SSH known hosts cleanup script. Combines same fingerprints to one line.

ssh-knownhosts-cleanup.py

#!/usr/bin/env python3
# encoding: utf-8
# By Joonas Kuorilehto 2011, MIT license
#
# The script combines .ssh/known_hosts so that each fingerprint is only
# listed once.

import re
import sys
import os
import shutil

def main():
    if len(sys.argv) != 2:
        print("Usage: {} $HOME/.ssh/known_hosts".format(sys.argv[0]))
        sys.exit(1)
    clean_known_hosts(sys.argv[1])

def clean_known_hosts(known_hosts_file):
    # Backup known hosts file
    shutil.copyfile(known_hosts_file, known_hosts_file+".old")

    # Read known hosts to memory
    with open(known_hosts_file) as f:
        knownhosts = dict()
        oldlines = 0
        for line in f:
            if line.strip() == "" or line.strip().startswith("#"):
                continue
            oldlines += 1
            hosts, keytype, fingerprint = line.strip().split(" ")[:3]
            dictkey = keytype + fingerprint
            hosts = hosts.split(",")
            if knownhosts.get(dictkey) == None:
                knownhosts[dictkey] = dict(hosts=set(), keytype=keytype,
                                           fingerprint=fingerprint)
            knownhosts[dictkey]['hosts'].update(hosts)

    # Replace known hosts with a cleaned version
    with open(known_hosts_file, 'w') as f:
        for key, host in knownhosts.items():
            host['hosts_joined'] = ",".join(host['hosts'])
            f.write("%(hosts_joined)s %(keytype)s %(fingerprint)s\n" % host)

    print("OK. Cleaned up", known_hosts_file)
    print("Change: from %d lines to %d lines." % (oldlines, len(knownhosts)))

    print("List of host names in the file (aliases on the same line):")
    i = 0
    for key, host in knownhosts.items():
        i += 1
        def ipv6(x):
            if ':' in x:
                return True
        def hostname(x):
            if not ipv6(x) and re.match("[a-z]", x, re.I):
                return True
        hosts = list(filter(hostname, host['hosts']))
        if len(hosts):
            print("%-4d"%i + " ".join(sorted(hosts)))


if __name__ == '__main__':
    main()

https://gist.github.com/joneskoo/1306614/96dd6d75a086bebbfa34def081867447850b616c KNOWN_HOSTS is now unknown (had to revert to your last commit).

Thanks for the script (so much tidier now).