Archery database audit platform deployment

One 【 Introduce 】

Official website ：https://archerydms.com/

github:https://github.com/hhyo/Archery

Archery yes archer Branch project of , Located at SQL Audit query platform , Designed to improve DBA Work efficiency , Support multi database SQL Go online and check , At the same time, it supports rich MySQL Operation and maintenance functions , All functions are compatible with mobile phone operation .

Two 【 Deploy 】

download  Releases file , Decompress and enter docker-compose Folder   # start-up docker-compose -f docker-compose.yml up -d # Table structure initialization docker exec -ti archery /bin/bash cd /opt/archery source /opt/venv4archery/bin/activate python3 manage.py makemigrations sql python3 manage.py migrate # Data initialization python3 manage.py dbshell<sql/fixtures/auth_group.sql python3 manage.py dbshell<src/init_sql/mysql_slow_query_review.sql # Create administrative users python3 manage.py createsuperuser # Restart the service docker restart archery # Log viewing and troubleshooting docker logs archery -f --tail=10 logs/archery.log

 3、 ... and  【 Feature list 】

 Four  【 Plug in processing 】
4.1  Database slow log plug-in 
 The server builds the database mysql processing method :

Archery Cannot automatically collect slow queries , You need to deploy the script manually . according to github The documents on the are as follows ：

1 Execute the script to create the table , stay archery The database creates two tables about collecting slow query information ：mysql_slow_query_review and mysql_slow_query_review_history

Create database ：

route ：/opt/archery/src/init_sql/mysql_slow_query_review.sql

pt Tool script execution ：

route ：/opt/archery/src/script/analysis_slow_query.sh

Copy the script SQL, stay archery The associated mysql In the implementation of

2 Deploy to the target database, that is, the server that needs to view the slow log pt Tools .

install pt Tools

download wget https://www.percona.com/downloads/percona-toolkit/3.0.12/binary/redhat/6/x86_64/percona-toolkit-3.0.12-1.el6.x86_64.rpm install yum install -y percona-toolkit-3.0.12-1.el6.x86_64.rpm The command defaults to /usr/bin/pt-query-digest Next

 Modify the script 

#!/bin/bash DIR="$( cd "$( dirname "$0" )" && pwd )" cd ${DIR} # To configure Archery Database connection address archery_db_host="192.168.1.1" archery_db_port=3306 archery_db_user="root" archery_db_password="*****" archery_db_database="archery" # Slow log location of the analyzed instance , It is recommended to clean the log file regularly , Otherwise, the analysis efficiency will be affected slowquery_file="/data/mysql/mysql-slow.log" #pt-query-digest Executable file path pt_query_digest="/usr/bin/pt-query-digest" # Connection information of the analyzed instance hostname="192.168.1.2:3306" # Need and Archery The content in the instance configuration is consistent , For screening , Incorrect configuration will result in data not being displayed # Get the last analysis time , Please delete during initialization last_analysis_time_$hostname file , All log data can be analyzed if [[ -s last_analysis_time_${hostname} ]]; then last_analysis_time=`cat last_analysis_time_${hostname}` else last_analysis_time='1000-01-01 00:00:00' fi # Collect the logs #RDS Need to increase the --no-version-check Options ${pt_query_digest} \ --user=${archery_db_user} --password=${archery_db_password} --port=${archery_db_port} \ --review h=${archery_db_host},D=${archery_db_database},t=mysql_slow_query_review \ --history h=${archery_db_host},D=${archery_db_database},t=mysql_slow_query_review_history \ --no-report --limit=100% --charset=utf8 \ --since "$last_analysis_time" \ --filter="\$event->{Bytes} = length(\$event->{arg}) and \$event->{hostname}=\"$hostname\" and \$event->{client}=\$event->{ip} " \ ${slowquery_file} > /tmp/analysis_slow_query.log if [[ $? -ne 0 ]]; then echo "failed" else echo `date +"%Y-%m-%d %H:%M:%S"`>last_analysis_time_${hostname} fi

Adjust the timing task trigger pt Tools synchronized to archery 

crontab -e

3.rds Database synchronization

3.1 Alibaba cloud officially provides docking , Just fill in ak/sk key that will do .

3.2 Other cloud databases , For example, Hua Weiyun / Tencent, cloud, etc. ,

Method 1： You can modify the source code yourself slowlog.py The corresponding sdk, Synchronize database slow logs .

Method 2: Get the slow log of Huawei cloud database api Interface , Get slow log file , Through PT Tools synchronized to archery.