bigip-to-terraform

Transform running F5 BigIP config into Terraform resources.

(C) 2022 Nicolai Langfeldt nicolai.langfeldt@schibsted.com. This code is developed on Schibsteds time and dime.

This is Open Source with a Apache2 license. See LICENSE

Contributions welcome: please make a PR.

Usage

This script covers three use cases

• ./runner -u: list unreferenced pool and node resources to help you see what can be removed from your BigIP configuration.
• ./runner -v <vipspec>: Pick out data from your BigIP to construct a terraform config to configure the VIP, pool and nodes needed for the specified VIP(s). Vipspec can be a /regular expression/ or just a 'substring'. It will be searched for in the VIP name, the VIP "fullpath" (which includes /Common or other partition name) and the destination, i.e. the IP of the VIP. All matching VIPs and their pool and node resources will be extracted. Adds to terraform state already present in the terraformer directory.
• ./runner - just extract all of it. By default this also documents the unreferenced resources. Clears any previous state. This can be quite time consuming depending on your configuration size.

Add -c to clear the terraform state before importing the newly discovered resources.

Running with -v is additive (unless you do -c) so you can export all the VIPs and attendant resources you'd like.

If you run several times in a way that results in overlapping VIP, pool or node resources that will result in a conflict and the import of the resources will be ungraciously aborted mid stream.

Installing requirements

You will need

• Python 3
• pip
• terraform, at least version 0.13 from what I can see
• jq

pip install -r requirements.txt

If your pip is quarrelsome about "This environment is externally managed" you might or might not want to run

pip install --break-system-packages -r requirements.txt

If you decide to install using a venv you might need to make other changes as well.

Restrictions

• Only tested with a /Common partition. If you use partitions you may need to debug and fix some.
• The script only handles VIPs, pools and nodes. Policies and other things are not handled. Patches welcome.
• I also assume this will break if you have some IPv6. We do have some IPv6 VIPs but I would not consider this well tested.

Running

In the root directory create login.json with contents like this:

{
    "bigip": "172.16.17.18",
    "user": "someuser",
    "password": "the-password"
}

The user can be a "audit" user, it does not need r/w permissions on the BigIP.

Now execute ./runner as described in Usage above. Unless you specify -u this will:

1. run the python script that dumps the resources we (I) care about: VIPs, pools and nodes.
2. Skeleton declarations for the resources are written to import.tf and then copied to the terraformer directory.
3. In the import.tf file are embedded comments that contain the commands to import all the resources from the running BigIP to the local terraform.tfstate file. These commands are executed.
4. When all the resources are imported into the terraform state it is used to generate resources.tf which contain the full resource declarations for your use.
5. Finaly import.tf is deleted from the terraformer directory and a importer-<time-date>.tf is left in the root directory.

After all this you should be able to use terraform to configure your BigIP, either for the resources you specified with -v or the whole config.

If you grep for "not referenced" in the importer-<time-date>.tf file you will get a list of pools and nodes that are not in use. You can delete these from your BigIP if you like.