1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
|
# Splitops
================
Splitops is a script to allow execution of commands only after the approval of
multiple users.
It is intended to be used with OpenSSH by specifiying it as the "command" option
for authorized users in `~/.ssh/authorized_keys` (explained below).
Server requirements:
=====================
GNU/linux server.
The Splitops program is written in the Python programming language, so you will need to have installed the
python3 package in your server, please check in your terminal by typing "python3 -V", to see if you have this package already installed.
Please do before using the program
=====================================
1) From your local computer git clone the deployment.git repository, and within the "splitops" folder,
copy the file "splitops" from your computer to the remote server (path:/usr/local/bin/). There is no need for you
to clone the whole deployment.git repository from the remote server to grab this program.
You can use the command "scp" for this.
scp /home/user/deployment/splitops/splitops root@server:/usr/local/bin/
2) In your remote server SSH configuration (/etc/ssh/sshd_config),
please make sure you have the option "PubkeyAuthentication yes" is
de-commented. If not, decomment the line, and --reload your ssh service.
3) In your /root/.ssh/authorized_keys, please add the next lines:
command="/usr/local/bin/splitops alice" [... key of alice ...]
command="/usr/local/bin/splitops bob" [... key of bob ...]
(one for each user, that you want to have approval from for each command request)
"Alice" and "Bob" are just usernames, they don't need to be real user system accounts in the remote server.
The way the splitops program works, is by associating a public SSH key to a username, but this username doesn't need to match,
with the real username you are using for a specific SSH public key.
Having these 2 lines in the .ssh/authorized_key files, will *force* --anyone trying to login in as root through SSH,
to execute the program: "splitops", so any other user not listed in this root/.ssh/authorized_key file, won't be able to do anything.
4) Logout from the remote server, and from your client machine, in order to start using the program Splitops,
try to "request" your very first command. This is done by using the splitops sub-command "propose".
e.g: ssh root@server propose rm -rf /opt/something
After executing this, you will have some sort of answer from the server such as next:
- authenticated as: bob
- requested command: ['rm', '-rf', '/opt/something']
- assigned id: ccafbd
That means, your new command request, is waiting for approval.
Now, you (if you are either Alice or Bob), you can "approve" this command request, and afterwards wait for
others to do the same thing (until the approval is complete for the number of users listed on the .ssh/authorized keys of the remote server).
Let's say you are Bob for now,
---------------------------------
bob@computer:~$ root@server get # To see if you have anything pending to approve
And you will see a list of pending requests. Now get the the ID of any request you want to approve,
let's say "ccafbd", then type:
bob@computer:~$ root@server approve ccafbd
Now you have to ask Alice, to approve the same request "ccafbd"
Alice will do,
alice@computer:~$ root@server approve ccafbd
And finally you will be able to --remotely execute "as root", that specific and --approved command:
bob@computer:~$ ssh root@server run ccafbd
Summary
========
Once the production remote server is up an running using the "splitops" command, a routine for you as "double-check systems administrator"
would be to use the next splitops sub-commands (get, propose,approve, discargd, run...)
alice@computer:~$ root@server get # To see if you any pending commands to approve
alice@computer:~$ root@server approve "some ID" # To approve a specific command
alice@computer:~$ root@server run "some ID" # To run, a specific approved command
alice@computer:~$ root@server propose cp /opt/something /usr/local/something # (and wait)
e.g 1) IF you try to "run" a command, but you don't have enough approvals, you will receive some output as this:
- authenticated as: --user=javier
- running command with ID bcb6a5
- not enough approvals, got 1 but need 2
e.g 2) If you try to connect to the remote server as --root, and without using any Splitops sub-command, you will receive the
next answer from the server:
alice@computer:~$ ssh root@server
- authenticated as: --user=alice
- no command provided, try help
- Connection to 78.141.227.64 closed.
Full command list
===================
CMDID= The hash assigned to each command request.
whoami: Check authentication
propose CMD ...: Propose a new command
get: Get the currently proposed command
approve CMDID: Approve a command
run CMDID: Run a sufficiently approved command
discard: Discard the currently proposed command
Please remember you have to use this sub-command remotely from your computer, towards the remote server, by using the "root" user.
Or with the user of your choice, who has on its .ssh/authorized_key file, the users list, with the OpenSSH "command= ..." option.
[*]- Also take into account, if any other user "propose" a new command, the last "proposed" command will be discarded. The program doesn't use a stack
to store all requests, it works just with 1 single request. So even if you have been approved by several users, a specific CMDID, you won't be able
to run it, if a new command --proposal has been requested.
[*] We encourage all users interested in using the Splitops command to use it first with a --normal user, and not with the root user.
So this way you can learn how the program behaves, without losing the connection with your server as the root user.
|
