PID and PID files and Hadoop change PID file storage location under Linux system

1. Understanding PID:

PID full name is process identification.
PID is the code of the process, and each process has a unique PID number. It is randomly assigned by the process runtime and does not represent a specialized process. The PID does not change the identifier at run time, but when you terminate the program and then run the PID identifier, it will be reclaimed by the system, and it may continue to be assigned to the new running program.

2.pid file

• Content of the PID file
• Viewed with the cat command, you can see that there is only one row of content and the ID of the process is logged
• The role of PID files
• Prevent multiple copies of a process from starting
• Principle of PID File
• When the process runs, it adds a file lock to the. pid file, and only the process that obtains the lock has write access (F_WRLCK), and writes its own PID to the file.
  Other processes that attempt to obtain the lock will automatically exit. 3.Hadoop in PID file storage
• Brief introduction:
• The default configuration of the PID file for the related process after Hadoop startup is saved in the/tmp directory, using the stop-all.sh,stop-dfs.sh,stop-yarn.sh script to stop the related process is stopped by the corresponding PID file, and Linux under/tmp The directory has a mechanism for timed cleanup, so to prevent errors such as no namenode to stop when the process is stopped, we should change the location where the process PID file is stored.
• Change method:

  
  
  
  [[email protected] ~]$ mkdir -p /opt/software/hadoop-2.8.1/pids
  [[email protected] ~]$ jps
  10512 ResourceManager
  8018 Jps
  10004 SecondaryNameNode
  9655 NameNode
  9806 DataNode
  10623 NodeManager
  [[email protected] ~]$ cd /opt/software/hadoop-2.8.1/sbin/
  [[email protected] sbin]$ ./stop-dfs.sh
  [[email protected] sbin]$ ./stop-yarn.sh
  [[email protected] sbin]$ vi hadoop-daemon.sh
  HADOOP_PID_DIR=/opt/software/hadoop-2.8.1/pids
  [[email protected] sbin]$ vi yarn-daemon.sh
  YARN_PID_DIR=/opt/software/hadoop-2.8.1/pids
  [[email protected] sbin]$ ./start-dfs.sh
  [[email protected] sbin]$ ./start-yarn.sh
  [[email protected] sbin]$ jps
  9201 SecondaryNameNode
  9425 ResourceManager
  9540 NodeManager
  9828 Jps
  8852 NameNode
  8973 DataNode
  #Found that the pid is stored in a new location
  [[email protected] sbin]$ cd /opt/software/hadoop-2.8.1/pids/
  [[email protected] pids]$ ll
  total 20
  -rw-rw-r-- 1 hadoop hadoop 5 May 22 20:14 hadoop-hadoop-datanode.pid
  -rw-rw-r-- 1 hadoop hadoop 5 May 22 20:14 hadoop-hadoop-namenode.pid
  -rw-rw-r-- 1 hadoop hadoop 5 May 22 20:14 hadoop-hadoop-secondarynamenode.pid
  -rw-rw-r-- 1 hadoop hadoop 5 May 22 20:15 yarn-hadoop-nodemanager.pid
  -rw-rw-r-- 1 hadoop hadoop 5 May 22 20:15 yarn-hadoop-resourcemanager.pid
  

PID and PID files and Hadoop change PID file storage location under Linux system