbd will refuse to start if
# this file already exists.
CONTROL_FIFO="/var/run/glbd.fifo"

# Number of threads (connection pools) to use. It is always a good idea
# to have a few per CPU core.
THREADS="3"

# Maximum connections. System open files limit will be modified to accommodate
# at least that many client connections provided sufficient privileges.
# Normally you should set it if you plan to go above 500.
#MAX_CONN=

# Target servers for client connections, space separated.
# Target specification: IP[:PORT[:WEIGHT]] (WEIGHT and PORT optional)
# WEIGHT defaults to 1, PORT defaults to LISTEN_ADDR port.
DEFAULT_TARGETS="node1_ip:3306:1 node2_ip:3306:1 node3_ip:3306:1"

# Other glbd options if any as they would appear on the command line.
OTHER_OPTIONS="--round"

我们假设LISTEN_ADDR是3306，因为我们需要与安装了Galera集群负载均衡器的新服务器交换数据库连接&监听3306来代替单个服务器(运行MariaDB)。如果我们对端口的猜测是错误的，请纠正我们。

Section A定义了当前的体系结构，Section B在下面的图中定义了拟议的架构更改：

以下是服务启动时Galera集群负载均衡器的日志。

Jan 24 12:49:54 galera-lb-server systemd[1]: Starting LSB: run glbd daemon...
Jan 24 12:49:54 galera-lb-server glb[12112]: [Thu Jan 24 12:49:54 UTC 2019] glbd: starting...
Jan 24 12:49:54 galera-lb-server glb[12112]:    INFO: glb_daemon.c:44: Changing effective user to 'daemon'
Jan 24 12:49:54 galera-lb-server glb[12112]: glb v1.0.1 (epoll)
Jan 24 12:49:54 galera-lb-server glb[12112]: Incoming address: 0.0.0.0:3306, control FIFO: /var/run/glbd.fifo
Jan 24 12:49:54 galera-lb-server glb[12112]: Control  address:  127.0.0.1:4444
Jan 24 12:49:54 galera-lb-server glb[12112]: Number of threads: 3, max conn: 493, nodelay: ON, keepalive: ON, defer accept: OFF, linger: OFF, daemon: YES, lat.count: 0, policy: 'round-robin', top: NO, verbose: NO
Jan 24 12:49:54 galera-lb-server glb[12112]: Destinations: 3
Jan 24 12:49:54 galera-lb-server glb[12112]:    0:   node1_ip:3306 , w: 1.000
Jan 24 12:49:54 galera-lb-server glb[12112]:    1:  node2_ip:3306 , w: 1.000
Jan 24 12:49:54 galera-lb-server glb[12112]:    2:  node3_ip:3306 , w: 1.000
Jan 24 12:49:54 galera-lb-server glb[12112]: glb v1.0.1 (epoll)
Jan 24 12:49:54 galera-lb-server glb[12112]: Incoming address: 0.0.0.0:3306, control FIFO: /var/run/glbd.fifo
Jan 24 12:49:54 galera-lb-server glb[12112]: Control  address:  127.0.0.1:4444
Jan 24 12:49:54 galera-lb-server glb[12112]: Number of threads: 3, max conn: 493, nodelay: ON, keepalive: ON, defer accept: OFF, linger: OFF, daemon: YES, lat.count: 0, policy: 'round-robin', top: NO, verbose: NO
Jan 24 12:49:54 galera-lb-server glb[12112]: Destinations: 3
Jan 24 12:49:54 galera-lb-server glb[12112]:    0:   node1_ip:3306 , w: 1.000
Jan 24 12:49:54 galera-lb-server glb[12112]:    1:  node2_ip:3306 , w: 1.000
Jan 24 12:49:54 galera-lb-server glb[12112]:    2:  node2_ip:3306 , w: 1.000
Jan 24 12:49:54 galera-lb-server glbd[12133]: glb_main.c:194: Started.
Jan 24 12:49:54 galera-lb-server glb[12112]: [Thu Jan 24 12:49:54 UTC 2019] glbd: started, pid=12133
Jan 24 12:49:54 galera-lb-server systemd[1]: Started LSB: run glbd daemon.

但是我们注意到日志中的错误如下：

Jan 24 13:36:58 galera-lb-server glbd[12133]: glb_pool.c:687: Async connection to node1_ip:3306 failed: 111 (Connection refused)
Jan 24 13:36:58 galera-lb-server glbd[12133]: glb_pool.c:699: Reconnecting to node1_ip:3306

这是service glb getinfo的输出

root@galera-lb-server:/# service glb getinfo
Router:
------------------------------------------------------
        Address       :   weight   usage    map  conns
 node1_ip:3306  :    1.000   0.000    N/A      0
 node2_ip:3306  :    1.000   0.000    N/A      0
 node3_ip:3306  :    1.000   0.000    N/A      0
------------------------------------------------------
Destinations: 3, total connections: 0 of 493 max

我们可以通过复制直接将不同的网站组连接到任何一个节点，但这不是正确的方法。请告诉我们如何部署，如果我们在这里漏掉了什么，请告诉我们。

Answer 1

我们在所有节点中绑定了IP地址，这对于所有节点来说都是一样的。

[mysqld]
bind-address = IP

Granted privileges适用于所有MariaDB to load balancer IP用户，并更改了所有网站的连接。设置正确的MariaDB用户访问权限是愚蠢的错误。答案似乎很愚蠢，但希望能对其他用户有所帮助。