[1. Kmonitor拆包安装部署. 3](#1. Kmonitor拆包安装部署. 3)

[1.1.环境准备. 3](#1.1.环境准备. 3)

[1.2.拷贝并解压. 4](#1.2.拷贝并解压. 4)

[1.3.kadb_exporter 4](#1.3.kadb_exporter 4)

[1.3.1 修改application.yml文件. 4](#1.3.1 修改application.yml文件. 4)

[1.3.2 修改连接池. 5](#1.3.2 修改连接池. 5)

[1.3.3 修改启动文件(可选) 6](#1.3.3 修改启动文件(可选) 6)

[1.4.H2库. 7](#1.4.H2库. 7)

[1.4.1进入h2db并修改启动文件(可选) 7](#1.4.1进入h2db并修改启动文件(可选) 7)

[1.4.2 打开h2库网址并连接. 8](#1.4.2 打开h2库网址并连接. 8)

[1.4.3 启动kadb_exporter 8](#1.4.3 启动kadb_exporter 8)

[1.5.node_exporter 8](#1.5.node_exporter 8)

[1.6.Prometheus 10](#1.6.Prometheus 10)

[1.6.1 node_conf 10](#1.6.1 node_conf 10)

[1.6.2 修改prometheus.yml配置文件并启动. 14](#1.6.2 修改prometheus.yml配置文件并启动. 14)

[1.6.3 打开prometheus网址并查看探针状态. 16](#1.6.3 打开prometheus网址并查看探针状态. 16)

[1.7.Grafana 16](#1.7.Grafana 16)

[1.7.1 插件所在目录（可选）. 16](#1.7.1 插件所在目录（可选）. 16)

[1.7.2 启动grafana面板. 17](#1.7.2 启动grafana面板. 17)

[1.7.3 打开grafana面板并查看状态. 17](#1.7.3 打开grafana面板并查看状态. 17)

[2. 注意事项. 20](#2. 注意事项. 20)

[2.1.h2库. 20](#2.1.h2库. 20)

[2.2.node_exporter 20](#2.2.node_exporter 20)

[2.3.kadb_exporter 20](#2.3.kadb_exporter 20)

[2.4.KADB 20](#2.4.KADB 20)

[2.5.grafana面板参数. 20](#2.5.grafana面板参数. 20)

Kmonitor拆包安装部署
环境准备

操作系统：centos7+

集群主机名称和IP地址对应关系：

|---------------|----------|--------------|----------|-----------|
| IP地址（内网） | 内网网卡 | IP地址(外网） | 外网网卡 | 主机名称 |
| 172.18.35.208 | bondib0 | 10.1.35.208 | bond0 | dwabamg01 |
| 172.18.35.209 | bondib0 | 10.1.35.209 | bond0 | dwabamg02 |
| 172.18.35.211 | bondib0 | 10.1.35.211 | bond0 | dwabasg01 |
| 172.18.35.212 | bondib0 | 10.1.35.212 | bond0 | dwabasg02 |
| 172.18.35.213 | bondib0 | 10.1.35.213 | bond0 | dwabasg03 |
| 172.18.35.214 | bondib0 | 10.1.35.214 | bond0 | dwabasg04 |
| 172.18.35.215 | bondib0 | 10.1.35.215 | bond0 | dwabasg05 |
| 172.18.35.216 | bondib0 | 10.1.35.216 | bond0 | dwabasg06 |
| 172.18.35.217 | bondib0 | 10.1.35.217 | bond0 | dwabasg07 |
| 172.18.35.218 | bondib0 | 10.1.35.218 | bond0 | dwabasg08 |
| 浮动IP：10.1.35.210 |||||
| 网关地址：10.1.35.254 |||||

数据库：postgres

集群用户名：xinjiang

密码：统一为123456，如需更改，需要修改编译之后的密码

在集群所有节点上创建操作系统用户：xinjiang，使用root用户在下面的机器上执行：

|-----------|------------------------------|
| 主机名称 | 执行操作 |
| dwabamg01 | useradd -g mppadmin xinjiang |
| dwabamg02 | useradd -g mppadmin xinjiang |
| dwabasg01 | useradd -g mppadmin xinjiang |
| dwabasg02 | useradd -g mppadmin xinjiang |
| dwabasg03 | useradd -g mppadmin xinjiang |
| dwabasg04 | useradd -g mppadmin xinjiang |
| dwabasg05 | useradd -g mppadmin xinjiang |
| dwabasg06 | useradd -g mppadmin xinjiang |
| dwabasg07 | useradd -g mppadmin xinjiang |
| dwabasg08 | useradd -g mppadmin xinjiang |

拷贝并解压

拷贝centos7_amd64.tar.gz到集群节点10.1.35.209的用户xinjiang的home目录下并解压，必须在监控的集群用户下，root用户执行
su - xinjiang

tar -xvf centos7_amd64.tar.gz

kadb_exporter

1.3.1 修改application.yml文件

使用xinjiang用户编辑/home/xinjiang/centos7_amd64/kadb_exporter文件
[xinjiang@dwabamg02 kadb_exporter]$ pwd

/home/xinjiang/centos7_amd64/kadb_exporter

[xinjiang@dwabamg02 kadb_exporter]$ vi application.yml

spring:

profiles:

active: development

server:

port: 10000 #kadb_exporter端口号

spring:

profiles: development

datasource:

url:jdbc:h2:tcp://10.1.35.209:10002//home/xinjiang/centos7_amd64/h2db/data/operator

#h2库url串

username: root

password: ENC(rVBkqsNjKhfSrkZazEoIQzMUlEejr6qNfP6U8m66JS2nSupFuAJpeRReWH_w_y39eGI6pZwbKenptjFxD4KJiuTIncrK2h3mBCiaTOTQHKX32rXD6NrW1gmGSVmE0blBXOdLZZkEfTVEWOAHR8IdldVCkK8anzOEC7em68qJZ98) #这里的密码是h2库连接的root密码是123456

hikari:

pool-name: default

connection-test-query: select current_timestamp;

minimum-idle: 2

maximum-pool-size: 10

1.3.2 修改连接池

使用xinjiang用户编辑/home/xinjiang/centos7_amd64/kadb_exporter/conf文件
[xinjiang@dwabamg02 conf]$ pwd

/home/xinjiang/centos7_amd64/kadb_exporter/conf

[xinjiang@dwabamg02 conf]$ vi jdbc_pool_default.xml

<?xml version="1.0" encoding="UTF-8"?>

<!DOCTYPE properties SYSTEM "http://java.sun.com/dtd/properties.dtd">

<comment>动态数据源池中默认JDBC连接池配置参数</comment>

#数据库访问url，格式：jdbc：postgresql://ip:port/db_name

gzkadb@sx0sxrf

#数据库用户密码为加密密码123456，获取加密密码使用的脚本为：

#./pass_enc.sh --prikey conf/conf_encrypt.pri --dbpass $PASSWORD

</properties>

1.3.3 修改启动文件(可选)

如果限制kadb_exporter进程使用的内存为2.5GB，执行下面的操作

使用xinjiang用户编辑/home/xinjiang/centos7_amd64/kadb_exporter/start.sh文件
[xinjiang@dwabamg02 kadb_exporter]$ vi start.sh

#!/bin/bash

source ~/.bashrc

if [ ! -d logs/prometheus ]; then

mkdir -p logs/prometheus

Start PrometheusExporter log.

nohup python PrometheusExporter.py > /dev/null 2>&1 &

Start KADB monitor.

nohup java -Xmx2048m -Xms2048m -cp .:./lib/*:./conf/* cn.com.kingbase.kmonitor.kadb.KMonitor &

#内存限制为一个kadb_exporter2.5g的话，添加-Xmx2048m -Xms2048m限制

修改完毕source ~/.bashrc刷新

H2库

1.4.1进入h2db并修改启动文件(可选)

如果限制h2db进程使用的内存，执行下面的操作，加入-Xmx2048m -Xms2048m，否则无需修改

使用xinjiang用户编辑/home/xinjiang/centos7_amd64/h2db/start.sh文件
[xinjiang@dwabamg02 h2db]$ pwd

/home/xinjiang/centos7_amd64/h2db

[xinjiang@mpp170 h2db]$ vi start.sh

#!/bin/sh

dir=$(dirname "$0")

nohup java -cp " $dir/kadb_h2.jar:$ H2DRIVERS: $CLASSPATH" org.h2.tools.Server -ifNotExists -tcpAllowOthers -webAllowOthers -webPort 10001 -tcpPort 10002 "$ @" &

#10002为h2库tcp连接端口

[xinjiang@dwabamg02 h2db]$ sh start.sh

[xinjiang@dwabamg02 h2db]$ cat nohup.out

TCP server running at tcp://192.168.0.30:10002 (others can connect)

PG server running at pg://192.168.0.30:5435 (only local connections)

Web Console server running at http://192.168.0.30:10001 (others can connect)

#http://192.168.0.30:10001为进入h2库的网址

1.4.2 打开h2库网址并连接

#账号/密码：root/123456

1.4.3 启动kadb_exporter

[xinjiang@dwabamg02 kadb_exporter]$ pwd

/home/xinjiang/centos7_amd64/kadb_exporter

[xinjiang@mpp170 kadb_exporter]$ sh start.sh

node_exporter

分别在各个节点下xinjiang用户启动node_exporter
[xinjiang@dwabamg02 centos7_amd64]$ pwd

/home/xinjiang/centos7_amd64

******将node_exporter执行程序拷贝到集群每个节点的/home/xinjiang/目录下*********

[xinjiang@dwabamg02 centos7_amd64]$ scp node_exporter/ xinjiang@ dwabamg01:/home/xinjiang/

[xinjiang@dwabamg02 centos7_amd64]$ scp node_exporter/ xinjiang@ dwabasg01:/home/xinjiang/

[xinjiang@dwabamg02 centos7_amd64]$ scp node_exporter/ xinjiang@ dwabasg02:/home/xinjiang/

[xinjiang@dwabamg02 centos7_amd64]$ scp node_exporter/ xinjiang@ dwabasg03:/home/xinjiang/

[xinjiang@dwabamg02 centos7_amd64]$ scp node_exporter/ xinjiang@ dwabasg04:/home/xinjiang/

[xinjiang@dwabamg02 centos7_amd64]$ scp node_exporter/ xinjiang@ dwabasg05:/home/xinjiang/

[xinjiang@dwabamg02 centos7_amd64]$ scp node_exporter/ xinjiang@ dwabasg06:/home/xinjiang/

[xinjiang@dwabamg02 centos7_amd64]$ scp node_exporter/ xinjiang@ dwabasg07:/home/xinjiang/

[xinjiang@dwabamg02 centos7_amd64]$ scp node_exporter/ xinjiang@ dwabasg08:/home/xinjiang/

************分别登录到集群的每个节点，启动node_exporter******************

******** dwabamg01启动node_exporter**************

[xinjiang@dwabamg02 node_exporter]$ ssh dwabamg01

[xinjiang@dwabamg01 centos7_amd64]$ cd /home/xinjiang/node_exporter/

[xinjiang@dwabamg01 node_exporter]$ sh start.sh

[xinjiang@dwabamg01 node_exporter]$ exit

******** dwabasg01启动node_exporter**************

[xinjiang@dwabamg02 node_exporter]$ ssh dwabasg01

[xinjiang@dwabasg01 centos7_amd64]$ cd /home/xinjiang/node_exporter/

[xinjiang@dwabasg01 node_exporter]$ sh start.sh

[xinjiang@dwabasg01 node_exporter]$ exit

******** dwabasg02启动node_exporter**************

[xinjiang@dwabamg02 node_exporter]$ ssh dwabasg02

[xinjiang@dwabasg02 centos7_amd64]$ cd /home/xinjiang/node_exporter/

[xinjiang@dwabasg02 node_exporter]$ sh start.sh

[xinjiang@dwabasg02 node_exporter]$ exit

******** dwabasg03启动node_exporter**************

[xinjiang@dwabamg02 node_exporter]$ ssh dwabasg03

[xinjiang@dwabasg03 centos7_amd64]$ cd /home/xinjiang/node_exporter/

[xinjiang@dwabasg03 node_exporter]$ sh start.sh

[xinjiang@dwabasg03 node_exporter]$ exit

******** dwabasg04启动node_exporter**************

[xinjiang@dwabamg02 node_exporter]$ ssh dwabasg04

[xinjiang@dwabasg04 centos7_amd64]$ cd /home/xinjiang/node_exporter/

[xinjiang@dwabasg04 node_exporter]$ sh start.sh

[xinjiang@dwabasg04 node_exporter]$ exit

******** dwabasg05启动node_exporter**************

[xinjiang@dwabamg02 node_exporter]$ ssh dwabasg05

[xinjiang@dwabasg05 centos7_amd64]$ cd /home/xinjiang/node_exporter/

[xinjiang@dwabasg05 node_exporter]$ sh start.sh

[xinjiang@dwabasg05 node_exporter]$ exit

******** dwabasg06启动node_exporter**************

[xinjiang@dwabamg02 node_exporter]$ ssh dwabasg06

[xinjiang@dwabasg06 centos7_amd64]$ cd /home/xinjiang/node_exporter/

[xinjiang@dwabasg06 node_exporter]$ sh start.sh

[xinjiang@dwabasg06 node_exporter]$ exit

******** dwabasg07启动node_exporter**************

[xinjiang@dwabamg02 node_exporter]$ ssh dwabasg07

[xinjiang@dwabasg07 centos7_amd64]$ cd /home/xinjiang/node_exporter/

[xinjiang@dwabasg07 node_exporter]$ sh start.sh

[xinjiang@dwabasg07 node_exporter]$ exit

******** dwabasg08启动node_exporter**************

[xinjiang@dwabamg02 node_exporter]$ ssh dwabasg08

[xinjiang@dwabasg08 centos7_amd64]$ cd /home/xinjiang/node_exporter/

[xinjiang@dwabasg08 node_exporter]$ sh start.sh

[xinjiang@dwabasg08 node_exporter]$ exit

Prometheus

1.6.1 node_conf

使用用户xinjiang新建并进入node_conf目录，创建node_kadb_info.json文件
[xinjiang@dwabamg02 prometheus]$ pwd

/home/xinjiang/centos7_amd64/prometheus

[xinjiang@dwabamg02 prometheus]$ mkdir -p node_conf

[xinjiang@dwabamg02 prometheus]$ cd node_conf/

[xinjiang@dwabamg02 node_conf]$ vim node_kadb_info.json

[

{

"labels": {

"desc": "集群standby节点",

"project": "新疆农信",

"system": "xinjiang",

"instance": "10.1.35.209",

"hostname": "dwabamg02",

"cluster": "kadb_cluster",

"service": "node_exporter"

"targets": [

"10.1.35.209:10003"

]

{

"labels": {

"desc": "集群master节点",

"project": "新疆农信",

"system": " xinjiang ",

"instance": "10.1.35.208",

"hostname": " dwabamg01",

"cluster": "kadb_cluster",

"service": "node_exporter"

"targets": [

"10.1.35.208:10003"

]

{

"labels": {

"desc": "集群计算节点1",

"project": "新疆农信",

"system": " xinjiang ",

"instance": "10.1.35.211",

"hostname": "dwabasg01",

"cluster": "kadb_cluster",

"service": "node_exporter"

"targets": [

"10.1.35.211:10003"

]

{

"labels": {

"desc": "集群计算节点2",

"project": "新疆农信",

"system": " xinjiang ",

"instance": "10.1.35.212",

"hostname": "dwabasg02",

"cluster": "kadb_cluster",

"service": "node_exporter"

"targets": [

"10.1.35.212:10003"

]

{

"labels": {

"desc": "集群计算节点3",

"project": "新疆农信",

"system": " xinjiang ",

"instance": "10.1.35.212",

"hostname": "dwabasg03",

"cluster": "kadb_cluster",

"service": "node_exporter"

"targets": [

"10.1.35.213:10003"

]

{

"labels": {

"desc": "集群计算节点4",

"project": "新疆农信",

"system": " xinjiang ",

"instance": "10.1.35.214",

"hostname": "dwabasg04",

"cluster": "kadb_cluster",

"service": "node_exporter"

"targets": [

"10.1.35.214:10003"

]

{

"labels": {

"desc": "集群计算节点5",

"project": "新疆农信",

"system": " xinjiang ",

"instance": "10.1.35.215",

"hostname": "dwabasg05",

"cluster": "kadb_cluster",

"service": "node_exporter"

"targets": [

"10.1.35.215:10003"

]

{

"labels": {

"desc": "集群计算节点6",

"project": "新疆农信",

"system": " xinjiang ",

"instance": "10.1.35.216",

"hostname": "dwabasg06",

"cluster": "kadb_cluster",

"service": "node_exporter"

"targets": [

"10.1.35.216:10003"

]

{

"labels": {

"desc": "集群计算节点7",

"project": "新疆农信",

"system": " xinjiang ",

"instance": "10.1.35.217",

"hostname": "dwabasg07",

"cluster": "kadb_cluster",

"service": "node_exporter"

"targets": [

"10.1.35.217:10003"

]

{

"labels": {

"desc": "集群计算节点8",

"project": "新疆农信",

"system": " xinjiang ",

"instance": "10.1.35.218",

"hostname": "dwabasg08",

"cluster": "kadb_cluster",

"service": "node_exporter"

"targets": [

"10.1.35.218:10003"

]

}

]

1.6.2 修改prometheus.yml配置文件并启动

[xinjiang@dwabamg02 prometheus]$ pwd

/home/xinjiang/centos7_amd64/prometheus

[xinjiang@dwabamg02 prometheus]$ vi prometheus.yml

global:

scrape_interval: 10s

evaluation_interval: 10s

scrape_configs:

job_name: 'consul'

static_configs:

targets: ['10.1.35.209:10000']

labels:

cluster: 'kadb_cluster'

service: 'kadb_exporter'

kadburl: 'http://192.168.0.30:10000'

targets: ['10.1.35.209:10003']

labels:

cluster: 'kadb_cluster'

service: 'node_exporter'

kadburl: 'http://10.1.35.209:10003'

targets: ['10.1.35.208:10003']

labels:

cluster: 'kadb_cluster'

service: 'node_exporter'

kadburl: 'http:// 10.1.35.208:10003'

targets: ['10.1.35.211:10003']

labels:

cluster: 'kadb_cluster'

service: 'node_exporter'

kadburl: 'http:// 10.1.35.211:10003'

targets: ['10.1.35.212:10003']

labels:

cluster: 'kadb_cluster'

service: 'node_exporter'

kadburl: 'http://10.1.35.212:10003'

targets: ['10.1.35.213:10003']

labels:

cluster: 'kadb_cluster'

service: 'node_exporter'

kadburl: 'http://10.1.35.213:10003'

targets: ['10.1.35.214:10003']

labels:

cluster: 'kadb_cluster'

service: 'node_exporter'

kadburl: 'http://10.1.35.214:10003'

targets: ['10.1.35.215:10003']

labels:

cluster: 'kadb_cluster'

service: 'node_exporter'

kadburl: 'http://10.1.35.215:10003'

targets: ['10.1.35.216:10003']

labels:

cluster: 'kadb_cluster'

service: 'node_exporter'

kadburl: 'http:/10.1.35.216:10003'

targets: ['10.1.35.217:10003']

labels:

cluster: 'kadb_cluster'

service: 'node_exporter'

kadburl: 'http://10.1.35.217:10003'

targets: ['10.1.35.218:10003']

labels:

cluster: 'kadb_cluster'

service: 'node_exporter'

kadburl: 'http://10.1.35.218:10003'

file_sd_configs:

files:
/home/xinjiang/centos7_amd64/prometheus/node_conf/node_kadb_info.json

[xinjiang@mpp170 prometheus]$ sh start.sh #启动prometheus

1.6.3 打开prometheus网址并查看探针状态

图例显示所以的探针处于"UP"状态为正常

Grafana

1.7.1 插件所在目录（可选）

无新增监控面板，可不做修改
[xinjiang@dwabamg02 plugins]$ pwd

/home/xinjiang/centos7_amd64/kadb_monitor/plugins

[xinjiang@dwabamg02 kadb_monitor]$ cd plugins/

[xinjiang@mpp170 plugins]$ ll

总用量 20

drwxrwxr-x. 3 xinjiang xinjiang 70 7月 5 16:01 AlertManagerPanel

drwxrwxr-x. 3 xinjiang xinjiang 38 10月 20 16:09 ClusterVersionStatusPanel

drwxrwxr-x. 3 xinjiang xinjiang 4096 7月 5 16:01 data-table-plugin

drwxrwxr-x. 3 xinjiang xinjiang 70 7月 5 16:01 InstanceStatusPanel

drwxrwxr-x. 3 xinjiang xinjiang 98 10月 9 09:44 kadb_AlertNowList_plugin

drwxrwxr-x. 3 xinjiang xinjiang 98 9月 24 11:35 kadb_ClusterInfoTable_plugin

drwxrwxr-x. 3 xinjiang xinjiang 4096 7月 5 16:01 kadb_piechart_panel

drwxrwxr-x. 3 xinjiang xinjiang 4096 10月 9 11:15 kadb_TopologyTable_plugin

drwxrwxr-x. 3 xinjiang xinjiang 4096 7月 5 16:01 selected-table-plugin

drwxrwxr-x. 3 xinjiang xinjiang 70 7月 5 16:01 SessionListPanel

drwxrwxr-x. 3 xinjiang xinjiang 4096 7月 5 16:01 topology-plugin

#集群版本和拓扑信息显示，插件为：kadb_AlertNowList_plugin、ClusterVersionStatusPanel、kadb_TopologyTable_plugin

1.7.2 启动grafana面板

[xinjiang@dwabamg02 kadb_monitor]$ pwd

/home/xinjiang/centos7_amd64/kadb_monitor

[xinjiang@mpp170 kadb_monitor]$ sh start.sh

1.7.3 打开grafana面板并查看状态

浏览器地址：http://10.1.35.208:3000

建议使用谷歌浏览器，如果打开grafana是可以的，则不需要修改参数，参数修改视情况而定

如果集群拓扑信息不能正常显示，则单击图中"下箭头"，选择"编辑"

在面板右边"可视化"菜单中，将IP地址修改为：10.1.35.209。

注意：这里的ip地址为kadb_export的地址和端口（10000），如果是内网地址，需要将kadb_export的地址映射为浏览器的外网地址后，进行修改。B端需要直接和kadb_export进行通讯

"最新警报"和"集群版本"两个监控面板也做同样的修改

如果节点资源信息信息不能正常显示，表现为监控界面打开缓慢，有如下报错，并且kadb_monitor.log日志文件有报错，不能连接192.168.0.30:10004，如图

则是kmonitor没有正确设置prometheus的地址信息，需要在上面左侧图中选择

齿轮，配置实际的prometheus地址

能curl到kadb_export:10000地址的下面信息，说明浏览器和kadb_export通讯正常

监控面板编辑页面：

节点监控页面：

主机监控页面：

注意事项

2.1.h2库

h2一定要起来，h2库网址连接正常

2.2.node_exporter

每个监控的ip下要有一个node_exporter目录并启动

2.3.kadb_exporter

一个集群要有一个kadb_exporter，在主节点上，并且一个kadb_exporter至少占实际内存2.5G

2.4.KADB

Kadb数据库要正常启动

2.6. 修改默认参数

选择这里的设置

选择这里的变量

修改：request_url为prometheus节点地址

2.7 频繁扫描日志造成磁盘I/O繁忙

修改kadb_export的配置文件：/home/mppadmin/centos7_amd64/kadb_exporter/conf/schedules.xml

将以下部分删除

日志抓取调度：

和磁盘数据分布相关的调度：

人大金仓KADB监控工具及问题处理

1.3.1 修改application.yml文件

1.3.2 修改连接池

1.3.3 修改启动文件(可选)

Start PrometheusExporter log.

nohup python PrometheusExporter.py > /dev/null 2>&1 &

Start KADB monitor.

1.4.1进入h2db并修改启动文件(可选)

1.4.2 打开h2库网址并连接

1.4.3 启动kadb_exporter

1.6.1 node_conf

1.6.2 修改prometheus.yml配置文件并启动

1.6.3 打开prometheus网址并查看探针状态

1.7.1 插件所在目录（可选）

1.7.2 启动grafana面板

1.7.3 打开grafana面板并查看状态

2.1.h2库

2.2.node_exporter

2.3.kadb_exporter

2.4.KADB

2.6. 修改默认参数

2.7 频繁扫描日志造成磁盘I/O繁忙