checkitbaby 1.7

Package checkitbaby

Checkitbaby

Definition

Checkitbaby is a tool to allow automatic test validations in a lab. It uses some Agents to interact with the setup, for instance to play a simple client/server role, to change the setup topology or even to query the DUT. Agents are connected using ssh. It is recommended to use ssh keys.

Playbooks are defined as a collection of Testcases, each testcase is a simple text file where each line defines an action applied to an Agent. Each line of the testcase can either trigger an action and/or get some information and see if some requirements are met (checks).
Test scenario syntax is simple and evolutive, commands are defined keywords and depend on the type of agents targeted.
Multiple simultaneous connections to agents are supported. Variables are allowed in testcases. A variable is just a keyword encompassed with dollar signs '$' and defined in a variable file. During testcases execution, each Run information such as agent terminal ouputs are collected in log files. Test verifications are always done from log file parsing, like a human would do. With this, it is possible to easily double-check the test result post-run. Marks can be used as a delimeter for check verification within the agent log file.

Checkitbaby can be simply run as a script to run all or some testcases against the setup. It is possible to run the testcase in Dry-Run mode to only validate the scenarios file syntax for staging purpose.

Checkitbaby focus is to run against a FortiPoc setup, either from withing the PoC (from a testing lxc) or externaly to PoC (from user PC). It can however be used in other contexts.

There are multiple ways to run a playbook:

• Run the entire playbook :
  python3 run_playbook --playbook <playbook_name> [--feedback <feedback_file> --run <run_id> --debug]

• Run a single testcase from the playbook by its id :
  python3 run_playbook --playbook <playbook_name> --testcase <test_case_id> [--feedback <feedback_file> ]

• Run a playlist of testcases from the playbook, defined in conf/playlists.yml:
  python3 run_playbook --playbook <playbook_name> --playlist <playlist_id> [--feedback <feedback_file> --run <run_id> --debug]

For each possibility, it is possible to set options:

--feedback <feedback> : Feedback file expected from taskwatcher
--run : Designate the run directory where results will be stored
--dryrun : Only performs a dry-run to validate scenario syntax (no connections to agents)
--debug : Turns debugging on (debug output in file debug.log)

When all the testcases from a Playbook have run, a Report in a json format is created. The report is organized by testcases and includes all checks results from the testcase. A general Pass/Fail covering all testcases is also included.

Author

Cedric GUSTAVE

Installation

• Requirements : netcontrol python library: pip install -I netcontrol
• Package is hosted on pypi, it should be installed using pip: pip install -I checkitbaby

Usage

Command line:

Usage: python3 run_playbook.py -playbook <playbookName>

Optional settings:

    --playlist <playlist_id>
    --testcase <tescase_id>
    --run <run_id>
    --dryrun
    --debug

Examples

• Running all testcases from a playbook :
  ex : python3 run_playbook.py --playbook myPlaybook --path PlayBookPath

• Running a specific testcase from a playbook :
  ex : python3 run_playbook.py --playbook myPlaybook --testcase 003 --path PlayBookPath

• Running a specific playlist from a playbook : ex : python3 run_playbook.py --playbook myPlaybook --playlist PL01 --path PlayBookPath

Web integration:

There is currently no web integration, this is however in the pipe.
A web GUI would be provided to : load, select and run the defined testcases and provide a formated report.

Organization

File tree structure

The following directory tree structure is used to organize the tests :

• Directory structure:

/PLAYBOOK_BASE_PATH : The base name of the playbook location
	  ex : /fortipoc/playbooks

/PLAYBOOK_BASE_PATH/ANY_PLAYBOOK_NAME
	  ex : /fortipoc/playbooks/advpn

/PLAYBOOK_BASE_PATH/ANY_PLAYBOOK_NAME/conf/agents.yml: files with agents definitions
	  ex : /fortipoc/playbooks/advpn/conf/agents.yml

/PLAYBOOK_BASE_PATH/ANY_PLAYBOOK_NAME/conf/variables.yml : files with variables definitions
	  ex : /fortipoc/playbooks/advpn/conf/variables.yml

/PLAYBOOK_BASE_PATH/ANY_PLAYBOOK_NAME/conf/macros.txt : files with macro definitions
	  ex : /fortipoc/playbooks/advpn/conf/macros.txt

/PLAYBOOK_BASE_PATH/ANY_PLAYBOOK_NAME/testcases : directory containing testcases
	  ex : /fortipoc/playbooks/advpn/testcases

/PLAYBOOK_BASE_PATH/ANY_PLAYBOOK_NAME/testcases/NNN_TESTCASE_NAME : one testcase file  
     with NNN : a number starting from 000 (to order testcases)
	 TESTCASE_NAME : any name for the testcase

      ex : /fortipoc/playbooks/advpn/testcases/001_spoke_to_hub_connectivity.txt
      ex : /fortipoc/playbooks/advpn/testcases/002_spoke_ipsec_tunnel.txt
      ex : /fortipoc/playbooks/advpn/testcases/003_spoke_routing.txt

• Creating a new playbook:

Use create_new_playbook.sh to create a new playbook file tree:
Usage : ./create_new_playbook.sh <playbook_name>

Agents

Currently supported agents are :

• Debian LXC (aka Linux host)
• Vyos routers
• FortiGate devices
• FortiPoC VMs

A few agent-less functions are defined (for instance to wait or append some comments or standard marks in the logs)

The generic syntax of each line of a testcase is as follow :
<AGENT_NAME>:<AGENT_CONNECTION_ID> <COMMAND> <COMMAND_SPECIFIC_DATA>

Each test/validation uses command 'check' followed by the test reference between square bracket [TEST_NAME].
The TEST_NAME should be unique in the testcase file.
A check may include a single or a list of Requirements. Requirements follow keyword 'has', they are provided as key=value pairs separated by spaces. A test pass if all provided requirements are met. If no requirements are provided, the test check would be succefull if an occurence was found.

Each line starting with comment sign '#' are ignored.
Lines are ran sequentially.

SSH connections to agents are automatically opened and closed at the end of the testcase.

The following chapter defines each agent command syntax and support.

Generic commands

Following commands are not agent specific and can be used with all agents

message "my message"

# Append a message on the user output when running the testcase
# This message is not append on the agent log file.
message "set Branch 1 connections delays and losses"

mark "mark_id"

# Appends a mark on the agent log file (but not on user output)
# This should be used to delimit checks parsing start (see check command 'since')
HOSTS-B2:1 mark "receive_ready"

skip all

# Skip all following lines from the testcase
skip all

wait (in seconds)

# Wait a number of seconds
wait 30

Debian LXC

ping test

optional maxloss and maxdelay requirements

# Ping test, pass if at least one packet is not lost
# Delay and loss are added in the report
LXC1-1:1 ping [con_test] 10.0.2.1

# Ping test, pass if maximum packet loss under 50 %
LXC1-1:1 ping [con_test] 10.0.2.1 maxloss 50

# Ping test, pass if delay is < 10 ms
# Uses variables $google_dns$ defined in conf/variables.yml

clt-1:2 ping [fail_delay_conntest] $google_dns$ maxdelay 10

connection test

Connection (UDP or TCP) one way or two-way test.
Open, connect, close connections and send data. It is recommended to use 'marks' to limit the check parsing area.

# Open a tcp server on port 8000  on agent LXC-1 from its connection 1
LXC-1:1 open tcp 8000

# Append a mark in LXC-1:1 log file
LXC-1:1 mark "server ready"

# Connect to a tcp server at ip 10.0.2.1 on port 8000
LXC-2:1 connect tcp 10.0.2.1 8000

# Send data string 'alice' on tcp connection from client side
LXC-2:1 data send "alice"

# Check data 'alice' is received server. Test is called '1_traffic_origin_direction'
# Parsing on server log file starts at mark "server ready"
LXC-1:1 check [1_traffic_origin_direction] data receive "alice" since "server ready"

# Append a mark "client ready" on client log file
LXC-2:1 mark "client ready"

# Send data string 'bob' on tcp connection from server side:
LXC-1:1 data send "bob"

# Check data 'bob' is received on client. Test is called '2_traffic_reply_direction'
# Search scope on client log file start at mark "client ready"
LXC-2:1 check [2_traffic_reply_direction] data receive "bob" since "client ready"

# Close tcp socket from client side:
LXC-1:1 close tcp

Vyos

Interact with Vyos routers. Does not generate tests results in reports.

traffic-policy

Changes the defined traffic-policy values

# Change vyos device R1-B1 traffic-policy named 'WAN' settings 
R1-B1:1 traffic-policy WAN delay 10 loss 0

FortiGate

Interact with FortiGate device, generates test results and retrieve information added to the report.
Vdoms are supported.

Status

# Check that FGT-B1 VM license is Valid
FGT-B1-1:1 check [FGT-B1_license] status has license=True

# Get FortiGate firmware version and VM license status
# Added in the reports as respectively 'version' and 'license'
FGT-B1-1:1 get status

vdom support

When a command need to be run inside a vdom, add vdom=VDOM_NAME just after the 1st command keyword.
Some examples :

• to check vdom 'customer' has 12 bgp routes with 8 of them recursive routes:
  F1B2:1 check [bgp_routes] bgp vdom=customer has total=12 recursive=8

• to check ssh session exist in vdom customer:
  F1B2:1 check [ssh_session_exists] session vdom=customer filter dport=22

• to check SDWAN member is alive in SDWAN Rule 1 on vdom 'customer':
  check [sdwan_service] sdwan vdom=customer service 1 member 1 has status=alive

Sessions

Checks on FortiGate session table. This command has a first 'filter' section to select the sessions. An implicit 'diag sys session filter clear' is done before the command. Allowed keywords are :
['vd','sintf','dintf','src','nsrc','dst','proto','sport','nport','dport','policy','expire','duration','proto-state','session-state1','session-state2','ext-src','ext-dst','ext-src-negate','ext-dst-negate','negate']. Multiple selectors can be used if separated with space.

Supported requirements : 'state', 'src','dest','sport','dport','proto','proto_state','duration','expire','timeout','dev','gwy','total' (number of sessions)

# Checks that a least a session with destination port 9000 exists
FGT-B1-1:1 check [session_tcp9000] session filter dport=9000

# Checks that a least a session with dport 22 and dest ip 192.168.0.1 exists
FGT-B1-1 check [ssh_session_exist] session filter dport=22 dest=192.168.0.1

# Checks that session with destination port 5000 has dirty flag set
FGT-B1-1 check [session_is_dirty] session filter dport=5000 has flag=dirty

IPsec tunnel

• Generic checks on IPsec based on diagnose vpn ike status
• flush all ike gateway

# Flush all ike gateways ('diagnose vpn ike gateway flush') 
FGT-B1-1:1 flush ike gateway

# Check number of established ike tunnels
FGT-B1-1:1 check [B1_tunnels] ike status has ike_established=3

# Check number of established IPsec tunnels (created and established)
FGT-B1-1:1 check [B1_tunnels] ike status has ipsec_created=3 ipsec_established=3

BGP routes

Checks on routing table BGP from get router info routing-table bgp

# number of bgp routes is 4 :
FGT-B1-1 check [bgp_4_routes] bgp has total=4

# bgp route for subnet 10.0.0.0/24 exist :
FGT-B1-1 check [bgp_subnet_10.0.0.0] bgp has subnet=10.0.0.0/24

# bgp nexthop 10.255.1.253 exist
FGT-B1-1 check [bgp_nexthop_10.255.1.253] bgp has nexthop=10.255.1.253

# bgp has route toward interface vpn_mpls
FGT-B1-1 check [bgp_subnet_10.0.0.0] bgp has interface=vpn_mpls

# multiple requirements can be combined
FGT-B1-1 check [multi] bgp has nexthop=10.255.1.253 nexthop=10.255.2.253 subnet=10.0.0.0/24
FGT-A:1 check [route_ok] routing table bgp 10.0.0.0/24 next-hop 1.1.1.1 interface port1

SD-WAN

Various checks from diagnose sys virtual-wan-link service <SERVICE>

# check alive members :
FGT-B1-1 check [sdwan_1_member1_alive] sdwan service 1 member 1 has state=alive

# check sla value for a particular member (only available for sla type rule)
FGT-B1-1 check [sdwan_1_member1_sla] sdwan service 1 member 1 has sla=0x1

# check that member seq 1 is the preferred member on service 1 (aka rule 1)
FGT-B1-1 check [sdwan_1_preferred] sdwan service 1 member 1 has preferred=1

FortiPoC

Interact with FortiPoC to bring ports up or down Using fpoc link up/down device port

link up / link down

# Bring up link for FGT-B1-port1 switch side
fpoc:1 link up FGT-B1-1 port1

# Bring down link for FGT-B1-port1 switch side
fpoc:1 link down FGT-B1-1 port1

Debug

All debugs are stored in file 'debug.log'
Usefull messages (for instance to track syntax error in testcases definition) should be with level WARNING or ERROR. Program is aborted for level ERROR.

Log level is 'INFO' by default but it can be adjusted to DEBUG using optional --debug

sample

20200317:17:25:30,198 DEBUG   [playbook  .    get_agent_type      :  319] name=HOSTS-B2 type=lxc
20200317:17:25:30,198 DEBUG   [playbook  .    _get_agent_from_tc_l:  347] Found corresponding type=lxc
20200317:17:25:30,198 DEBUG   [playbook  .    run_testcase        :  199] agent_name=HOSTS-B2 agent_type=lxc agent_conn=1
20200317:17:25:30,198 INFO    [playbook  .    _create_agent_conn  :  273] Enter with name=HOSTS-B2 type=lxc conn=1
20200317:17:25:30,198 DEBUG   [playbook  .    _create_agent_conn  :  283] agent=HOSTS-B2 is already in our list
20200317:17:25:30,198 DEBUG   [playbook  .    _create_agent_conn  :  303] Connection to HOSTS-B2:1 already exists
20200317:17:25:30,198 DEBUG   [playbook  .    run_testcase        :  219] Agent already existing

Feedback file values

Lists of possible feedback values :

[playbook_path]     : Path to playbook top directory 
[playbook]          : playbook name
[run]               : run id
[testcase_id]	    : id of testcase
[testcase_name]     : name of testcase
[testcase_progress] : percentage of progress for the testcase
[playlist_id]       : name of the playlist
[progress]			: Overall progress for playbook, playlist or individual testcase
[start_time]        : starting time in unix time
[end_time]          : ending time in unix time

Files samples

conf/macros.txt

This is a sample macro file, in playbooks/PLAYBOOK_NAME/conf/macros.txt
Note that a variable 'server' is used in the macro, it is emcompassed with double-dollars '$$'.
This is made sso $$server$$ is translated to $server$ in the scenario during macro expansion, then variable 'server' (in conf/variables.xml) will be used to replace $server$ in the scenario.

Example of a macro call

Use & to reference a macro. &tcp_connection_check(H1B1,1,H1B2,1,9000)

Definition:

Macros are defined in conf/macros.txt

# Macro for connectivity check, one-way
macro tcp_connection_check(client,client_conn,server,server_conn,port):

# create a random message
set server_ready = random_string(8)
set message = random_string(8)
set test_id = random_string(4)

# Open server on the given port
$server$:$server_conn$ open tcp $port$
$server$:$server_conn$ mark "$server_ready$"

# Client connects
$client$:$client_conn$ connect tcp $$server$$ $port$

# Client send data on forward direction
$client$:$client_conn$ send "$message$"

# Server checks message is received
$server$:$server_conn$ check [tcp_forward_$test_id$] "$message$" since "$server_ready$"

# Closing
$client$:$client_conn$ close
$server$:$server_conn$ close
end

# ---

# Macro for connectivity check, two-way
# uses a double-translation for client and server that need to be defined

macro tcp_connection_twoway_check(client,client_conn,server,server_conn,port):

# create a random message
set server_ready = random_string(8)
set client_ready = random_string(8)
set message1 = random_string(8)
set message2 = random_string(8)
set test_id = random_string(4)

# Open server on the given port
$server$:$server_conn$ open tcp $port$
$server$:$server_conn$ mark "$server_ready$"

# Client connects
$client$:$client_conn$ connect tcp $$server$$ $port$

# Client send data on forward direction
$client$:$client_conn$ send "$message1$"
$client$:$client_conn$ mark "$client_ready$"

# Server checks message1 is received
$server$:$server_conn$ check [tcp_forward_$test_id$] "$message1$" since "$server_ready$"

# Server send data on reply direction
$server$:$server_conn$ send "$message2$"

# Client checks message2 is received
$client$:$client_conn$ check [tcp_reply_$test_id$] "$message2$" since "$client_ready$"


# Closing
$client$:$client_conn$ close
$server$:$server_conn$ close
end

conf/variables.yml

---
# checkitbaby variables definition file

srv-1: 127.0.0.1
cli-1: 127.0.0.1
localhost: 127.0.0.1
google_dns: 8.8.8.8
unreachable_host: 169.254.33.111

conf/agents.yml

---
# checkitbaby agents definition file

# LXC

clt-1:
  type: lxc
  ip: 127.0.0.1
  port: 22
  login: cgustave
  password: ''
  ssh_key_file: "/home/cgustave/.ssh/id_rsa"

srv-1:
  type: lxc
  ip: 127.0.0.1
  port: 22
  login: cgustave
  password: ''
  ssh_key_file: "/home/cgustave/.ssh/id_rsa"

# Vyos

R1-B1:
  type: vyos
  ip: 192.168.122.178
  port: '10111'
  login: vyos
  password: vyos
  ssh_key_file: "/home/cgustave/github/python/checkitbaby/checkitbaby/playbooks/test/conf/id_rsa"

R2-B1:
  type: vyos
  ip: 192.168.122.178
  port: '10113'
  login: vyos
  password: vyos
  ssh_key_file: "/home/cgustave/github/python/checkitbaby/checkitbaby/playbooks/test/conf/id_rsa"

# FortiPoC

fpoc:
  type: fortipoc
  ip: 192.168.122.178
  port: '22'
  login: admin
  password: ''
  ssh_key_file: ''

# FortiGate

FGT-B1-1:
  type: fortigate
  ip: 192.168.122.178
  port: '10101'
  login: admin
  password: ''
  ssh_key_file: ''

FGT-B2-9:
  type: fortigate
  ip: 192.168.122.178
  port: '10102'
  login: admin
  password: ''
  ssh_key_file: ''

conf/playlists.yml

This is a sample of a playlist file

---
# checkitbaby playlist definition file

PL01:
  comment: All LXC tests
  list:
  - '010'
  - '011'

run/1/report.json

This is a sample of a report after a run with --run 1

{
    "result": true,
    "testcases": {
	    "010": {
            "result": true,
            "check": {
			    "origin": true,
                "reply": true
            },
		    "get": {}
        },
        "011": {
		    "result": true,
            "check": {
                "tcp_forward_T69I": true,
				"tcp_reply_T69I": true
            },
            "get": {}
		}
    }
}