7544717b96
Previously, when reaping child processes, if there were no remaining reapable children, the parent strand would only nap 1, which puts unnecessary load on respirate unless at least one child strand exits in the next second. Change this approach by having the exiting child strands, after they release the lease, schedule their parent immediately if the parent has no non-exited child strands. When doing this, you need to be careful to make sure there are not race conditions that would delay the scheduling of the parent. There are two potential situations you need to handle: 1. Multiple children exiting at the same time 2. Parent currently running while child is exiting By waiting until after the child strand leases are released, you still have a race condition with 1, but the race condition is that multiple child strands exiting concurrently could both reschedule the parent strand. However, that isn't a problem. You want to avoid the case where neither child strand schedules the parent, which rescheduling after releasing the lease should do. To handle 2, inside reap use Model#lock! to lock the parent strand. This will make exiting child strands block if they UPDATE the parent strand with a new schedule, until the parent strand's transaction commits. However, it's possible that a child strand already UPDATED the parent. To handle this situation, before calling lock!, store the cached schedule value in a local variable. lock! implicitly does a reload, so compare the schedule value after reload. If the schedule has changed, likely a child scheduled the parent for immediate execution, so nap 0 in that case. Just in case there are unforeseen race conditions that are not handled, only nap for 120 seconds if there are active children. Worst case scenario, this results in a 2 minute delay for running the parent. However, this can potentially result in 120x less load from parent strands polling children.
208 lines
8.6 KiB
Ruby
208 lines
8.6 KiB
Ruby
# frozen_string_literal: true
|
|
|
|
require_relative "../../model/spec_helper"
|
|
|
|
RSpec.describe Prog::Kubernetes::UpgradeKubernetesNode do
|
|
subject(:prog) { described_class.new(st) }
|
|
|
|
let(:st) { Strand.new }
|
|
|
|
let(:project) {
|
|
Project.create(name: "default")
|
|
}
|
|
let(:subnet) {
|
|
Prog::Vnet::SubnetNexus.assemble(Config.kubernetes_service_project_id, name: "test", ipv4_range: "172.19.0.0/16", ipv6_range: "fd40:1a0a:8d48:182a::/64").subject
|
|
}
|
|
|
|
let(:kubernetes_cluster) {
|
|
kc = KubernetesCluster.create_with_id(
|
|
name: "k8scluster",
|
|
version: "v1.32",
|
|
cp_node_count: 3,
|
|
private_subnet_id: subnet.id,
|
|
location_id: Location::HETZNER_FSN1_ID,
|
|
project_id: project.id,
|
|
target_node_size: "standard-4",
|
|
target_node_storage_size_gib: 37
|
|
)
|
|
|
|
lb = LoadBalancer.create(private_subnet_id: subnet.id, name: "somelb", health_check_endpoint: "/foo", project_id: Config.kubernetes_service_project_id)
|
|
LoadBalancerPort.create(load_balancer_id: lb.id, src_port: 123, dst_port: 456)
|
|
kc.add_cp_vm(create_vm)
|
|
kc.add_cp_vm(create_vm)
|
|
|
|
kc.update(api_server_lb_id: lb.id)
|
|
kc
|
|
}
|
|
|
|
let(:kubernetes_nodepool) {
|
|
kn = KubernetesNodepool.create(name: "k8stest-np", node_count: 2, kubernetes_cluster_id: kubernetes_cluster.id, target_node_size: "standard-8", target_node_storage_size_gib: 78)
|
|
kn.add_vm(create_vm)
|
|
kn.add_vm(create_vm)
|
|
kn.reload
|
|
}
|
|
|
|
before do
|
|
allow(Config).to receive(:kubernetes_service_project_id).and_return(Project.create(name: "UbicloudKubernetesService").id)
|
|
allow(prog).to receive(:kubernetes_cluster).and_return(kubernetes_cluster)
|
|
end
|
|
|
|
describe "#before_run" do
|
|
before do
|
|
Strand.create(id: kubernetes_cluster.id, label: "wait", prog: "KubernetesClusterNexus")
|
|
end
|
|
|
|
it "exits when kubernetes cluster is deleted and has no children itself" do
|
|
st.update(prog: "Kubernetes::UpgradeKubernetesNode", label: "somestep", stack: [{}])
|
|
prog.before_run # Nothing happens
|
|
|
|
kubernetes_cluster.strand.label = "destroy"
|
|
expect { prog.before_run }.to exit({"msg" => "upgrade cancelled"})
|
|
end
|
|
|
|
it "donates when kubernetes cluster is deleted and but has a child" do
|
|
st.update(prog: "Kubernetes::UpgradeKubernetesNode", label: "somestep", stack: [{}])
|
|
Strand.create(parent_id: st.id, prog: "Kubernetes::ProvisionKubernetesNode", label: "start", stack: [{}], lease: Time.now + 10)
|
|
kubernetes_cluster.strand.label = "destroy"
|
|
expect { prog.before_run }.to nap(120)
|
|
end
|
|
end
|
|
|
|
describe "#start" do
|
|
it "provisions a new kubernetes node" do
|
|
expect(prog).to receive(:frame).and_return({})
|
|
expect(prog).to receive(:bud).with(Prog::Kubernetes::ProvisionKubernetesNode, {})
|
|
expect { prog.start }.to hop("wait_new_node")
|
|
|
|
expect(prog).to receive(:frame).and_return({"nodepool_id" => kubernetes_nodepool.id})
|
|
expect(prog).to receive(:bud).with(Prog::Kubernetes::ProvisionKubernetesNode, {"nodepool_id" => kubernetes_nodepool.id})
|
|
expect { prog.start }.to hop("wait_new_node")
|
|
end
|
|
end
|
|
|
|
describe "#wait_new_node" do
|
|
it "donates if there are sub-programs running" do
|
|
st.update(prog: "Kubernetes::UpgradeKubernetesNode", label: "wait_new_node", stack: [{}])
|
|
Strand.create(parent_id: st.id, prog: "Kubernetes::ProvisionKubernetesNode", label: "start", stack: [{}], lease: Time.now + 10)
|
|
expect { prog.wait_new_node }.to nap(120)
|
|
end
|
|
|
|
it "hops to assign_role if there are no sub-programs running" do
|
|
st.update(prog: "Kubernetes::UpgradeKubernetesNode", label: "wait_new_node", stack: [{}])
|
|
Strand.create(parent_id: st.id, prog: "Kubernetes::ProvisionKubernetesNode", label: "start", stack: [{}], exitval: {"vm_id" => "12345"})
|
|
expect { prog.wait_new_node }.to hop("drain_old_node")
|
|
expect(prog.strand.stack.first["new_vm_id"]).to eq "12345"
|
|
end
|
|
end
|
|
|
|
describe "#drain_old_node" do
|
|
let(:sshable) { instance_double(Sshable) }
|
|
let(:old_vm) { create_vm }
|
|
|
|
before do
|
|
allow(prog).to receive(:frame).and_return({"old_vm_id" => old_vm.id})
|
|
expect(prog.old_vm.id).to eq(old_vm.id)
|
|
cp_vm = create_vm
|
|
expect(kubernetes_cluster).to receive(:cp_vms).and_return([old_vm, cp_vm])
|
|
allow(cp_vm).to receive(:sshable).and_return(sshable)
|
|
|
|
expect(prog).to receive(:register_deadline).with("remove_old_node_from_cluster", 60 * 60)
|
|
end
|
|
|
|
it "starts the drain process when run for the first time and naps" do
|
|
expect(sshable).to receive(:d_check).with("drain_node").and_return("NotStarted")
|
|
expect(sshable).to receive(:d_run).with("drain_node", "sudo", "kubectl", "--kubeconfig=/etc/kubernetes/admin.conf", "drain", old_vm.name, "--ignore-daemonsets", "--delete-emptydir-data")
|
|
expect { prog.drain_old_node }.to nap(10)
|
|
end
|
|
|
|
it "naps when the node is getting drained" do
|
|
expect(sshable).to receive(:d_check).with("drain_node").and_return("InProgress")
|
|
expect { prog.drain_old_node }.to nap(10)
|
|
end
|
|
|
|
it "restarts when it fails" do
|
|
expect(sshable).to receive(:d_check).with("drain_node").and_return("Failed")
|
|
expect(sshable).to receive(:d_restart).with("drain_node")
|
|
expect { prog.drain_old_node }.to nap(10)
|
|
end
|
|
|
|
it "naps when daemonizer something unexpected and waits for the page" do
|
|
expect(sshable).to receive(:d_check).with("drain_node").and_return("UnexpectedState")
|
|
expect { prog.drain_old_node }.to nap(60 * 60)
|
|
end
|
|
|
|
it "drains the old node and hops to drop the old node" do
|
|
expect(sshable).to receive(:d_check).with("drain_node").and_return("Succeeded")
|
|
expect { prog.drain_old_node }.to hop("remove_old_node_from_cluster")
|
|
end
|
|
end
|
|
|
|
describe "#remove_old_node_from_cluster" do
|
|
before do
|
|
old_vm = create_vm
|
|
new_vm = create_vm
|
|
allow(prog).to receive(:frame).and_return({"old_vm_id" => old_vm.id, "new_vm_id" => new_vm.id})
|
|
expect(prog.old_vm.id).to eq(old_vm.id)
|
|
expect(prog.new_vm.id).to eq(new_vm.id)
|
|
|
|
expect(kubernetes_nodepool.vms.count).to eq(2)
|
|
expect(kubernetes_cluster.reload.all_vms.map(&:id)).to eq((kubernetes_cluster.cp_vms + kubernetes_nodepool.vms).map(&:id))
|
|
|
|
mock_sshable = instance_double(Sshable)
|
|
expect(prog.old_vm).to receive(:sshable).and_return(mock_sshable)
|
|
expect(mock_sshable).to receive(:cmd).with("sudo kubeadm reset --force")
|
|
end
|
|
|
|
it "removes the old node from the CP while upgrading the control plane" do
|
|
expect(prog.kubernetes_nodepool).to be_nil
|
|
api_server_lb = instance_double(LoadBalancer)
|
|
expect(kubernetes_cluster).to receive(:api_server_lb).and_return(api_server_lb)
|
|
expect(kubernetes_cluster).to receive(:remove_cp_vm).with(prog.old_vm)
|
|
expect(api_server_lb).to receive(:detach_vm).with(prog.old_vm)
|
|
expect { prog.remove_old_node_from_cluster }.to hop("delete_node_object")
|
|
end
|
|
|
|
it "removes the old node from the nodepool while upgrading the node pool" do
|
|
allow(prog).to receive(:frame).and_return({"old_vm_id" => prog.old_vm.id, "new_vm_id" => prog.new_vm.id, "nodepool_id" => kubernetes_nodepool.id})
|
|
expect(prog.kubernetes_nodepool).not_to be_nil
|
|
expect(prog.kubernetes_nodepool).to receive(:remove_vm).with(prog.old_vm)
|
|
expect { prog.remove_old_node_from_cluster }.to hop("delete_node_object")
|
|
end
|
|
end
|
|
|
|
describe "#delete_node_object" do
|
|
before do
|
|
old_vm = create_vm
|
|
allow(prog).to receive(:frame).and_return({"old_vm_id" => old_vm.id})
|
|
expect(prog.old_vm.id).to eq(old_vm.id)
|
|
end
|
|
|
|
it "deletes the node object from kubernetes" do
|
|
client = instance_double(Kubernetes::Client)
|
|
expect(kubernetes_cluster).to receive(:client).and_return(client)
|
|
expect(client).to receive(:delete_node).with(prog.old_vm.name).and_return(Net::SSH::Connection::Session::StringWithExitstatus.new("success", 0))
|
|
expect { prog.delete_node_object }.to hop("destroy_node")
|
|
end
|
|
|
|
it "raises if the error of delete node command is not successful" do
|
|
client = instance_double(Kubernetes::Client)
|
|
expect(kubernetes_cluster).to receive(:client).and_return(client)
|
|
expect(client).to receive(:delete_node).with(prog.old_vm.name).and_return(Net::SSH::Connection::Session::StringWithExitstatus.new("failed", 1))
|
|
expect { prog.delete_node_object }.to raise_error(RuntimeError)
|
|
end
|
|
end
|
|
|
|
describe "#destroy_node" do
|
|
before do
|
|
old_vm = create_vm
|
|
allow(prog).to receive(:frame).and_return({"old_vm_id" => old_vm.id})
|
|
expect(prog.old_vm.id).to eq(old_vm.id)
|
|
end
|
|
|
|
it "destroys the old node" do
|
|
expect(prog.old_vm).to receive(:incr_destroy)
|
|
expect { prog.destroy_node }.to exit({"msg" => "upgraded node"})
|
|
end
|
|
end
|
|
end
|