Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Rdma functional tests improvements #825

Open
wants to merge 2 commits into
base: master
Choose a base branch
from
+81 −53
Open

Rdma functional tests improvements #825

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
64a9cff
extend functional tests timeout
SchSeba Dec 31, 2024
9c7d149
Make rdma functional tests robust for single node environments
SchSeba Dec 31, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion hack/run-e2e-conformance.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -7,4 +7,4 @@ GOPATH="${GOPATH:-$HOME/go}"
JUNIT_OUTPUT="${JUNIT_OUTPUT:-/tmp/artifacts}"
export PATH=$PATH:$GOPATH/bin

${root}/bin/ginkgo -output-dir=$JUNIT_OUTPUT --junit-report "unit_report.xml" -v "$SUITE" -- -report=$JUNIT_OUTPUT
${root}/bin/ginkgo --timeout 3h -output-dir=$JUNIT_OUTPUT --junit-report "unit_report.xml" -v "$SUITE" -- -report=$JUNIT_OUTPUT
132 changes: 80 additions & 52 deletions test/conformance/tests/test_networkpool.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -27,6 +27,7 @@ import (
var _ = Describe("[sriov] NetworkPool", Ordered, func() {
var testNode string
var interfaces []*sriovv1.InterfaceExt
var resourceName = "testrdma"

BeforeAll(func() {
err := namespaces.Create(namespaces.Test, clients)
Expand Down Expand Up @@ -68,10 +69,12 @@ var _ = Describe("[sriov] NetworkPool", Ordered, func() {
By("waiting for operator to finish the configuration")
WaitForSRIOVStable()
nodeState := &sriovv1.SriovNetworkNodeState{}
err = clients.Get(context.Background(), client.ObjectKey{Name: testNode, Namespace: operatorNamespace}, nodeState)
Expect(err).ToNot(HaveOccurred())
Expect(nodeState.Spec.System.RdmaMode).To(Equal(consts.RdmaSubsystemModeExclusive))
Expect(nodeState.Status.System.RdmaMode).To(Equal(consts.RdmaSubsystemModeExclusive))
Eventually(func(g Gomega) {
err = clients.Get(context.Background(), client.ObjectKey{Name: testNode, Namespace: operatorNamespace}, nodeState)
g.Expect(err).ToNot(HaveOccurred())
g.Expect(nodeState.Spec.System.RdmaMode).To(Equal(consts.RdmaSubsystemModeExclusive))
g.Expect(nodeState.Status.System.RdmaMode).To(Equal(consts.RdmaSubsystemModeExclusive))
}, 20*time.Minute, 5*time.Second).Should(Succeed())

By("Checking rdma mode and kernel args")
cmdlineOutput, _, err := runCommandOnConfigDaemon(testNode, "/bin/bash", "-c", "cat /host/proc/cmdline")
Expand All @@ -85,14 +88,22 @@ var _ = Describe("[sriov] NetworkPool", Ordered, func() {
Expect(strings.HasPrefix(output, "1")).To(BeTrue())

By("configure rdma mode to shared")
networkPool.Spec.RdmaMode = consts.RdmaSubsystemModeShared
err = clients.Update(context.Background(), networkPool)
Expect(err).ToNot(HaveOccurred())
Eventually(func(g Gomega) {
err = clients.Get(context.Background(), client.ObjectKey{Name: testNode, Namespace: operatorNamespace}, networkPool)
g.Expect(err).ToNot(HaveOccurred())
networkPool.Spec.RdmaMode = consts.RdmaSubsystemModeShared
err = clients.Update(context.Background(), networkPool)
g.Expect(err).ToNot(HaveOccurred())
}, time.Minute, 5*time.Second).Should(Succeed())

By("waiting for operator to finish the configuration")
WaitForSRIOVStable()
err = clients.Get(context.Background(), client.ObjectKey{Name: testNode, Namespace: operatorNamespace}, nodeState)
Expect(err).ToNot(HaveOccurred())
Expect(nodeState.Spec.System.RdmaMode).To(Equal(consts.RdmaSubsystemModeShared))
Expect(nodeState.Status.System.RdmaMode).To(Equal(consts.RdmaSubsystemModeShared))
Eventually(func(g Gomega) {
err = clients.Get(context.Background(), client.ObjectKey{Name: testNode, Namespace: operatorNamespace}, nodeState)
g.Expect(err).ToNot(HaveOccurred())
g.Expect(nodeState.Spec.System.RdmaMode).To(Equal(consts.RdmaSubsystemModeShared))
g.Expect(nodeState.Status.System.RdmaMode).To(Equal(consts.RdmaSubsystemModeShared))
}, 20*time.Minute, 5*time.Second).Should(Succeed())

By("Checking rdma mode and kernel args")
cmdlineOutput, _, err = runCommandOnConfigDaemon(testNode, "/bin/bash", "-c", "cat /host/proc/cmdline")
Expand All @@ -107,16 +118,20 @@ var _ = Describe("[sriov] NetworkPool", Ordered, func() {

By("removing rdma mode configuration")
Eventually(func(g Gomega) {
err = clients.Get(context.Background(), client.ObjectKey{Name: testNode, Namespace: operatorNamespace}, networkPool)
g.Expect(err).ToNot(HaveOccurred())
err = clients.Delete(context.Background(), networkPool)
g.Expect(err).ToNot(HaveOccurred())
}, 5*time.Minute, 5*time.Second).Should(Succeed())

By("waiting for operator to finish the configuration")
WaitForSRIOVStable()

err = clients.Get(context.Background(), client.ObjectKey{Name: testNode, Namespace: operatorNamespace}, nodeState)
Expect(err).ToNot(HaveOccurred())
Expect(nodeState.Spec.System.RdmaMode).To(Equal(""))
Expect(nodeState.Status.System.RdmaMode).To(Equal(consts.RdmaSubsystemModeShared))
Eventually(func(g Gomega) {
err = clients.Get(context.Background(), client.ObjectKey{Name: testNode, Namespace: operatorNamespace}, nodeState)
g.Expect(err).ToNot(HaveOccurred())
g.Expect(nodeState.Spec.System.RdmaMode).To(Equal(""))
g.Expect(nodeState.Status.System.RdmaMode).To(Equal(consts.RdmaSubsystemModeShared))
}, 20*time.Minute, 5*time.Second).Should(Succeed())

By("Checking rdma mode and kernel args")
cmdlineOutput, _, err = runCommandOnConfigDaemon(testNode, "/bin/bash", "-c", "cat /host/proc/cmdline")
Expand Down Expand Up @@ -150,25 +165,6 @@ var _ = Describe("[sriov] NetworkPool", Ordered, func() {
Skip("no mellanox card available to test rdma")
}

networkPool := &sriovv1.SriovNetworkPoolConfig{
ObjectMeta: metav1.ObjectMeta{Name: testNode, Namespace: operatorNamespace},
Spec: sriovv1.SriovNetworkPoolConfigSpec{RdmaMode: consts.RdmaSubsystemModeExclusive,
NodeSelector: &metav1.LabelSelector{MatchLabels: map[string]string{"kubernetes.io/hostname": testNode}}}}

err = clients.Create(context.Background(), networkPool)
Expect(err).ToNot(HaveOccurred())
By("waiting for operator to finish the configuration")
WaitForSRIOVStable()
})

It("should run pod with RDMA cni and expose nic metrics and another one without rdma info", func() {
By("creating a policy")
resourceName := "testrdma"
_, err := network.CreateSriovPolicy(clients, "test-policy-", operatorNamespace, iface.Name, testNode, 5, resourceName, "netdevice",
func(policy *sriovv1.SriovNetworkNodePolicy) { policy.Spec.IsRdma = true })
Expect(err).ToNot(HaveOccurred())
WaitForSRIOVStable()

By("Creating sriov network to use the rdma device")
sriovNetwork := &sriovv1.SriovNetwork{
ObjectMeta: metav1.ObjectMeta{
Expand Down Expand Up @@ -201,6 +197,32 @@ var _ = Describe("[sriov] NetworkPool", Ordered, func() {
Expect(err).ToNot(HaveOccurred())
waitForNetAttachDef("test-nordmanetwork", namespaces.Test)

networkPool := &sriovv1.SriovNetworkPoolConfig{
ObjectMeta: metav1.ObjectMeta{Name: testNode, Namespace: operatorNamespace},
Spec: sriovv1.SriovNetworkPoolConfigSpec{RdmaMode: consts.RdmaSubsystemModeExclusive,
NodeSelector: &metav1.LabelSelector{MatchLabels: map[string]string{"kubernetes.io/hostname": testNode}}}}
err = clients.Create(context.Background(), networkPool)
Expect(err).ToNot(HaveOccurred())

By("waiting for operator to finish the configuration")
WaitForSRIOVStable()
nodeState := &sriovv1.SriovNetworkNodeState{}
Eventually(func(g Gomega) {
err = clients.Get(context.Background(), client.ObjectKey{Name: testNode, Namespace: operatorNamespace}, nodeState)
g.Expect(err).ToNot(HaveOccurred())
g.Expect(nodeState.Spec.System.RdmaMode).To(Equal(consts.RdmaSubsystemModeExclusive))
g.Expect(nodeState.Status.System.RdmaMode).To(Equal(consts.RdmaSubsystemModeExclusive))
}, 20*time.Minute, 5*time.Second).Should(Succeed())
})

It("should run pod with RDMA cni and expose nic metrics and another one without rdma info", func() {
By("creating a policy")
_, err := network.CreateSriovPolicy(clients, "test-policy-", operatorNamespace, iface.Name, testNode, 5, resourceName, "netdevice",
func(policy *sriovv1.SriovNetworkNodePolicy) { policy.Spec.IsRdma = true })
Expect(err).ToNot(HaveOccurred())

By("waiting for operator to finish the configuration")
WaitForSRIOVStable()
podDefinition := pod.DefineWithNetworks([]string{"test-rdmanetwork"})
firstPod, err := clients.Pods(namespaces.Test).Create(context.Background(), podDefinition, metav1.CreateOptions{})
Expect(err).ToNot(HaveOccurred())
Expand Down Expand Up @@ -287,6 +309,22 @@ var _ = Describe("[sriov] NetworkPool", Ordered, func() {
Skip("no mellanox card available to test rdma")
}

By("Creating sriov network to use the rdma device")
sriovNetwork := &sriovv1.SriovNetwork{
ObjectMeta: metav1.ObjectMeta{
Name: "test-rdmanetwork",
Namespace: operatorNamespace,
},
Spec: sriovv1.SriovNetworkSpec{
ResourceName: resourceName,
IPAM: `{"type":"host-local","subnet":"10.10.10.0/24","rangeStart":"10.10.10.171","rangeEnd":"10.10.10.181"}`,
NetworkNamespace: namespaces.Test,
}}

err = clients.Create(context.Background(), sriovNetwork)
Expect(err).ToNot(HaveOccurred())
waitForNetAttachDef("test-rdmanetwork", namespaces.Test)

networkPool := &sriovv1.SriovNetworkPoolConfig{
ObjectMeta: metav1.ObjectMeta{Name: testNode, Namespace: operatorNamespace},
Spec: sriovv1.SriovNetworkPoolConfigSpec{RdmaMode: consts.RdmaSubsystemModeShared,
Expand All @@ -296,32 +334,22 @@ var _ = Describe("[sriov] NetworkPool", Ordered, func() {
Expect(err).ToNot(HaveOccurred())
By("waiting for operator to finish the configuration")
WaitForSRIOVStable()
nodeState := &sriovv1.SriovNetworkNodeState{}
Eventually(func(g Gomega) {
err = clients.Get(context.Background(), client.ObjectKey{Name: testNode, Namespace: operatorNamespace}, nodeState)
g.Expect(err).ToNot(HaveOccurred())
g.Expect(nodeState.Spec.System.RdmaMode).To(Equal(consts.RdmaSubsystemModeShared))
g.Expect(nodeState.Status.System.RdmaMode).To(Equal(consts.RdmaSubsystemModeShared))
}, 20*time.Minute, 5*time.Second).Should(Succeed())
})

It("should run pod without RDMA cni and not expose nic metrics", func() {
By("creating a policy")
resourceName := "testrdma"
_, err := network.CreateSriovPolicy(clients, "test-policy-", operatorNamespace, iface.Name, testNode, 5, resourceName, "netdevice",
func(policy *sriovv1.SriovNetworkNodePolicy) { policy.Spec.IsRdma = true })
Expect(err).ToNot(HaveOccurred())
WaitForSRIOVStable()

By("Creating sriov network to use the rdma device")
sriovNetwork := &sriovv1.SriovNetwork{
ObjectMeta: metav1.ObjectMeta{
Name: "test-rdmanetwork",
Namespace: operatorNamespace,
},
Spec: sriovv1.SriovNetworkSpec{
ResourceName: resourceName,
IPAM: `{"type":"host-local","subnet":"10.10.10.0/24","rangeStart":"10.10.10.171","rangeEnd":"10.10.10.181"}`,
NetworkNamespace: namespaces.Test,
}}

err = clients.Create(context.Background(), sriovNetwork)
Expect(err).ToNot(HaveOccurred())
waitForNetAttachDef("test-rdmanetwork", namespaces.Test)

podDefinition := pod.DefineWithNetworks([]string{"test-rdmanetwork"})
firstPod, err := clients.Pods(namespaces.Test).Create(context.Background(), podDefinition, metav1.CreateOptions{})
Expect(err).ToNot(HaveOccurred())
Expand Down
Loading