Skip to content

Latest commit

 

History

History
155 lines (109 loc) · 5.26 KB

README.md

File metadata and controls

155 lines (109 loc) · 5.26 KB
Raw

Deploy metallb "pools" in Kubernetes using Kustomize

This means IPAddressPools, L2Advertisements, and BGPAdvertisements.

This does NOT install or configure metallb.

Requirements

  • A functional k8s cluster.
  • metallb must already be installed. There are a lot of configuration knobs and dials to turn with an installation. This project isn't for deploying and configuring metallb, but rather the IP pools and advertisements.

Deployment (without overlays)

NOTE: Do not change the namespace from metallb-system. I believe this is a requirement for metallb itself.

Update kustomization.yaml with changes

IPAddressPool section

  • The base IPAddressPool name is pool. To have more than one pool, you will need to rename it. You will probably want to do so anyway. Choose the new name in the /metadata/name rename op.
  • The IP address pool itself is a single line from an array in the base definition. You can add more here if you like, but the default is to replace just the one entry. Feel free to use CIDR notation, as the AvoidBuggyIPs flag is set to true (which will avoid .0 and .255 addresses). Replace /spec/addresses/0 value with your IP range.

L2Advertisement

  • Edit patches/advertisement.yaml and put in whatever changes you need. The existing example sets an ethernet interface and 3 nodes. There are many more configuration options available to you. Do not change the /metadata/name in this patch! The renames will take place later in kustomization.yaml as the patch file must match the name. Yes, we could set target and get around that. It's problematic to troubleshoot this sometimes as folks don't know what they've done, so we keep it fairly pure here.
  • Edit the value for /metadata/name. This is the rename op for the this L2Advertisement.
  • Edit the value for /spec/ipAddressPools/0. This value should match what you renamed /metadata/name to in the IPAddressPool section.

BGPAdvertisement

This one requires replacing the L2Advertisement section with the commented out BGPAdvertisement section. Everything else should remain similar to the previous instructions.

NOTE: This also requires commenting out the l2 line and uncommenting the bgp one:

  # - base/advertisements/l2/advertisement.yaml
  - base/advertisements/bgp/advertisement.yaml
  • Edit patches/advertisement.yaml and put in whatever changes you need. The existing example sets an ethernet interface and 3 nodes. There are many more configuration options available to you. Do not change the /metadata/name in this patch! The renames will take place later in kustomization.yaml as the patch file must match the name. Yes, we could set target and get around that. It's problematic to troubleshoot this sometimes as folks don't know what they've done, so we keep it fairly pure here.
  • Edit the value for /metadata/name. This is the rename op for the this BGPAdvertisement.
  • Edit the value for /spec/ipAddressPools/0. This value should match what you renamed /metadata/name to in the IPAddressPool section.

Both L2 and BGP at the same time

As metallb themselves tell you, you can use both advertisement types at once, for the same IPAddressPool. To do so requires both lines will need to be uncommented:

  - base/advertisements/l2/advertisement.yaml
  - base/advertisements/bgp/advertisement.yaml

You will also need to copy patches/advertisement.yaml to another file, e.g. patches/bgp_advertisement.yaml. So long as you can distinguish between the two, that's all that matters.

You would then need to uncomment both sections in kustomization.yaml, and ensure that the correct patches/FILENAME.yaml patches the correct advertisement type.

Test

  • (Optional) To preview generated configuration before deploying:

kubectl kustomize .

Apply

  • Run the following command to build and deploy:

kubectl apply -k .

Deployment using overlays

Copy the basic structure to each overlay

  • Copy kustomization.yaml, and the various patches/* files to overlays/NAME, or whatever directory structure you prefer.
  • For complex deployments, this could look like overlays/GROUP/SUBGROUP, where GROUP could be teamname, and SUBGROUP could be prod, qa, or dev. This allows fine granular control over which blocks of IPs are allocated to which teams and environments.
  • Each NAME or SUBGROUP should have its own kustomization.yaml and patches subdirectory.
  • Be sure to update the resources section of kustomization.yaml to be able to reach the base directory.

For NAME (one level)

resources:
  - ../../base
  - ../../base/advertisements/l2

For GROUP/SUBGROUP (two levels)

resources:
  - ../../../base
  - ../../../base/advertisements/l2

Edit each overlay path accordingly

Apply the same configuration steps as above for each overlay path you create.

Test

  • (Optional) To preview generated configuration before deploying:

kubectl kustomize overlays/NAME

or

kubectl kustomize overlays/GROUP/SUBGROUP

Apply

  • Run the following command to build and deploy:

kubectl apply -k overlays/NAME

or

kubectl apply -k overlays/GROUP/SUBGROUP