From 7080a206496256d4aff52ee6a6351322f87e3bf4 Mon Sep 17 00:00:00 2001 From: Simon Murray Date: Fri, 30 Aug 2024 11:08:48 +0100 Subject: [PATCH] Increase Machine Check Timeouts (#18) Ironic can take a long time, like between 15-30 minutes to spin up a machine. The timeout value is deliberately very lax e.g. 2h initially to allow time to SSH in and debug errors. Eventually we should be able to specify timeouts based on machine type, e.g. VMs will be pretty much instant, whereas baremetal will take longer. --- charts/cluster-api-cluster-openstack/Chart.yaml | 2 +- charts/cluster-api-cluster-openstack/README.md | 2 +- .../templates/machinehealthcheck.yaml | 3 ++- 3 files changed, 4 insertions(+), 3 deletions(-) diff --git a/charts/cluster-api-cluster-openstack/Chart.yaml b/charts/cluster-api-cluster-openstack/Chart.yaml index 67deeaf..11b3c83 100644 --- a/charts/cluster-api-cluster-openstack/Chart.yaml +++ b/charts/cluster-api-cluster-openstack/Chart.yaml @@ -2,5 +2,5 @@ apiVersion: v2 name: cluster-api-cluster-openstack description: A Helm chart to deploy a Kubernetes Cluster type: application -version: v0.5.3 +version: v0.5.4 icon: https://raw.githubusercontent.com/unikorn-cloud/helm-cluster-api/main/icons/default.png diff --git a/charts/cluster-api-cluster-openstack/README.md b/charts/cluster-api-cluster-openstack/README.md index 68e5c1d..485332c 100644 --- a/charts/cluster-api-cluster-openstack/README.md +++ b/charts/cluster-api-cluster-openstack/README.md @@ -36,7 +36,7 @@ spec: source: repoURL: https://unikorn-cloud.github.io/helm-cluster-api chart: cluster-api-cluster-openstack - targetRevision: v0.5.3 + targetRevision: v0.5.4 helm: releaseName: foo # Remove the default work queue. diff --git a/charts/cluster-api-cluster-openstack/templates/machinehealthcheck.yaml b/charts/cluster-api-cluster-openstack/templates/machinehealthcheck.yaml index f68108e..4568a4b 100644 --- a/charts/cluster-api-cluster-openstack/templates/machinehealthcheck.yaml +++ b/charts/cluster-api-cluster-openstack/templates/machinehealthcheck.yaml @@ -10,7 +10,8 @@ metadata: spec: clusterName: {{ include "cluster.name" . }} maxUnhealthy: 50% - nodeStartupTimeout: 10m0s + # Note this is really relaxed because of ironic taking forever. + nodeStartupTimeout: 2h0m0s selector: matchLabels: cluster.x-k8s.io/cluster-name: {{ include "cluster.name" . }}