Creating multiple SSH connections at a time using Paramiko

Question

The code below runs grep in one machine through SSH and prints the results:

import sys, os, string
import paramiko

cmd = "grep -h 'king' /opt/data/horror_20100810*"

ssh = paramiko.SSHClient()
ssh.set_missing_host_key_policy(paramiko.AutoAddPolicy())
ssh.connect('10.10.3.10', username='xy', password='xy')
stdin, stdout, stderr = ssh.exec_command(cmd)
stdin.write('xy\n')
stdin.flush()

print stdout.readlines()

How can I grep five machines all at once (so that I don't have major delay), than put all that in five variables and print them all out.

Answer 1

You'll need to put the calls into separate threads (or processes, but that would be overkill) which in turn requires the code to be in a function (which is a good idea anyway: don't have substantial code at a module's top level).

For example:

import sys, os, string, threading
import paramiko

cmd = "grep -h 'king' /opt/data/horror_20100810*"

outlock = threading.Lock()

def workon(host):

    ssh = paramiko.SSHClient()
    ssh.set_missing_host_key_policy(paramiko.AutoAddPolicy())
    ssh.connect(host, username='xy', password='xy')
    stdin, stdout, stderr = ssh.exec_command(cmd)
    stdin.write('xy\n')
    stdin.flush()

    with outlock:
        print stdout.readlines()

def main():
    hosts = ['10.10.3.10', '10.10.4.12', '10.10.2.15', ] # etc
    threads = []
    for h in hosts:
        t = threading.Thread(target=workon, args=(h,))
        t.start()
        threads.append(t)
    for t in threads:
        t.join()

main()

If you had many more than five hosts, I would recommend using instead a "thread pool" architecture and a queue of work units. But, for just five, it's simpler to stick to the "dedicated thread" model (especially since there is no thread pool in the standard library, so you'd need a third party package like threadpool... or a lot of subtle custom code of your own of course;-).

Answer 2

Just run everything in a for loop, and don't forget to close stdin before moving on to next iteration. That is, after line stdin.flush() add line stdin.close()