fix: Speed up daemonset upgrades

We had max_unavailable set to 1 which means it can take up to n x 30+ seconds to perform rolling upgrades. That's a long time for big environments with hundreds of nodes.
rackerlabs · May 10, 2024 · 5a39db2 · 5a39db2
1 parent 4019122
commit 5a39db2
Show file tree

Hide file tree

Showing 2 changed files with 2 additions and 2 deletions.
diff --git a/helm-configs/nova/nova-helm-overrides.yaml b/helm-configs/nova/nova-helm-overrides.yaml
@@ -2330,7 +2330,7 @@ pod:
         compute:
           enabled: true
           min_ready_seconds: 0
-          max_unavailable: 1
+          max_unavailable: 20%
     disruption_budget:
       metadata:
         min_available: 0

diff --git a/kustomize/libvirt/helm/libvirt-helm-overrides.yaml b/kustomize/libvirt/helm/libvirt-helm-overrides.yaml
@@ -222,7 +222,7 @@ pod:
         libvirt:
           enabled: true
           min_ready_seconds: 0
-          max_unavailable: 1
+          max_unavailable: 20%
   resources:
     enabled: false
     libvirt: