open-nomad/nomad/structs/csi.go

635 lines
16 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 structs
import (
"fmt"
"strings"
"time"
)
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
// CSISocketName is the filename that Nomad expects plugins to create inside the
// PluginMountDir.
const CSISocketName = "csi.sock"
// CSIIntermediaryDirname is the name of the directory inside the PluginMountDir
// where Nomad will expect plugins to create intermediary mounts for volumes.
const CSIIntermediaryDirname = "volumes"
// VolumeTypeCSI is the type in the volume stanza of a TaskGroup
const VolumeTypeCSI = "csi"
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
// CSIPluginType is an enum string that encapsulates the valid options for a
// CSIPlugin stanza's Type. These modes will allow the plugin to be used in
// different ways by the client.
type CSIPluginType string
const (
// CSIPluginTypeNode indicates that Nomad should only use the plugin for
// performing Node RPCs against the provided plugin.
CSIPluginTypeNode CSIPluginType = "node"
// CSIPluginTypeController indicates that Nomad should only use the plugin for
// performing Controller RPCs against the provided plugin.
CSIPluginTypeController CSIPluginType = "controller"
// CSIPluginTypeMonolith indicates that Nomad can use the provided plugin for
// both controller and node rpcs.
CSIPluginTypeMonolith CSIPluginType = "monolith"
)
// CSIPluginTypeIsValid validates the given CSIPluginType string and returns
// true only when a correct plugin type is specified.
func CSIPluginTypeIsValid(pt CSIPluginType) bool {
switch pt {
case CSIPluginTypeNode, CSIPluginTypeController, CSIPluginTypeMonolith:
return true
default:
return false
}
}
// TaskCSIPluginConfig contains the data that is required to setup a task as a
// CSI plugin. This will be used by the csi_plugin_supervisor_hook to configure
// mounts for the plugin and initiate the connection to the plugin catalog.
type TaskCSIPluginConfig struct {
// ID is the identifier of the plugin.
// Ideally this should be the FQDN of the plugin.
ID string
// Type instructs Nomad on how to handle processing a plugin
Type CSIPluginType
// MountDir is the destination that nomad should mount in its CSI
// directory for the plugin. It will then expect a file called CSISocketName
// to be created by the plugin, and will provide references into
// "MountDir/CSIIntermediaryDirname/{VolumeName}/{AllocID} for mounts.
MountDir string
}
func (t *TaskCSIPluginConfig) Copy() *TaskCSIPluginConfig {
if t == nil {
return nil
}
nt := new(TaskCSIPluginConfig)
*nt = *t
return nt
}
// CSIVolumeAttachmentMode chooses the type of storage api that will be used to
// interact with the device.
type CSIVolumeAttachmentMode string
const (
CSIVolumeAttachmentModeUnknown CSIVolumeAttachmentMode = ""
CSIVolumeAttachmentModeBlockDevice CSIVolumeAttachmentMode = "block-device"
CSIVolumeAttachmentModeFilesystem CSIVolumeAttachmentMode = "file-system"
)
func ValidCSIVolumeAttachmentMode(attachmentMode CSIVolumeAttachmentMode) bool {
switch attachmentMode {
case CSIVolumeAttachmentModeBlockDevice, CSIVolumeAttachmentModeFilesystem:
return true
default:
return false
}
}
// CSIVolumeAccessMode indicates how a volume should be used in a storage topology
// e.g whether the provider should make the volume available concurrently.
type CSIVolumeAccessMode string
const (
CSIVolumeAccessModeUnknown CSIVolumeAccessMode = ""
CSIVolumeAccessModeSingleNodeReader CSIVolumeAccessMode = "single-node-reader-only"
CSIVolumeAccessModeSingleNodeWriter CSIVolumeAccessMode = "single-node-writer"
CSIVolumeAccessModeMultiNodeReader CSIVolumeAccessMode = "multi-node-reader-only"
CSIVolumeAccessModeMultiNodeSingleWriter CSIVolumeAccessMode = "multi-node-single-writer"
CSIVolumeAccessModeMultiNodeMultiWriter CSIVolumeAccessMode = "multi-node-multi-writer"
)
// ValidCSIVolumeAccessMode checks to see that the provided access mode is a valid,
// non-empty access mode.
func ValidCSIVolumeAccessMode(accessMode CSIVolumeAccessMode) bool {
switch accessMode {
case CSIVolumeAccessModeSingleNodeReader, CSIVolumeAccessModeSingleNodeWriter,
CSIVolumeAccessModeMultiNodeReader, CSIVolumeAccessModeMultiNodeSingleWriter,
CSIVolumeAccessModeMultiNodeMultiWriter:
return true
default:
return false
}
}
// ValidCSIVolumeAccessMode checks for a writable access mode
func ValidCSIVolumeWriteAccessMode(accessMode CSIVolumeAccessMode) bool {
switch accessMode {
case CSIVolumeAccessModeSingleNodeWriter,
CSIVolumeAccessModeMultiNodeSingleWriter,
CSIVolumeAccessModeMultiNodeMultiWriter:
return true
default:
return false
}
}
// CSIVolume is the full representation of a CSI Volume
type CSIVolume struct {
// ID is a namespace unique URL safe identifier for the volume
ID string
// Name is a display name for the volume, not required to be unique
Name string
// ExternalID identifies the volume for the CSI interface, may be URL unsafe
ExternalID string
Namespace string
Topologies []*CSITopology
AccessMode CSIVolumeAccessMode
AttachmentMode CSIVolumeAttachmentMode
// Allocations, tracking claim status
ReadAllocs map[string]*Allocation
WriteAllocs map[string]*Allocation
// Schedulable is true if all the denormalized plugin health fields are true, and the
// volume has not been marked for garbage collection
Schedulable bool
PluginID string
Provider string
ProviderVersion string
ControllerRequired bool
ControllersHealthy int
ControllersExpected int
NodesHealthy int
NodesExpected int
ResourceExhausted time.Time
CreateIndex uint64
ModifyIndex uint64
}
// CSIVolListStub is partial representation of a CSI Volume for inclusion in lists
type CSIVolListStub struct {
ID string
Namespace string
Name string
ExternalID string
Topologies []*CSITopology
AccessMode CSIVolumeAccessMode
AttachmentMode CSIVolumeAttachmentMode
CurrentReaders int
CurrentWriters int
Schedulable bool
PluginID string
Provider string
ControllersHealthy int
ControllersExpected int
NodesHealthy int
NodesExpected int
CreateIndex uint64
ModifyIndex uint64
}
// NewCSIVolume creates the volume struct. No side-effects
func NewCSIVolume(pluginID string, index uint64) *CSIVolume {
out := &CSIVolume{
ID: pluginID,
CreateIndex: index,
ModifyIndex: index,
}
out.newStructs()
return out
}
func (v *CSIVolume) newStructs() {
if v.Topologies == nil {
v.Topologies = []*CSITopology{}
}
v.ReadAllocs = map[string]*Allocation{}
v.WriteAllocs = map[string]*Allocation{}
}
func (v *CSIVolume) RemoteID() string {
if v.ExternalID != "" {
return v.ExternalID
}
return v.ID
}
func (v *CSIVolume) Stub() *CSIVolListStub {
stub := CSIVolListStub{
ID: v.ID,
Namespace: v.Namespace,
Name: v.Name,
ExternalID: v.ExternalID,
Topologies: v.Topologies,
AccessMode: v.AccessMode,
AttachmentMode: v.AttachmentMode,
CurrentReaders: len(v.ReadAllocs),
CurrentWriters: len(v.WriteAllocs),
Schedulable: v.Schedulable,
PluginID: v.PluginID,
Provider: v.Provider,
ControllersHealthy: v.ControllersHealthy,
NodesHealthy: v.NodesHealthy,
NodesExpected: v.NodesExpected,
CreateIndex: v.CreateIndex,
ModifyIndex: v.ModifyIndex,
}
return &stub
}
func (v *CSIVolume) CanReadOnly() bool {
if !v.Schedulable {
return false
}
return v.ResourceExhausted == time.Time{}
}
func (v *CSIVolume) CanWrite() bool {
if !v.Schedulable {
return false
}
switch v.AccessMode {
case CSIVolumeAccessModeSingleNodeWriter, CSIVolumeAccessModeMultiNodeSingleWriter:
return len(v.WriteAllocs) == 0
case CSIVolumeAccessModeMultiNodeMultiWriter:
return v.ResourceExhausted == time.Time{}
default:
return false
}
}
// Copy returns a copy of the volume, which shares only the Topologies slice
func (v *CSIVolume) Copy() *CSIVolume {
copy := *v
out := &copy
out.newStructs()
for k, v := range v.ReadAllocs {
out.ReadAllocs[k] = v
}
for k, v := range v.WriteAllocs {
out.WriteAllocs[k] = v
}
return out
}
// Claim updates the allocations and changes the volume state
func (v *CSIVolume) Claim(claim CSIVolumeClaimMode, alloc *Allocation) bool {
switch claim {
case CSIVolumeClaimRead:
return v.ClaimRead(alloc)
case CSIVolumeClaimWrite:
return v.ClaimWrite(alloc)
case CSIVolumeClaimRelease:
return v.ClaimRelease(alloc)
}
return false
}
// ClaimRead marks an allocation as using a volume read-only
func (v *CSIVolume) ClaimRead(alloc *Allocation) bool {
if _, ok := v.ReadAllocs[alloc.ID]; ok {
return true
}
if !v.CanReadOnly() {
return false
}
// Allocations are copy on write, so we want to keep the id but don't need the
// pointer. We'll get it from the db in denormalize.
v.ReadAllocs[alloc.ID] = nil
delete(v.WriteAllocs, alloc.ID)
return true
}
// ClaimWrite marks an allocation as using a volume as a writer
func (v *CSIVolume) ClaimWrite(alloc *Allocation) bool {
if _, ok := v.WriteAllocs[alloc.ID]; ok {
return true
}
if !v.CanWrite() {
return false
}
// Allocations are copy on write, so we want to keep the id but don't need the
// pointer. We'll get it from the db in denormalize.
v.WriteAllocs[alloc.ID] = nil
delete(v.ReadAllocs, alloc.ID)
return true
}
// ClaimRelease is called when the allocation has terminated and already stopped using the volume
func (v *CSIVolume) ClaimRelease(alloc *Allocation) bool {
delete(v.ReadAllocs, alloc.ID)
delete(v.WriteAllocs, alloc.ID)
return true
}
// Equality by value
func (v *CSIVolume) Equal(o *CSIVolume) bool {
if v == nil || o == nil {
return v == o
}
// Omit the plugin health fields, their values are controlled by plugin jobs
if v.ID == o.ID &&
v.Namespace == o.Namespace &&
v.AccessMode == o.AccessMode &&
v.AttachmentMode == o.AttachmentMode &&
v.PluginID == o.PluginID {
// Setwise equality of topologies
var ok bool
for _, t := range v.Topologies {
ok = false
for _, u := range o.Topologies {
if t.Equal(u) {
ok = true
break
}
}
if !ok {
return false
}
}
return true
}
return false
}
// Validate validates the volume struct, returning all validation errors at once
func (v *CSIVolume) Validate() error {
errs := []string{}
if v.ID == "" {
errs = append(errs, "missing volume id")
}
if v.PluginID == "" {
errs = append(errs, "missing plugin id")
}
if v.Namespace == "" {
errs = append(errs, "missing namespace")
}
if v.AccessMode == "" {
errs = append(errs, "missing access mode")
}
if v.AttachmentMode == "" {
errs = append(errs, "missing attachment mode")
}
// TODO: Volume Topologies are optional - We should check to see if the plugin
// the volume is being registered with requires them.
// var ok bool
// for _, t := range v.Topologies {
// if t != nil && len(t.Segments) > 0 {
// ok = true
// break
// }
// }
// if !ok {
// errs = append(errs, "missing topology")
// }
if len(errs) > 0 {
return fmt.Errorf("validation: %s", strings.Join(errs, ", "))
}
return nil
}
// Request and response wrappers
type CSIVolumeRegisterRequest struct {
Volumes []*CSIVolume
WriteRequest
}
type CSIVolumeRegisterResponse struct {
QueryMeta
}
type CSIVolumeDeregisterRequest struct {
VolumeIDs []string
WriteRequest
}
type CSIVolumeDeregisterResponse struct {
QueryMeta
}
type CSIVolumeClaimMode int
const (
CSIVolumeClaimRead CSIVolumeClaimMode = iota
CSIVolumeClaimWrite
CSIVolumeClaimRelease
)
type CSIVolumeClaimRequest struct {
VolumeID string
AllocationID string
Claim CSIVolumeClaimMode
WriteRequest
}
type CSIVolumeClaimResponse struct {
// Opaque static publish properties of the volume. SP MAY use this
// field to ensure subsequent `NodeStageVolume` or `NodePublishVolume`
// calls calls have contextual information.
// The contents of this field SHALL be opaque to nomad.
// The contents of this field SHALL NOT be mutable.
// The contents of this field SHALL be safe for the nomad to cache.
// The contents of this field SHOULD NOT contain sensitive
// information.
// The contents of this field SHOULD NOT be used for uniquely
// identifying a volume. The `volume_id` alone SHOULD be sufficient to
// identify the volume.
// This field is OPTIONAL and when present MUST be passed to
// `NodeStageVolume` or `NodePublishVolume` calls on the client
PublishContext map[string]string
// Volume contains the expanded CSIVolume for use on the client after a Claim
// has completed.
Volume *CSIVolume
QueryMeta
}
type CSIVolumeListRequest struct {
PluginID string
NodeID string
QueryOptions
}
type CSIVolumeListResponse struct {
Volumes []*CSIVolListStub
QueryMeta
}
type CSIVolumeGetRequest struct {
ID string
QueryOptions
}
type CSIVolumeGetResponse struct {
Volume *CSIVolume
QueryMeta
}
// CSIPlugin collects fingerprint info context for the plugin for clients
type CSIPlugin struct {
ID string
Provider string // the vendor name from CSI GetPluginInfoResponse
Version string // the vendor verson from CSI GetPluginInfoResponse
ControllerRequired bool
// Map Node.IDs to fingerprint results, split by type. Monolith type plugins have
// both sets of fingerprinting results.
Controllers map[string]*CSIInfo
Nodes map[string]*CSIInfo
// Allocations are populated by denormalize to show running allocations
Allocations []*AllocListStub
// Cache the count of healthy plugins
ControllersHealthy int
NodesHealthy int
CreateIndex uint64
ModifyIndex uint64
}
// NewCSIPlugin creates the plugin struct. No side-effects
func NewCSIPlugin(id string, index uint64) *CSIPlugin {
out := &CSIPlugin{
ID: id,
CreateIndex: index,
ModifyIndex: index,
}
out.newStructs()
return out
}
func (p *CSIPlugin) newStructs() {
p.Controllers = map[string]*CSIInfo{}
p.Nodes = map[string]*CSIInfo{}
}
func (p *CSIPlugin) Copy() *CSIPlugin {
copy := *p
out := &copy
out.newStructs()
for k, v := range p.Controllers {
out.Controllers[k] = v
}
for k, v := range p.Nodes {
out.Nodes[k] = v
}
return out
}
// AddPlugin adds a single plugin running on the node. Called from state.NodeUpdate in a
// transaction
func (p *CSIPlugin) AddPlugin(nodeID string, info *CSIInfo) {
if info.ControllerInfo != nil {
p.ControllerRequired = info.RequiresControllerPlugin &&
info.ControllerInfo.SupportsAttachDetach
prev, ok := p.Controllers[nodeID]
if ok && prev.Healthy {
p.ControllersHealthy -= 1
}
p.Controllers[nodeID] = info
if info.Healthy {
p.ControllersHealthy += 1
}
}
if info.NodeInfo != nil {
prev, ok := p.Nodes[nodeID]
if ok && prev.Healthy {
p.NodesHealthy -= 1
}
p.Nodes[nodeID] = info
if info.Healthy {
p.NodesHealthy += 1
}
}
}
// DeleteNode removes all plugins from the node. Called from state.DeleteNode in a
// transaction
func (p *CSIPlugin) DeleteNode(nodeID string) {
prev, ok := p.Controllers[nodeID]
if ok && prev.Healthy {
p.ControllersHealthy -= 1
}
delete(p.Controllers, nodeID)
prev, ok = p.Nodes[nodeID]
if ok && prev.Healthy {
p.NodesHealthy -= 1
}
delete(p.Nodes, nodeID)
}
type CSIPluginListStub struct {
ID string
Provider string
ControllerRequired bool
ControllersHealthy int
ControllersExpected int
NodesHealthy int
NodesExpected int
CreateIndex uint64
ModifyIndex uint64
}
func (p *CSIPlugin) Stub() *CSIPluginListStub {
return &CSIPluginListStub{
ID: p.ID,
Provider: p.Provider,
ControllerRequired: p.ControllerRequired,
ControllersHealthy: p.ControllersHealthy,
ControllersExpected: len(p.Controllers),
NodesHealthy: p.NodesHealthy,
NodesExpected: len(p.Nodes),
CreateIndex: p.CreateIndex,
ModifyIndex: p.ModifyIndex,
}
}
func (p *CSIPlugin) IsEmpty() bool {
return len(p.Controllers) == 0 && len(p.Nodes) == 0
}
type CSIPluginListRequest struct {
QueryOptions
}
type CSIPluginListResponse struct {
Plugins []*CSIPluginListStub
QueryMeta
}
type CSIPluginGetRequest struct {
ID string
QueryOptions
}
type CSIPluginGetResponse struct {
Plugin *CSIPlugin
QueryMeta
}