I upgraded my homelab cluster using Talos Linux today.

I made a mistake and forgot to use the custom-built image using iscsi-tools.

Learning the hard way.

Here are my notes:

Talos upgrade

[!IMPORTANT] Remember to create a custom image including iscsi-tools

Set up CLI environment and set databases to maintenance mode

export TALOS_CP="192.168.100.107"
export TALOS_W1="192.168.100.245"
export TALOS_W2="192.168.100.60"

k cnp maintenance set --all-namespaces

[!WARNING] Use the --preserve flag on single-node control plane clusters. Only needed for the control plane node.

Upgrade path:

  1. from 1.6.4 to 1.6.8
  2. from 1.6.8 to 1.7.5

Upgrade 1

[!NOTE] I made a mistake and did not update to the custom-built image including the iscsi-tools This caused my cluster storage to break.

talosctl upgrade --preserve --nodes $TALOS_CP -e $TALOS_CP --image ghcr.io/siderolabs/installer:v1.6.8
talosctl upgrade --nodes $TALOS_W1 -e $TALOS_CP --image ghcr.io/siderolabs/installer:v1.6.8
talosctl upgrade --nodes $TALOS_W2 -e $TALOS_CP --image ghcr.io/siderolabs/installer:v1.6.8

Upgrade 2

I went to https://factory.talos.dev/ and created a new image including iscsi-tools.

Everything worked fine after that.

talosctl upgrade --preserve --nodes $TALOS_CP -e $TALOS_CP --image factory.talos.dev/installer/c9078f9419961640c712a8bf2bb9174933dfcf1da383fd8ea2b7dc21493f8bac:v1.7.5
talosctl upgrade --nodes $TALOS_W1 -e $TALOS_CP --image factory.talos.dev/installer/c9078f9419961640c712a8bf2bb9174933dfcf1da383fd8ea2b7dc21493f8bac:v1.7.5
talosctl upgrade --nodes $TALOS_W2 -e $TALOS_CP --image factory.talos.dev/installer/c9078f9419961640c712a8bf2bb9174933dfcf1da383fd8ea2b7dc21493f8bac:v1.7.5

Kubernetes Upgrade

[!WARNING] Make sure to upgrade the Talos CLI before proceeding. “It is advisable to use the same version of talosctl as the version of the boot media used.” - Talos Docs

Upgrade Path

  1. From 1.29.0 to 1.29.3
  2. From 1.29.3 to 1.30.3

[!NOTE]

[!NOTE] To trigger a Kubernetes upgrade, issue a command specifying the version of Kubernetes to upgrade to, such as: talosctl –nodes upgrade-k8s –to 1.30.0 Note that the –nodes parameter specifies the control plane node to send the API call to, but all members of the cluster will be upgraded. To check what will be upgraded you can run talosctl upgrade-k8s with the –dry-run flag

Round 1

k cnp maintenance set --all-namespaces
kubent
talosctl --nodes $TALOS_CP -e $TALOS_CP upgrade-k8s --to 1.29.3 --dry-run
talosctl --nodes $TALOS_CP -e $TALOS_CP upgrade-k8s --to 1.29.3

Round 2

k cnp maintenance set --all-namespaces
kubent
talosctl --nodes $TALOS_CP -e $TALOS_CP upgrade-k8s --to 1.30.3 --dry-run
talosctl --nodes $TALOS_CP -e $TALOS_CP upgrade-k8s --to 1.30.3

202407311407