open-nomad/plugins/csi/client_test.go

658 lines
17 KiB
Go
Raw Normal View History

CSI Plugin Registration (#6555) This changeset implements the initial registration and fingerprinting of CSI Plugins as part of #5378. At a high level, it introduces the following: * A `csi_plugin` stanza as part of a Nomad task configuration, to allow a task to expose that it is a plugin. * A new task runner hook: `csi_plugin_supervisor`. This hook does two things. When the `csi_plugin` stanza is detected, it will automatically configure the plugin task to receive bidirectional mounts to the CSI intermediary directory. At runtime, it will then perform an initial heartbeat of the plugin and handle submitting it to the new `dynamicplugins.Registry` for further use by the client, and then run a lightweight heartbeat loop that will emit task events when health changes. * The `dynamicplugins.Registry` for handling plugins that run as Nomad tasks, in contrast to the existing catalog that requires `go-plugin` type plugins and to know the plugin configuration in advance. * The `csimanager` which fingerprints CSI plugins, in a similar way to `drivermanager` and `devicemanager`. It currently only fingerprints the NodeID from the plugin, and assumes that all plugins are monolithic. Missing features * We do not use the live updates of the `dynamicplugin` registry in the `csimanager` yet. * We do not deregister the plugins from the client when they shutdown yet, they just become indefinitely marked as unhealthy. This is deliberate until we figure out how we should manage deploying new versions of plugins/transitioning them.
2019-10-22 13:20:26 +00:00
package csi
import (
"context"
2020-01-24 16:20:23 +00:00
"errors"
CSI Plugin Registration (#6555) This changeset implements the initial registration and fingerprinting of CSI Plugins as part of #5378. At a high level, it introduces the following: * A `csi_plugin` stanza as part of a Nomad task configuration, to allow a task to expose that it is a plugin. * A new task runner hook: `csi_plugin_supervisor`. This hook does two things. When the `csi_plugin` stanza is detected, it will automatically configure the plugin task to receive bidirectional mounts to the CSI intermediary directory. At runtime, it will then perform an initial heartbeat of the plugin and handle submitting it to the new `dynamicplugins.Registry` for further use by the client, and then run a lightweight heartbeat loop that will emit task events when health changes. * The `dynamicplugins.Registry` for handling plugins that run as Nomad tasks, in contrast to the existing catalog that requires `go-plugin` type plugins and to know the plugin configuration in advance. * The `csimanager` which fingerprints CSI plugins, in a similar way to `drivermanager` and `devicemanager`. It currently only fingerprints the NodeID from the plugin, and assumes that all plugins are monolithic. Missing features * We do not use the live updates of the `dynamicplugin` registry in the `csimanager` yet. * We do not deregister the plugins from the client when they shutdown yet, they just become indefinitely marked as unhealthy. This is deliberate until we figure out how we should manage deploying new versions of plugins/transitioning them.
2019-10-22 13:20:26 +00:00
"fmt"
"testing"
csipbv1 "github.com/container-storage-interface/spec/lib/go/csi"
"github.com/golang/protobuf/ptypes/wrappers"
fake "github.com/hashicorp/nomad/plugins/csi/testing"
"github.com/stretchr/testify/require"
)
2019-12-18 12:18:07 +00:00
func newTestClient() (*fake.IdentityClient, *fake.ControllerClient, *fake.NodeClient, CSIPlugin) {
ic := fake.NewIdentityClient()
cc := fake.NewControllerClient()
nc := fake.NewNodeClient()
CSI Plugin Registration (#6555) This changeset implements the initial registration and fingerprinting of CSI Plugins as part of #5378. At a high level, it introduces the following: * A `csi_plugin` stanza as part of a Nomad task configuration, to allow a task to expose that it is a plugin. * A new task runner hook: `csi_plugin_supervisor`. This hook does two things. When the `csi_plugin` stanza is detected, it will automatically configure the plugin task to receive bidirectional mounts to the CSI intermediary directory. At runtime, it will then perform an initial heartbeat of the plugin and handle submitting it to the new `dynamicplugins.Registry` for further use by the client, and then run a lightweight heartbeat loop that will emit task events when health changes. * The `dynamicplugins.Registry` for handling plugins that run as Nomad tasks, in contrast to the existing catalog that requires `go-plugin` type plugins and to know the plugin configuration in advance. * The `csimanager` which fingerprints CSI plugins, in a similar way to `drivermanager` and `devicemanager`. It currently only fingerprints the NodeID from the plugin, and assumes that all plugins are monolithic. Missing features * We do not use the live updates of the `dynamicplugin` registry in the `csimanager` yet. * We do not deregister the plugins from the client when they shutdown yet, they just become indefinitely marked as unhealthy. This is deliberate until we figure out how we should manage deploying new versions of plugins/transitioning them.
2019-10-22 13:20:26 +00:00
client := &client{
identityClient: ic,
controllerClient: cc,
2019-12-18 12:18:07 +00:00
nodeClient: nc,
CSI Plugin Registration (#6555) This changeset implements the initial registration and fingerprinting of CSI Plugins as part of #5378. At a high level, it introduces the following: * A `csi_plugin` stanza as part of a Nomad task configuration, to allow a task to expose that it is a plugin. * A new task runner hook: `csi_plugin_supervisor`. This hook does two things. When the `csi_plugin` stanza is detected, it will automatically configure the plugin task to receive bidirectional mounts to the CSI intermediary directory. At runtime, it will then perform an initial heartbeat of the plugin and handle submitting it to the new `dynamicplugins.Registry` for further use by the client, and then run a lightweight heartbeat loop that will emit task events when health changes. * The `dynamicplugins.Registry` for handling plugins that run as Nomad tasks, in contrast to the existing catalog that requires `go-plugin` type plugins and to know the plugin configuration in advance. * The `csimanager` which fingerprints CSI plugins, in a similar way to `drivermanager` and `devicemanager`. It currently only fingerprints the NodeID from the plugin, and assumes that all plugins are monolithic. Missing features * We do not use the live updates of the `dynamicplugin` registry in the `csimanager` yet. * We do not deregister the plugins from the client when they shutdown yet, they just become indefinitely marked as unhealthy. This is deliberate until we figure out how we should manage deploying new versions of plugins/transitioning them.
2019-10-22 13:20:26 +00:00
}
2019-12-18 12:18:07 +00:00
return ic, cc, nc, client
CSI Plugin Registration (#6555) This changeset implements the initial registration and fingerprinting of CSI Plugins as part of #5378. At a high level, it introduces the following: * A `csi_plugin` stanza as part of a Nomad task configuration, to allow a task to expose that it is a plugin. * A new task runner hook: `csi_plugin_supervisor`. This hook does two things. When the `csi_plugin` stanza is detected, it will automatically configure the plugin task to receive bidirectional mounts to the CSI intermediary directory. At runtime, it will then perform an initial heartbeat of the plugin and handle submitting it to the new `dynamicplugins.Registry` for further use by the client, and then run a lightweight heartbeat loop that will emit task events when health changes. * The `dynamicplugins.Registry` for handling plugins that run as Nomad tasks, in contrast to the existing catalog that requires `go-plugin` type plugins and to know the plugin configuration in advance. * The `csimanager` which fingerprints CSI plugins, in a similar way to `drivermanager` and `devicemanager`. It currently only fingerprints the NodeID from the plugin, and assumes that all plugins are monolithic. Missing features * We do not use the live updates of the `dynamicplugin` registry in the `csimanager` yet. * We do not deregister the plugins from the client when they shutdown yet, they just become indefinitely marked as unhealthy. This is deliberate until we figure out how we should manage deploying new versions of plugins/transitioning them.
2019-10-22 13:20:26 +00:00
}
func TestClient_RPC_PluginProbe(t *testing.T) {
cases := []struct {
Name string
ResponseErr error
ProbeResponse *csipbv1.ProbeResponse
ExpectedResponse bool
ExpectedErr error
}{
{
Name: "handles underlying grpc errors",
ResponseErr: fmt.Errorf("some grpc error"),
ExpectedErr: fmt.Errorf("some grpc error"),
},
{
Name: "returns false for ready when the provider returns false",
ProbeResponse: &csipbv1.ProbeResponse{
Ready: &wrappers.BoolValue{Value: false},
},
ExpectedResponse: false,
},
{
Name: "returns true for ready when the provider returns true",
ProbeResponse: &csipbv1.ProbeResponse{
Ready: &wrappers.BoolValue{Value: true},
},
ExpectedResponse: true,
},
{
/* When a SP does not return a ready value, a CO MAY treat this as ready.
We do so because example plugins rely on this behaviour. We may
re-evaluate this decision in the future. */
Name: "returns true for ready when the provider returns a nil wrapper",
ProbeResponse: &csipbv1.ProbeResponse{
Ready: nil,
},
ExpectedResponse: true,
},
}
for _, c := range cases {
t.Run(c.Name, func(t *testing.T) {
2019-12-18 12:18:07 +00:00
ic, _, _, client := newTestClient()
CSI Plugin Registration (#6555) This changeset implements the initial registration and fingerprinting of CSI Plugins as part of #5378. At a high level, it introduces the following: * A `csi_plugin` stanza as part of a Nomad task configuration, to allow a task to expose that it is a plugin. * A new task runner hook: `csi_plugin_supervisor`. This hook does two things. When the `csi_plugin` stanza is detected, it will automatically configure the plugin task to receive bidirectional mounts to the CSI intermediary directory. At runtime, it will then perform an initial heartbeat of the plugin and handle submitting it to the new `dynamicplugins.Registry` for further use by the client, and then run a lightweight heartbeat loop that will emit task events when health changes. * The `dynamicplugins.Registry` for handling plugins that run as Nomad tasks, in contrast to the existing catalog that requires `go-plugin` type plugins and to know the plugin configuration in advance. * The `csimanager` which fingerprints CSI plugins, in a similar way to `drivermanager` and `devicemanager`. It currently only fingerprints the NodeID from the plugin, and assumes that all plugins are monolithic. Missing features * We do not use the live updates of the `dynamicplugin` registry in the `csimanager` yet. * We do not deregister the plugins from the client when they shutdown yet, they just become indefinitely marked as unhealthy. This is deliberate until we figure out how we should manage deploying new versions of plugins/transitioning them.
2019-10-22 13:20:26 +00:00
defer client.Close()
ic.NextErr = c.ResponseErr
ic.NextPluginProbe = c.ProbeResponse
resp, err := client.PluginProbe(context.TODO())
if c.ExpectedErr != nil {
require.Error(t, c.ExpectedErr, err)
}
require.Equal(t, c.ExpectedResponse, resp)
})
}
}
func TestClient_RPC_PluginInfo(t *testing.T) {
cases := []struct {
Name string
ResponseErr error
InfoResponse *csipbv1.GetPluginInfoResponse
ExpectedResponseName string
ExpectedResponseVersion string
ExpectedErr error
CSI Plugin Registration (#6555) This changeset implements the initial registration and fingerprinting of CSI Plugins as part of #5378. At a high level, it introduces the following: * A `csi_plugin` stanza as part of a Nomad task configuration, to allow a task to expose that it is a plugin. * A new task runner hook: `csi_plugin_supervisor`. This hook does two things. When the `csi_plugin` stanza is detected, it will automatically configure the plugin task to receive bidirectional mounts to the CSI intermediary directory. At runtime, it will then perform an initial heartbeat of the plugin and handle submitting it to the new `dynamicplugins.Registry` for further use by the client, and then run a lightweight heartbeat loop that will emit task events when health changes. * The `dynamicplugins.Registry` for handling plugins that run as Nomad tasks, in contrast to the existing catalog that requires `go-plugin` type plugins and to know the plugin configuration in advance. * The `csimanager` which fingerprints CSI plugins, in a similar way to `drivermanager` and `devicemanager`. It currently only fingerprints the NodeID from the plugin, and assumes that all plugins are monolithic. Missing features * We do not use the live updates of the `dynamicplugin` registry in the `csimanager` yet. * We do not deregister the plugins from the client when they shutdown yet, they just become indefinitely marked as unhealthy. This is deliberate until we figure out how we should manage deploying new versions of plugins/transitioning them.
2019-10-22 13:20:26 +00:00
}{
{
Name: "handles underlying grpc errors",
ResponseErr: fmt.Errorf("some grpc error"),
ExpectedErr: fmt.Errorf("some grpc error"),
},
{
Name: "returns an error if we receive an empty `name`",
InfoResponse: &csipbv1.GetPluginInfoResponse{
Name: "",
VendorVersion: "",
CSI Plugin Registration (#6555) This changeset implements the initial registration and fingerprinting of CSI Plugins as part of #5378. At a high level, it introduces the following: * A `csi_plugin` stanza as part of a Nomad task configuration, to allow a task to expose that it is a plugin. * A new task runner hook: `csi_plugin_supervisor`. This hook does two things. When the `csi_plugin` stanza is detected, it will automatically configure the plugin task to receive bidirectional mounts to the CSI intermediary directory. At runtime, it will then perform an initial heartbeat of the plugin and handle submitting it to the new `dynamicplugins.Registry` for further use by the client, and then run a lightweight heartbeat loop that will emit task events when health changes. * The `dynamicplugins.Registry` for handling plugins that run as Nomad tasks, in contrast to the existing catalog that requires `go-plugin` type plugins and to know the plugin configuration in advance. * The `csimanager` which fingerprints CSI plugins, in a similar way to `drivermanager` and `devicemanager`. It currently only fingerprints the NodeID from the plugin, and assumes that all plugins are monolithic. Missing features * We do not use the live updates of the `dynamicplugin` registry in the `csimanager` yet. * We do not deregister the plugins from the client when they shutdown yet, they just become indefinitely marked as unhealthy. This is deliberate until we figure out how we should manage deploying new versions of plugins/transitioning them.
2019-10-22 13:20:26 +00:00
},
ExpectedErr: fmt.Errorf("PluginGetInfo: plugin returned empty name field"),
},
{
Name: "returns the name when successfully retrieved and not empty",
InfoResponse: &csipbv1.GetPluginInfoResponse{
Name: "com.hashicorp.storage",
VendorVersion: "1.0.1",
CSI Plugin Registration (#6555) This changeset implements the initial registration and fingerprinting of CSI Plugins as part of #5378. At a high level, it introduces the following: * A `csi_plugin` stanza as part of a Nomad task configuration, to allow a task to expose that it is a plugin. * A new task runner hook: `csi_plugin_supervisor`. This hook does two things. When the `csi_plugin` stanza is detected, it will automatically configure the plugin task to receive bidirectional mounts to the CSI intermediary directory. At runtime, it will then perform an initial heartbeat of the plugin and handle submitting it to the new `dynamicplugins.Registry` for further use by the client, and then run a lightweight heartbeat loop that will emit task events when health changes. * The `dynamicplugins.Registry` for handling plugins that run as Nomad tasks, in contrast to the existing catalog that requires `go-plugin` type plugins and to know the plugin configuration in advance. * The `csimanager` which fingerprints CSI plugins, in a similar way to `drivermanager` and `devicemanager`. It currently only fingerprints the NodeID from the plugin, and assumes that all plugins are monolithic. Missing features * We do not use the live updates of the `dynamicplugin` registry in the `csimanager` yet. * We do not deregister the plugins from the client when they shutdown yet, they just become indefinitely marked as unhealthy. This is deliberate until we figure out how we should manage deploying new versions of plugins/transitioning them.
2019-10-22 13:20:26 +00:00
},
ExpectedResponseName: "com.hashicorp.storage",
ExpectedResponseVersion: "1.0.1",
CSI Plugin Registration (#6555) This changeset implements the initial registration and fingerprinting of CSI Plugins as part of #5378. At a high level, it introduces the following: * A `csi_plugin` stanza as part of a Nomad task configuration, to allow a task to expose that it is a plugin. * A new task runner hook: `csi_plugin_supervisor`. This hook does two things. When the `csi_plugin` stanza is detected, it will automatically configure the plugin task to receive bidirectional mounts to the CSI intermediary directory. At runtime, it will then perform an initial heartbeat of the plugin and handle submitting it to the new `dynamicplugins.Registry` for further use by the client, and then run a lightweight heartbeat loop that will emit task events when health changes. * The `dynamicplugins.Registry` for handling plugins that run as Nomad tasks, in contrast to the existing catalog that requires `go-plugin` type plugins and to know the plugin configuration in advance. * The `csimanager` which fingerprints CSI plugins, in a similar way to `drivermanager` and `devicemanager`. It currently only fingerprints the NodeID from the plugin, and assumes that all plugins are monolithic. Missing features * We do not use the live updates of the `dynamicplugin` registry in the `csimanager` yet. * We do not deregister the plugins from the client when they shutdown yet, they just become indefinitely marked as unhealthy. This is deliberate until we figure out how we should manage deploying new versions of plugins/transitioning them.
2019-10-22 13:20:26 +00:00
},
}
for _, c := range cases {
t.Run(c.Name, func(t *testing.T) {
2019-12-18 12:18:07 +00:00
ic, _, _, client := newTestClient()
CSI Plugin Registration (#6555) This changeset implements the initial registration and fingerprinting of CSI Plugins as part of #5378. At a high level, it introduces the following: * A `csi_plugin` stanza as part of a Nomad task configuration, to allow a task to expose that it is a plugin. * A new task runner hook: `csi_plugin_supervisor`. This hook does two things. When the `csi_plugin` stanza is detected, it will automatically configure the plugin task to receive bidirectional mounts to the CSI intermediary directory. At runtime, it will then perform an initial heartbeat of the plugin and handle submitting it to the new `dynamicplugins.Registry` for further use by the client, and then run a lightweight heartbeat loop that will emit task events when health changes. * The `dynamicplugins.Registry` for handling plugins that run as Nomad tasks, in contrast to the existing catalog that requires `go-plugin` type plugins and to know the plugin configuration in advance. * The `csimanager` which fingerprints CSI plugins, in a similar way to `drivermanager` and `devicemanager`. It currently only fingerprints the NodeID from the plugin, and assumes that all plugins are monolithic. Missing features * We do not use the live updates of the `dynamicplugin` registry in the `csimanager` yet. * We do not deregister the plugins from the client when they shutdown yet, they just become indefinitely marked as unhealthy. This is deliberate until we figure out how we should manage deploying new versions of plugins/transitioning them.
2019-10-22 13:20:26 +00:00
defer client.Close()
ic.NextErr = c.ResponseErr
ic.NextPluginInfo = c.InfoResponse
name, version, err := client.PluginGetInfo(context.TODO())
CSI Plugin Registration (#6555) This changeset implements the initial registration and fingerprinting of CSI Plugins as part of #5378. At a high level, it introduces the following: * A `csi_plugin` stanza as part of a Nomad task configuration, to allow a task to expose that it is a plugin. * A new task runner hook: `csi_plugin_supervisor`. This hook does two things. When the `csi_plugin` stanza is detected, it will automatically configure the plugin task to receive bidirectional mounts to the CSI intermediary directory. At runtime, it will then perform an initial heartbeat of the plugin and handle submitting it to the new `dynamicplugins.Registry` for further use by the client, and then run a lightweight heartbeat loop that will emit task events when health changes. * The `dynamicplugins.Registry` for handling plugins that run as Nomad tasks, in contrast to the existing catalog that requires `go-plugin` type plugins and to know the plugin configuration in advance. * The `csimanager` which fingerprints CSI plugins, in a similar way to `drivermanager` and `devicemanager`. It currently only fingerprints the NodeID from the plugin, and assumes that all plugins are monolithic. Missing features * We do not use the live updates of the `dynamicplugin` registry in the `csimanager` yet. * We do not deregister the plugins from the client when they shutdown yet, they just become indefinitely marked as unhealthy. This is deliberate until we figure out how we should manage deploying new versions of plugins/transitioning them.
2019-10-22 13:20:26 +00:00
if c.ExpectedErr != nil {
require.Error(t, c.ExpectedErr, err)
}
require.Equal(t, c.ExpectedResponseName, name)
require.Equal(t, c.ExpectedResponseVersion, version)
CSI Plugin Registration (#6555) This changeset implements the initial registration and fingerprinting of CSI Plugins as part of #5378. At a high level, it introduces the following: * A `csi_plugin` stanza as part of a Nomad task configuration, to allow a task to expose that it is a plugin. * A new task runner hook: `csi_plugin_supervisor`. This hook does two things. When the `csi_plugin` stanza is detected, it will automatically configure the plugin task to receive bidirectional mounts to the CSI intermediary directory. At runtime, it will then perform an initial heartbeat of the plugin and handle submitting it to the new `dynamicplugins.Registry` for further use by the client, and then run a lightweight heartbeat loop that will emit task events when health changes. * The `dynamicplugins.Registry` for handling plugins that run as Nomad tasks, in contrast to the existing catalog that requires `go-plugin` type plugins and to know the plugin configuration in advance. * The `csimanager` which fingerprints CSI plugins, in a similar way to `drivermanager` and `devicemanager`. It currently only fingerprints the NodeID from the plugin, and assumes that all plugins are monolithic. Missing features * We do not use the live updates of the `dynamicplugin` registry in the `csimanager` yet. * We do not deregister the plugins from the client when they shutdown yet, they just become indefinitely marked as unhealthy. This is deliberate until we figure out how we should manage deploying new versions of plugins/transitioning them.
2019-10-22 13:20:26 +00:00
})
}
}
func TestClient_RPC_PluginGetCapabilities(t *testing.T) {
cases := []struct {
Name string
ResponseErr error
Response *csipbv1.GetPluginCapabilitiesResponse
ExpectedResponse *PluginCapabilitySet
ExpectedErr error
}{
{
Name: "handles underlying grpc errors",
ResponseErr: fmt.Errorf("some grpc error"),
ExpectedErr: fmt.Errorf("some grpc error"),
},
{
Name: "HasControllerService is true when it's part of the response",
Response: &csipbv1.GetPluginCapabilitiesResponse{
Capabilities: []*csipbv1.PluginCapability{
{
Type: &csipbv1.PluginCapability_Service_{
Service: &csipbv1.PluginCapability_Service{
Type: csipbv1.PluginCapability_Service_CONTROLLER_SERVICE,
},
},
},
},
},
ExpectedResponse: &PluginCapabilitySet{hasControllerService: true},
},
{
Name: "HasTopologies is true when it's part of the response",
Response: &csipbv1.GetPluginCapabilitiesResponse{
Capabilities: []*csipbv1.PluginCapability{
{
Type: &csipbv1.PluginCapability_Service_{
Service: &csipbv1.PluginCapability_Service{
Type: csipbv1.PluginCapability_Service_VOLUME_ACCESSIBILITY_CONSTRAINTS,
},
},
},
},
},
ExpectedResponse: &PluginCapabilitySet{hasTopologies: true},
},
}
for _, c := range cases {
t.Run(c.Name, func(t *testing.T) {
2019-12-18 12:18:07 +00:00
ic, _, _, client := newTestClient()
CSI Plugin Registration (#6555) This changeset implements the initial registration and fingerprinting of CSI Plugins as part of #5378. At a high level, it introduces the following: * A `csi_plugin` stanza as part of a Nomad task configuration, to allow a task to expose that it is a plugin. * A new task runner hook: `csi_plugin_supervisor`. This hook does two things. When the `csi_plugin` stanza is detected, it will automatically configure the plugin task to receive bidirectional mounts to the CSI intermediary directory. At runtime, it will then perform an initial heartbeat of the plugin and handle submitting it to the new `dynamicplugins.Registry` for further use by the client, and then run a lightweight heartbeat loop that will emit task events when health changes. * The `dynamicplugins.Registry` for handling plugins that run as Nomad tasks, in contrast to the existing catalog that requires `go-plugin` type plugins and to know the plugin configuration in advance. * The `csimanager` which fingerprints CSI plugins, in a similar way to `drivermanager` and `devicemanager`. It currently only fingerprints the NodeID from the plugin, and assumes that all plugins are monolithic. Missing features * We do not use the live updates of the `dynamicplugin` registry in the `csimanager` yet. * We do not deregister the plugins from the client when they shutdown yet, they just become indefinitely marked as unhealthy. This is deliberate until we figure out how we should manage deploying new versions of plugins/transitioning them.
2019-10-22 13:20:26 +00:00
defer client.Close()
ic.NextErr = c.ResponseErr
ic.NextPluginCapabilities = c.Response
resp, err := client.PluginGetCapabilities(context.TODO())
if c.ExpectedErr != nil {
require.Error(t, c.ExpectedErr, err)
}
require.Equal(t, c.ExpectedResponse, resp)
})
}
}
CSI Plugin Registration (#6555) This changeset implements the initial registration and fingerprinting of CSI Plugins as part of #5378. At a high level, it introduces the following: * A `csi_plugin` stanza as part of a Nomad task configuration, to allow a task to expose that it is a plugin. * A new task runner hook: `csi_plugin_supervisor`. This hook does two things. When the `csi_plugin` stanza is detected, it will automatically configure the plugin task to receive bidirectional mounts to the CSI intermediary directory. At runtime, it will then perform an initial heartbeat of the plugin and handle submitting it to the new `dynamicplugins.Registry` for further use by the client, and then run a lightweight heartbeat loop that will emit task events when health changes. * The `dynamicplugins.Registry` for handling plugins that run as Nomad tasks, in contrast to the existing catalog that requires `go-plugin` type plugins and to know the plugin configuration in advance. * The `csimanager` which fingerprints CSI plugins, in a similar way to `drivermanager` and `devicemanager`. It currently only fingerprints the NodeID from the plugin, and assumes that all plugins are monolithic. Missing features * We do not use the live updates of the `dynamicplugin` registry in the `csimanager` yet. * We do not deregister the plugins from the client when they shutdown yet, they just become indefinitely marked as unhealthy. This is deliberate until we figure out how we should manage deploying new versions of plugins/transitioning them.
2019-10-22 13:20:26 +00:00
func TestClient_RPC_ControllerGetCapabilities(t *testing.T) {
cases := []struct {
Name string
ResponseErr error
Response *csipbv1.ControllerGetCapabilitiesResponse
ExpectedResponse *ControllerCapabilitySet
ExpectedErr error
}{
{
Name: "handles underlying grpc errors",
ResponseErr: fmt.Errorf("some grpc error"),
ExpectedErr: fmt.Errorf("some grpc error"),
},
{
Name: "ignores unknown capabilities",
Response: &csipbv1.ControllerGetCapabilitiesResponse{
Capabilities: []*csipbv1.ControllerServiceCapability{
{
Type: &csipbv1.ControllerServiceCapability_Rpc{
Rpc: &csipbv1.ControllerServiceCapability_RPC{
Type: csipbv1.ControllerServiceCapability_RPC_GET_CAPACITY,
},
},
},
},
},
ExpectedResponse: &ControllerCapabilitySet{},
},
{
Name: "detects list volumes capabilities",
Response: &csipbv1.ControllerGetCapabilitiesResponse{
Capabilities: []*csipbv1.ControllerServiceCapability{
{
Type: &csipbv1.ControllerServiceCapability_Rpc{
Rpc: &csipbv1.ControllerServiceCapability_RPC{
Type: csipbv1.ControllerServiceCapability_RPC_LIST_VOLUMES,
},
},
},
{
Type: &csipbv1.ControllerServiceCapability_Rpc{
Rpc: &csipbv1.ControllerServiceCapability_RPC{
Type: csipbv1.ControllerServiceCapability_RPC_LIST_VOLUMES_PUBLISHED_NODES,
},
},
},
},
},
ExpectedResponse: &ControllerCapabilitySet{
HasListVolumes: true,
HasListVolumesPublishedNodes: true,
},
},
{
Name: "detects publish capabilities",
Response: &csipbv1.ControllerGetCapabilitiesResponse{
Capabilities: []*csipbv1.ControllerServiceCapability{
{
Type: &csipbv1.ControllerServiceCapability_Rpc{
Rpc: &csipbv1.ControllerServiceCapability_RPC{
Type: csipbv1.ControllerServiceCapability_RPC_PUBLISH_READONLY,
},
},
},
{
Type: &csipbv1.ControllerServiceCapability_Rpc{
Rpc: &csipbv1.ControllerServiceCapability_RPC{
Type: csipbv1.ControllerServiceCapability_RPC_PUBLISH_UNPUBLISH_VOLUME,
},
},
},
},
},
ExpectedResponse: &ControllerCapabilitySet{
HasPublishUnpublishVolume: true,
HasPublishReadonly: true,
},
},
}
for _, tc := range cases {
t.Run(tc.Name, func(t *testing.T) {
2019-12-18 12:18:07 +00:00
_, cc, _, client := newTestClient()
defer client.Close()
cc.NextErr = tc.ResponseErr
cc.NextCapabilitiesResponse = tc.Response
resp, err := client.ControllerGetCapabilities(context.TODO())
if tc.ExpectedErr != nil {
require.Error(t, tc.ExpectedErr, err)
}
require.Equal(t, tc.ExpectedResponse, resp)
})
}
}
2019-12-18 12:18:07 +00:00
func TestClient_RPC_NodeGetCapabilities(t *testing.T) {
cases := []struct {
Name string
ResponseErr error
Response *csipbv1.NodeGetCapabilitiesResponse
ExpectedResponse *NodeCapabilitySet
ExpectedErr error
}{
{
Name: "handles underlying grpc errors",
ResponseErr: fmt.Errorf("some grpc error"),
ExpectedErr: fmt.Errorf("some grpc error"),
},
{
Name: "ignores unknown capabilities",
Response: &csipbv1.NodeGetCapabilitiesResponse{
Capabilities: []*csipbv1.NodeServiceCapability{
{
Type: &csipbv1.NodeServiceCapability_Rpc{
Rpc: &csipbv1.NodeServiceCapability_RPC{
Type: csipbv1.NodeServiceCapability_RPC_EXPAND_VOLUME,
},
},
},
},
},
ExpectedResponse: &NodeCapabilitySet{},
},
{
Name: "detects stage volumes capability",
Response: &csipbv1.NodeGetCapabilitiesResponse{
Capabilities: []*csipbv1.NodeServiceCapability{
{
Type: &csipbv1.NodeServiceCapability_Rpc{
Rpc: &csipbv1.NodeServiceCapability_RPC{
Type: csipbv1.NodeServiceCapability_RPC_STAGE_UNSTAGE_VOLUME,
},
},
},
},
},
ExpectedResponse: &NodeCapabilitySet{
HasStageUnstageVolume: true,
},
},
}
for _, tc := range cases {
t.Run(tc.Name, func(t *testing.T) {
_, _, nc, client := newTestClient()
defer client.Close()
nc.NextErr = tc.ResponseErr
nc.NextCapabilitiesResponse = tc.Response
resp, err := client.NodeGetCapabilities(context.TODO())
if tc.ExpectedErr != nil {
require.Error(t, tc.ExpectedErr, err)
}
require.Equal(t, tc.ExpectedResponse, resp)
})
}
}
func TestClient_RPC_ControllerPublishVolume(t *testing.T) {
cases := []struct {
Name string
Request *ControllerPublishVolumeRequest
ResponseErr error
Response *csipbv1.ControllerPublishVolumeResponse
ExpectedResponse *ControllerPublishVolumeResponse
ExpectedErr error
}{
{
Name: "handles underlying grpc errors",
Request: &ControllerPublishVolumeRequest{},
ResponseErr: fmt.Errorf("some grpc error"),
ExpectedErr: fmt.Errorf("some grpc error"),
},
{
Name: "Handles missing NodeID",
Request: &ControllerPublishVolumeRequest{},
Response: &csipbv1.ControllerPublishVolumeResponse{},
ExpectedErr: fmt.Errorf("missing NodeID"),
},
{
Name: "Handles PublishContext == nil",
Request: &ControllerPublishVolumeRequest{VolumeID: "vol", NodeID: "node"},
Response: &csipbv1.ControllerPublishVolumeResponse{},
ExpectedResponse: &ControllerPublishVolumeResponse{},
},
{
Name: "Handles PublishContext != nil",
Request: &ControllerPublishVolumeRequest{VolumeID: "vol", NodeID: "node"},
Response: &csipbv1.ControllerPublishVolumeResponse{
PublishContext: map[string]string{
"com.hashicorp/nomad-node-id": "foobar",
"com.plugin/device": "/dev/sdc1",
},
},
ExpectedResponse: &ControllerPublishVolumeResponse{
PublishContext: map[string]string{
"com.hashicorp/nomad-node-id": "foobar",
"com.plugin/device": "/dev/sdc1",
},
},
},
}
for _, c := range cases {
t.Run(c.Name, func(t *testing.T) {
2019-12-18 12:18:07 +00:00
_, cc, _, client := newTestClient()
defer client.Close()
cc.NextErr = c.ResponseErr
cc.NextPublishVolumeResponse = c.Response
resp, err := client.ControllerPublishVolume(context.TODO(), c.Request)
if c.ExpectedErr != nil {
require.Error(t, c.ExpectedErr, err)
}
require.Equal(t, c.ExpectedResponse, resp)
})
}
}
func TestClient_RPC_ControllerUnpublishVolume(t *testing.T) {
cases := []struct {
Name string
Request *ControllerUnpublishVolumeRequest
ResponseErr error
Response *csipbv1.ControllerUnpublishVolumeResponse
ExpectedResponse *ControllerUnpublishVolumeResponse
ExpectedErr error
}{
{
Name: "Handles underlying grpc errors",
Request: &ControllerUnpublishVolumeRequest{},
ResponseErr: fmt.Errorf("some grpc error"),
ExpectedErr: fmt.Errorf("some grpc error"),
},
{
Name: "Handles missing NodeID",
Request: &ControllerUnpublishVolumeRequest{},
ExpectedErr: fmt.Errorf("missing NodeID"),
ExpectedResponse: nil,
},
{
Name: "Handles successful response",
Request: &ControllerUnpublishVolumeRequest{VolumeID: "vol", NodeID: "node"},
ExpectedErr: fmt.Errorf("missing NodeID"),
ExpectedResponse: &ControllerUnpublishVolumeResponse{},
},
}
for _, c := range cases {
t.Run(c.Name, func(t *testing.T) {
_, cc, _, client := newTestClient()
defer client.Close()
cc.NextErr = c.ResponseErr
cc.NextUnpublishVolumeResponse = c.Response
resp, err := client.ControllerUnpublishVolume(context.TODO(), c.Request)
if c.ExpectedErr != nil {
require.Error(t, c.ExpectedErr, err)
}
require.Equal(t, c.ExpectedResponse, resp)
})
}
CSI Plugin Registration (#6555) This changeset implements the initial registration and fingerprinting of CSI Plugins as part of #5378. At a high level, it introduces the following: * A `csi_plugin` stanza as part of a Nomad task configuration, to allow a task to expose that it is a plugin. * A new task runner hook: `csi_plugin_supervisor`. This hook does two things. When the `csi_plugin` stanza is detected, it will automatically configure the plugin task to receive bidirectional mounts to the CSI intermediary directory. At runtime, it will then perform an initial heartbeat of the plugin and handle submitting it to the new `dynamicplugins.Registry` for further use by the client, and then run a lightweight heartbeat loop that will emit task events when health changes. * The `dynamicplugins.Registry` for handling plugins that run as Nomad tasks, in contrast to the existing catalog that requires `go-plugin` type plugins and to know the plugin configuration in advance. * The `csimanager` which fingerprints CSI plugins, in a similar way to `drivermanager` and `devicemanager`. It currently only fingerprints the NodeID from the plugin, and assumes that all plugins are monolithic. Missing features * We do not use the live updates of the `dynamicplugin` registry in the `csimanager` yet. * We do not deregister the plugins from the client when they shutdown yet, they just become indefinitely marked as unhealthy. This is deliberate until we figure out how we should manage deploying new versions of plugins/transitioning them.
2019-10-22 13:20:26 +00:00
}
2020-01-23 11:24:37 +00:00
func TestClient_RPC_NodeStageVolume(t *testing.T) {
cases := []struct {
Name string
ResponseErr error
Response *csipbv1.NodeStageVolumeResponse
ExpectedErr error
}{
{
Name: "handles underlying grpc errors",
ResponseErr: fmt.Errorf("some grpc error"),
ExpectedErr: fmt.Errorf("some grpc error"),
},
{
Name: "handles success",
ResponseErr: nil,
ExpectedErr: nil,
},
}
for _, c := range cases {
t.Run(c.Name, func(t *testing.T) {
_, _, nc, client := newTestClient()
defer client.Close()
nc.NextErr = c.ResponseErr
nc.NextStageVolumeResponse = c.Response
err := client.NodeStageVolume(context.TODO(), "foo", nil, "/foo", &VolumeCapability{})
if c.ExpectedErr != nil {
require.Error(t, c.ExpectedErr, err)
} else {
require.Nil(t, err)
}
})
}
}
func TestClient_RPC_NodeUnstageVolume(t *testing.T) {
cases := []struct {
Name string
ResponseErr error
Response *csipbv1.NodeUnstageVolumeResponse
ExpectedErr error
}{
{
Name: "handles underlying grpc errors",
ResponseErr: fmt.Errorf("some grpc error"),
ExpectedErr: fmt.Errorf("some grpc error"),
},
{
Name: "handles success",
ResponseErr: nil,
ExpectedErr: nil,
},
}
for _, c := range cases {
t.Run(c.Name, func(t *testing.T) {
_, _, nc, client := newTestClient()
defer client.Close()
nc.NextErr = c.ResponseErr
nc.NextUnstageVolumeResponse = c.Response
err := client.NodeUnstageVolume(context.TODO(), "foo", "/foo")
if c.ExpectedErr != nil {
require.Error(t, c.ExpectedErr, err)
} else {
require.Nil(t, err)
}
})
}
}
2020-01-24 16:20:23 +00:00
func TestClient_RPC_NodePublishVolume(t *testing.T) {
cases := []struct {
Name string
Request *NodePublishVolumeRequest
ResponseErr error
Response *csipbv1.NodePublishVolumeResponse
ExpectedErr error
}{
{
Name: "handles underlying grpc errors",
Request: &NodePublishVolumeRequest{
VolumeID: "foo",
TargetPath: "/dev/null",
VolumeCapability: &VolumeCapability{},
},
ResponseErr: fmt.Errorf("some grpc error"),
ExpectedErr: fmt.Errorf("some grpc error"),
},
{
Name: "handles success",
Request: &NodePublishVolumeRequest{
VolumeID: "foo",
TargetPath: "/dev/null",
VolumeCapability: &VolumeCapability{},
},
ResponseErr: nil,
ExpectedErr: nil,
},
{
Name: "Performs validation of the publish volume request",
Request: &NodePublishVolumeRequest{
VolumeID: "",
},
ResponseErr: nil,
ExpectedErr: errors.New("missing VolumeID"),
},
}
for _, c := range cases {
t.Run(c.Name, func(t *testing.T) {
_, _, nc, client := newTestClient()
defer client.Close()
nc.NextErr = c.ResponseErr
nc.NextPublishVolumeResponse = c.Response
err := client.NodePublishVolume(context.TODO(), c.Request)
if c.ExpectedErr != nil {
require.Error(t, c.ExpectedErr, err)
} else {
require.Nil(t, err)
}
})
}
}
func TestClient_RPC_NodeUnpublishVolume(t *testing.T) {
cases := []struct {
Name string
VolumeID string
TargetPath string
ResponseErr error
Response *csipbv1.NodeUnpublishVolumeResponse
ExpectedErr error
}{
{
Name: "handles underlying grpc errors",
VolumeID: "foo",
TargetPath: "/dev/null",
ResponseErr: fmt.Errorf("some grpc error"),
ExpectedErr: fmt.Errorf("some grpc error"),
},
{
Name: "handles success",
VolumeID: "foo",
TargetPath: "/dev/null",
ResponseErr: nil,
ExpectedErr: nil,
},
{
Name: "Performs validation of the request args - VolumeID",
ResponseErr: nil,
ExpectedErr: errors.New("missing VolumeID"),
},
{
Name: "Performs validation of the request args - TargetPath",
VolumeID: "foo",
ResponseErr: nil,
ExpectedErr: errors.New("missing TargetPath"),
},
}
for _, c := range cases {
t.Run(c.Name, func(t *testing.T) {
_, _, nc, client := newTestClient()
defer client.Close()
nc.NextErr = c.ResponseErr
nc.NextUnpublishVolumeResponse = c.Response
err := client.NodeUnpublishVolume(context.TODO(), c.VolumeID, c.TargetPath)
if c.ExpectedErr != nil {
require.Error(t, c.ExpectedErr, err)
} else {
require.Nil(t, err)
}
})
}
}