Description
ACK node auto repair checks for supported node faults and performs recovery actions. When it is disabled, recovery requires manual work or separate automation.
Potential impact
Delayed recovery of failed nodes can reduce workload availability.
Remediation
Set management.enable = true and auto_repair = true for a managed node pool. Check the effect of restarts and concurrent disruption limits on workloads.
Examples
These excerpts enable managed features and auto repair. Node authentication and disk settings are omitted; review the other illustrated operational options, including auto upgrade.
Before
hcl
resource "alicloud_cs_kubernetes_node_pool" "default" {
name = var.name
cluster_id = alicloud_cs_managed_kubernetes.default.0.id
vswitch_ids = [alicloud_vswitch.default.id]
instance_types = [data.alicloud_instance_types.default.instance_types.0.id]
desired_size = 1
}
After
hcl
resource "alicloud_cs_kubernetes_node_pool" "default" {
name = var.name
cluster_id = alicloud_cs_managed_kubernetes.default.0.id
vswitch_ids = [alicloud_vswitch.default.id]
instance_types = [data.alicloud_instance_types.default.instance_types.0.id]
desired_size = 1
management {
enable = true
auto_repair = true
auto_upgrade = true
surge = 1
max_unavailable = 1
}
}