feat: allow configuring Sidero deployment strategy

Upgrading Sidero on a single-node cluster with host networking enabled fails because the used ports can only be bound once. By setting the deployment strategy to `Recreate` in these cases upgrades can succeed (but with a small amount of downtime). Signed-off-by: Gerard de Leeuw <gdeleeuw@leeuwit.nl> Signed-off-by: Andrey Smirnov <andrey.smirnov@talos-systems.com>
siderolabs · May 18, 2022 · 294fd62 · 294fd62
1 parent 41880ed
commit 294fd62
Show file tree

Hide file tree

Showing 6 changed files with 7 additions and 1 deletion.
diff --git a/app/sidero-controller-manager/config/manager/manager.yaml b/app/sidero-controller-manager/config/manager/manager.yaml
@@ -49,6 +49,8 @@ spec:
     matchLabels:
       control-plane: sidero-controller-manager
   replicas: 1
+  strategy:
+    type: ${SIDERO_CONTROLLER_MANAGER_DEPLOYMENT_STRATEGY:=RollingUpdate}
   template:
     metadata:
       labels:

diff --git a/sfyra/pkg/capi/capi.go b/sfyra/pkg/capi/capi.go
@@ -123,6 +123,7 @@ func (clusterAPI *Manager) Install(ctx context.Context) error {
 
 	// set template environment variables
 	os.Setenv("SIDERO_CONTROLLER_MANAGER_HOST_NETWORK", "true")
+	os.Setenv("SIDERO_CONTROLLER_MANAGER_DEPLOYMENT_STRATEGY", "Recreate")
 	os.Setenv("SIDERO_CONTROLLER_MANAGER_API_ENDPOINT", clusterAPI.cluster.SideroComponentsIP().String())
 	os.Setenv("SIDERO_CONTROLLER_MANAGER_SERVER_REBOOT_TIMEOUT", "30s") // wiping/reboot is fast in the test environment
 	os.Setenv("SIDERO_CONTROLLER_MANAGER_TEST_POWER_EXPLICIT_FAILURE", fmt.Sprintf("%f", clusterAPI.options.PowerSimulatedExplicitFailureProb))

diff --git a/website/content/docs/v0.5/Getting Started/install-clusterapi.md b/website/content/docs/v0.5/Getting Started/install-clusterapi.md
@@ -24,6 +24,7 @@ options.
 
 ```bash
 export SIDERO_CONTROLLER_MANAGER_HOST_NETWORK=true
+export SIDERO_CONTROLLER_MANAGER_DEPLOYMENT_STRATEGY=Recreate
 export SIDERO_CONTROLLER_MANAGER_API_ENDPOINT=192.168.1.150
 export SIDERO_CONTROLLER_MANAGER_SIDEROLINK_ENDPOINT=192.168.1.150
 

diff --git a/website/content/docs/v0.5/Guides/bootstrapping.md b/website/content/docs/v0.5/Guides/bootstrapping.md
@@ -147,6 +147,7 @@ To install Sidero and the other Talos providers, simply issue:
 
 ```bash
 SIDERO_CONTROLLER_MANAGER_HOST_NETWORK=true \
+  SIDERO_CONTROLLER_MANAGER_DEPLOYMENT_STRATEGY=Recreate \
   SIDERO_CONTROLLER_MANAGER_API_ENDPOINT=$PUBLIC_IP \
   clusterctl init -b talos -c talos -i sidero
 ```

diff --git a/website/content/docs/v0.5/Guides/sidero-on-rpi4.md b/website/content/docs/v0.5/Guides/sidero-on-rpi4.md
@@ -104,7 +104,7 @@ kubectl get nodes
 Install Sidero with host network mode, exposing the endpoints on the node's address:
 
 ```bash
-SIDERO_CONTROLLER_MANAGER_HOST_NETWORK=true SIDERO_CONTROLLER_MANAGER_API_ENDPOINT=${SIDERO_IP} clusterctl init -i sidero -b talos -c talos
+SIDERO_CONTROLLER_MANAGER_HOST_NETWORK=true SIDERO_CONTROLLER_MANAGER_DEPLOYMENT_STRATEGY=Recreate SIDERO_CONTROLLER_MANAGER_API_ENDPOINT=${SIDERO_IP} clusterctl init -i sidero -b talos -c talos
 ```
 
 Watch the progress of installation with:

diff --git a/website/content/docs/v0.5/Overview/installation.md b/website/content/docs/v0.5/Overview/installation.md
@@ -14,6 +14,7 @@ Sidero supports several variables to configure the installation, these variables
 variables or as variables in the `clusterctl` configuration:
 
 - `SIDERO_CONTROLLER_MANAGER_HOST_NETWORK` (`false`): run `sidero-controller-manager` on host network
+- `SIDERO_CONTROLLER_MANAGER_DEPLOYMENT_STRATEGY` (`RollingUpdate`): strategy to use when updating `sidero-controller-manager`, use `Recreate` when using a single node and `SIDERO_CONTROLLER_MANAGER_HOST_NETWORK` is `true`
 - `SIDERO_CONTROLLER_MANAGER_API_ENDPOINT` (empty): specifies the IP address controller manager API service can be reached on, defaults to the node IP (TCP)
 - `SIDERO_CONTROLLER_MANAGER_API_PORT` (8081): specifies the port controller manager can be reached on
 - `SIDERO_CONTROLLER_MANAGER_CONTAINER_API_PORT` (8081): specifies the controller manager internal container port