Skip to content

saparikh/bf_snapshot_collector

Repository files navigation

Batfish Snapshot Collector

This repository contains a script that will collect the running configuration from network elements and package them in the format required by Batfish.

Assumptions / Restrictions:

Supported devices

  • Cisco ASA, IOS, IOS-XE, NXOS, IOS-XR
  • Cumulus Linux
  • Arista EOS
  • Juniper JunOS
  • Checkpoint Gateway
  • A10 Loadbalancer (Ver 2.x and Ver 4.x)

Device inventory

The collection script expects a valid Ansible inventory file. There is an example_inventory.yml file to show you the expected format.

The device name must either be resolvable via DNS or you must specify the ansible_host variable for the device. Your inventory may have a mix of devices which are resolvable via DNS and devices that use the ansible_host variable.

Device access

All devices in the inventory MUST be accessible with the SAME username and password. That user MUST either be put into enable mode or be granted correct privilege level to retrieve running configuration without being in enable mode.

For Cumulus, make sure the user can read all the necessary files, such as /etc/frr/frr.conf, without using sudo

Collector setup

  1. Create a python (3.7+) virtual environment and install dependencies
pip install -r requirements.txt
  1. Create an inventory file for your devices

  2. Create a file with settings to connect to Batfish. See example_bf_settings.env for an example.

Taking network snapshots

You can collect network snapshots and upload to Batfish via this command:

BF_COLLECTOR_USER=<device-access-user> \
BF_COLLECTOR_PASSWORD=<device-access-passowrd> \ 
  bash snapshot_network.sh <inventory file> <batfish settings file> [<collection directory>]

In the first three lines of this command, we are passing in username and password to access the device as environment variables. They can be omitted if the environment variables are set via other means (e.g., the export command).

The third argument to the snapshot_network.sh script is optional and specifies where the output of the collector should be stored. The default value is the current working directory.

The snapshots are given a name based on collection time, such as 20211123_19:58:34. Collected snapshots are uploaded to Batfish but also left on local drive in the collection directory under a folder with the snapshot name. The logs will be under the folder logs/20211123_19:58:34 in the collection directory.

####Note: The script also collects some standard show command output. This data is useable natively with Batfish Enterprise, but not with Batfish. Show data collection can take a long time, so if you do not want or need that data, remove the appropriate lines from the bash script.

When using Batfish Enterprise

If you are using Batfish Enterprise:

  1. Set the BFE_ENTERPRISE variable to True in the batfish settings file.

  2. Supply the BFE_ACCESS_TOKEN and BFE_SSL_CERT environment variables to snapshot_network script. Thus, the command becomes

BF_COLLECTOR_USER=<device-access-user> \
BF_COLLECTOR_PASSWORD=<device-access-passowrd> \ 
BFE_ACCESS_TOKEN=<bfe-access-token> \ 
BFE_SSL_CERT=<bfe-ssl-cert> \ 
  bash snapshot_network.sh <inventory file> <batfish settings file> [<collection directory>]

BFE_SSL_CERT variable is optional. It is only needed if you used a self-signed cert when installing Batfish Enterprise, instead of using a valid certificate.

About

Batfish Snapshot Collector

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 3

  •  
  •  
  •