Some pits in the middleware installation process

1. General

• Question 1 : An error occurred when tar unzipped the package gzip: stdin: not in gzip format
  • Solution: no need to add z parameter
  • Reference blog: Error gzip: stdin: not in gzip format when tar unzips the package

2、Nginx

• Question 1 : Nginx appears 403 forbidden

  • Solution:
    • 1. Change the user of nginx.config to be consistent with the startup user: root
    • 2. Check if the index.html file is missing, and if the root path in the configuration file is correct, replace it with an absolute path.
  • Reference blog: Four ways to solve the 403 forbidden (13: Permission denied) error in Nginx

• instruction

进入Nginx所在的配置文件文件夹：
查看Nginx的版本号：nginx -V
启动Nginx：start nginx   或者  ./nginx
快速停止或关闭Nginx：nginx -s stop
正常停止或关闭Nginx：nginx -s quit
配置文件修改重装载命令：./nginx -s reload

• Question 2 : nginx startup error "/var/run/nginx/nginx.pid" failed"
  • Solution:
    • 1. Create a log path
    • 2. Modify the log of nginx.config to the creation path.
    • 3. Restart the service
  • Reference blog:
    nginx startup error "/var/run/nginx/nginx.pid" failed"

3 、 Redis

• Question 1 : Redis exit:

  • 1. Enter first: redis-cli
  • 2. Exit again: shutdown

• Question 2 : Redis sentry startup error No such master with specified name

  • Solution: Put the monitor mymaster in the front as much as possible, the auth-pass behind it, and the other configurations behind it.
  • Reference blog: Redis sentinel starts to report an error No such master with specified name

• Question 3 : Redis sentinel mode, the master node is down, and the slave node cannot be changed to the master node

  • Solution: notice whether the redis.conf file is written correctly:
    • bind 127.0.0.1 10.134.33.45 (the actual IP of the machine)

• Question 4 : Redis sentinel mode, the master node is down, and after the slave node changes to the master node, other slave nodes think that the node is down

  • Solution: configuration file problem, see Sentinel plus VIP drift to achieve high availability of Redis cluster

3、Rabbitmq

• 问题1 : ERROR: node with name “rabbit” already running on “lhg1”

  • Solution: It is caused by the background program being occupied, check the process and delete it: kill -9
  • Reference blog: RabbitMQ reports ERROR: node with name “rabbit” already running on “lhg1”

• Question 2 : After restarting rabbitmq, the username is lost and the user cannot log in

  • Solution: Find or create a new rabbitmq-env.conf under the specified path (for example: /etc/rabbitmq), the content is NODENAME=rabbit@rabbitmqnode01

• Question 3 : Rabbitmq startup encountered a lot of error correction ideas

  • Solution:
    • 1. Confirm the permissions and path of .erlang.cookic
      • Permission is generally 777 or 400
      • The path is either /root or the installation location (this is included in the error message)
    • 2. Force reset
      • rabbitmqctl force_reset
    • 3. Configure the ip address of each node in the cluster first, and then use the instruction to join the cluster
      • abbitmqctl join_cluster rabbit@A hostname