Day39 content summary

Process object properties and their methods

from multiprocessing import Process,current_process
import os


current_process().pid  #  View the current process number 
os.getpid()  #  View the current process number 
os.getppid()  #  View the parent process number of the current process 
"""
windows Terminal command 
		tasklist
		tasklist |findstr PID
mac Terminal command 
		ps aux
		ps aux|grep PID
"""

Daemon

"""
 After the daemon is finished, the daemon immediately ends 
"""
#  How to open    stay start Just write the following code before the statement 
p.daemon = True
p.start()

  The mutex

"""
 When multiple people operate the same data, the data will be disordered 
 We usually lock the above problems 

 effect :
	 Turn concurrency into serial , It sacrifices the efficiency of program operation but ensures the security of data 

 Be careful :
	 Only lock the operation data 
	 Try not to handle the lock by yourself   It is easy to cause deadlock 

 Expand ： Row lock   Table locks 
	 No one else can operate a row of data in the table 
	 No one else can operate a table 
"""
from multiprocessing import Lock
mutex = Lock()
#  Grab the lock 
mutex.acquire()
#  Release the lock 
mutex.release()

#  Simulated ticket grabbing 

queue Queue

 

• """
   queue : fifo 
   Stack : First in, then out 
  """
  from multiprocessing import Queue
  q = Queue()  #  Numbers can be placed in parentheses to limit the size of the queue 
  q.put()  #  Put the data    When the queue is full   Blocking 
  q.get()  #  Take the data    Wait until the queue is empty   Blocking 
  
  q.full()  #  Determine if the queue is full 
  q.empty()  #  Determine whether the queue is empty 
  q.get_nowait()  #  If there is no data when retrieving data, an error will be reported directly 
  q.get(timeout=5)  #  When retrieving data, if there is no data, etc 5s If not, it will directly report an error 
  

• Interprocess communication

  """
   There is no direct data interaction between processes , But data interaction can be realized through queues or pipelines 
  	 The Conduit :
  	 queue : The Conduit + lock 
  
   It is only possible to use it for local testing Queue, The actual production uses very powerful tools encapsulated by others 
  redis
  kafka
  RQ
  """
  

• The producer and consumer model

  """
   producer  +  Message queue  +  consumer 
   Why is there a message queue   To solve the problem of imbalance between supply and demand 
  """
  # JoinableQueue
  """
   Can be waited for q
   When you put data in the queue   There is a counter inside that automatically adds 1
   When you get data from a queue   call task_done()   The internal timer automatically decreases 1
  q.join()   When the counter is 0 And then continue to run 
  """
  

• Thread theory

  """
   process : Resource units 
   Threads : Executive unit 
   Threads are the real workers , The resources needed in the process of working are provided by the process where the thread is located 
  
   Each process must have its own thread  
  
   Multiple threads can be created in the same process 
  """
  
  """
   Open process 
  	 Apply for memory space   
  	” Copy code “
  	 It consumes a lot of resources 
   Start thread 
  	 Create multiple threads in the same process   These two operations are not required , The consumption of resources is relatively small 
  	
  
   Intelligence quotient (IQ) 
   EQ 
   Search business 
  """