Merge pull request #13660 from roosterfish/powerflex_sdc

Storage: Add Dell PowerFlex SDC operation mode

doc/.custom_wordlist.txt

@@ -126,6 +126,7 @@ macOS
 macvlan
 manpages
 Mbit
+MDM
 MiB
 Mibit
 MicroCeph
@@ -199,6 +200,7 @@ runtime
 SATA
 scalable
 scriptlet
+SDC
 SDN
 SDS
 SDT

doc/config_options.txt

@@ -5467,7 +5467,7 @@ This option is required only if {config:option}`storage-powerflex-pool-conf:powe
 :shortdesc: "How volumes are mapped to the local server"
 :type: "string"
 The mode gets discovered automatically if the system provides the necessary kernel modules.
-Currently, only `nvme` is supported.
+This can be either `nvme` or `sdc`.
 ```
 
 ```{config:option} powerflex.pool storage-powerflex-pool-conf

doc/reference/storage_powerflex.md

@@ -3,11 +3,15 @@
 
 [Dell PowerFlex](https://www.dell.com/en-us/shop/powerflex/sf/powerflex) is a software-defined storage solution from [Dell Technologies](https://www.dell.com/). Among other things it offers the consumption of redundant block storage across the network.
 
-LXD offers access to PowerFlex storage clusters by making use of the NVMe/TCP transport protocol.
+LXD offers access to PowerFlex storage clusters using either NVMe/TCP or Dell's Storage Data Client (SDC).
 In addition, PowerFlex offers copy-on-write snapshots, thin provisioning and other features.
 
-To use PowerFlex, make sure you have the required kernel modules installed on your host system.
+To use PowerFlex with NVMe/TCP, make sure you have the required kernel modules installed on your host system.
 On Ubuntu these are `nvme_fabrics` and `nvme_tcp`, which come bundled in the `linux-modules-extra-$(uname -r)` package.
+LXD takes care of connecting to the respective subsystem.
+
+When using the SDC, LXD requires it to already be connected to the Dell Metadata Manager (MDM) beforehand.
+As LXD doesn't set up the SDC, follow the official guides from Dell for configuration details.
 
 ## Terminology
 
@@ -17,9 +21,11 @@ A *protection domain* contains storage pools, which represent a set of physical
 LXD creates its volumes in those storage pools.
 
 You can take a snapshot of any volume in PowerFlex, which will create an independent copy of the parent volume.
-PowerFlex volumes get added as a NVMe drive to the respective LXD host the volume got mapped to.
-For this, the LXD host connects to one or multiple NVMe {abbr}`SDT (storage data targets)` provided by PowerFlex.
+PowerFlex volumes get added as a drive to the respective LXD host the volume got mapped to.
+In case of NVMe/TCP, the LXD host connects to one or multiple NVMe {abbr}`SDT (storage data targets)` provided by PowerFlex.
 Those SDT run as components on the PowerFlex storage layer.
+In case of SDC, the LXD hosts don't set up any connection by themselves.
+Instead they depend on the SDC to make the volumes available on the system for consumption.
 
 ## `powerflex` driver in LXD
 
@@ -35,14 +41,14 @@ This driver behaves differently than some of the other drivers in that it provid
 As a result and depending on the internal network, storage access might be a bit slower than for local storage.
 On the other hand, using remote storage has big advantages in a cluster setup, because all cluster members have access to the same storage pools with the exact same contents, without the need to synchronize storage pools.
 
-When creating a new storage pool using the `powerflex` driver, LXD tries to discover one of the SDT from the given storage pool.
+When creating a new storage pool using the `powerflex` driver in `nvme` mode, LXD tries to discover one of the SDT from the given storage pool.
 Alternatively, you can specify which SDT to use with {config:option}`storage-powerflex-pool-conf:powerflex.sdt`.
 LXD instructs the NVMe initiator to connect to all the other SDT when first connecting to the subsystem.
 
 Due to the way copy-on-write works in PowerFlex, snapshots of any volume don't rely on its parent.
 As a result, volume snapshots are fully functional volumes themselves, and it's possible to take additional snapshots from such volume snapshots.
 This tree of dependencies is called the *PowerFlex vTree*.
-Both volumes and their snapshots get added as standalone NVMe disks to the LXD host.
+Both volumes and their snapshots get added as standalone disks to the LXD host.
 
 (storage-powerflex-volume-names)=
 ### Volume names

go.mod

@@ -8,6 +8,7 @@ require (
 	github.com/canonical/go-dqlite v1.21.0
 	github.com/checkpoint-restore/go-criu/v6 v6.3.0
 	github.com/checkpoint-restore/go-criu/v7 v7.1.0
+	github.com/dell/goscaleio v1.14.1
 	github.com/digitalocean/go-qemu v0.0.0-20230711162256-2e3d0186973e
 	github.com/digitalocean/go-smbios v0.0.0-20180907143718-390a4f403a8e
 	github.com/dustinkirkland/golang-petname v0.0.0-20240428194347-eebcea082ee0

go.sum

@@ -105,6 +105,8 @@ github.com/davecgh/go-spew v1.1.0/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSs
 github.com/davecgh/go-spew v1.1.1/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
 github.com/davecgh/go-spew v1.1.2-0.20180830191138-d8f796af33cc h1:U9qPSI2PIWSS1VwoXQT9A3Wy9MM3WgvqSxFWenqJduM=
 github.com/davecgh/go-spew v1.1.2-0.20180830191138-d8f796af33cc/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
+github.com/dell/goscaleio v1.14.1 h1:SCHGLoOBKxQZ8EodChOLoIghcyhepbO5MLPRd4YQZ5c=
+github.com/dell/goscaleio v1.14.1/go.mod h1:h7SCmReARG/szFWBMQGETGkZObknhS45lQipQbtdmJ8=
 github.com/dgryski/go-farm v0.0.0-20200201041132-a6ae2369ad13 h1:fAjc9m62+UWV/WAFKLNi6ZS0675eEUC9y3AlwSbQu1Y=
 github.com/dgryski/go-farm v0.0.0-20200201041132-a6ae2369ad13/go.mod h1:SqUrOPUnsFjfmXRMNPybcSiG0BgUW2AuFH8PAnS2iTw=
 github.com/digitalocean/go-libvirt v0.0.0-20240610184155-f66fb3c0f6d7 h1:KMOLn19gbh7KbPEgu76ZIf/b2CnnYhC2GFLgLiN/YkA=

lxd/metadata/configuration.json

@@ -6154,7 +6154,7 @@
 					{
 						"powerflex.mode": {
 							"defaultdesc": "the discovered mode",
-							"longdesc": "The mode gets discovered automatically if the system provides the necessary kernel modules.\nCurrently, only `nvme` is supported.",
+							"longdesc": "The mode gets discovered automatically if the system provides the necessary kernel modules.\nThis can be either `nvme` or `sdc`.",
 							"shortdesc": "How volumes are mapped to the local server",
 							"type": "string"
 						}

lxd/storage/drivers/driver_powerflex.go

@@ -4,6 +4,8 @@ import (
 	"fmt"
 	"strings"
 
+	"github.com/dell/goscaleio"
+
 	deviceConfig "github.com/canonical/lxd/lxd/device/config"
 	"github.com/canonical/lxd/lxd/migration"
 	"github.com/canonical/lxd/lxd/operations"
@@ -18,6 +20,11 @@ const powerFlexDefaultUser = "admin"
 // powerFlexDefaultSize represents the default PowerFlex volume size.
 const powerFlexDefaultSize = "8GiB"
 
+const (
+	powerFlexModeNVMe = "nvme"
+	powerFlexModeSDC  = "sdc"
+)
+
 var powerFlexLoaded bool
 var powerFlexVersion string
 
@@ -27,6 +34,10 @@ type powerflex struct {
 	// Holds the low level HTTP client for the PowerFlex API.
 	// Use powerflex.client() to retrieve the client struct.
 	httpClient *powerFlexClient
+
+	// Holds the SDC GUID of this specific host.
+	// Use powerflex.getHostGUID() to retrieve the actual value.
+	sdcGUID string
 }
 
 // load is used to run one-time action per-driver rather than per-pool.
@@ -87,9 +98,14 @@ func (d *powerflex) FillConfig() error {
 		d.config["powerflex.user.name"] = powerFlexDefaultUser
 	}
 
+	// Try to discover the PowerFlex operation mode.
+	// First try if the NVMe/TCP kernel modules can be loaed.
+	// Second try if the SDC kernel module is setup.
 	if d.config["powerflex.mode"] == "" {
 		if d.loadNVMeModules() {
-			d.config["powerflex.mode"] = "nvme"
+			d.config["powerflex.mode"] = powerFlexModeNVMe
+		} else if goscaleio.DrvCfgIsSDCInstalled() {
+			d.config["powerflex.mode"] = powerFlexModeSDC
 		}
 	}
 
@@ -120,15 +136,11 @@ func (d *powerflex) Create() error {
 		return fmt.Errorf("The powerflex.gateway cannot be empty")
 	}
 
-	// Fail if no PowerFlex mode can be discovered.
-	if d.config["powerflex.mode"] == "" {
-		return fmt.Errorf("Failed to discover PowerFlex mode")
-	}
-
 	client := d.client()
 
-	// Discover one of the storage pools SDS services.
-	if d.config["powerflex.mode"] == "nvme" {
+	switch d.config["powerflex.mode"] {
+	case powerFlexModeNVMe:
+		// Discover one of the storage pools SDT services.
 		if d.config["powerflex.sdt"] == "" {
 			pool, err := d.resolvePool()
 			if err != nil {
@@ -150,6 +162,19 @@ func (d *powerflex) Create() error {
 
 			d.config["powerflex.sdt"] = relations[0].IPList[0].IP
 		}
+
+	case powerFlexModeSDC:
+		if d.config["powerflex.sdt"] != "" {
+			return fmt.Errorf("The powerflex.sdt config key is specific to the NVMe/TCP mode")
+		}
+
+		if !goscaleio.DrvCfgIsSDCInstalled() {
+			return fmt.Errorf("PowerFlex SDC is not available on the host")
+		}
+
+	default:
+		// Fail if no PowerFlex mode can be discovered.
+		return fmt.Errorf("Failed to discover PowerFlex mode")
 	}
 
 	return nil
@@ -209,12 +234,12 @@ func (d *powerflex) Validate(config map[string]string) error {
 		"powerflex.domain": validate.Optional(validate.IsAny),
 		// lxdmeta:generate(entities=storage-powerflex; group=pool-conf; key=powerflex.mode)
 		// The mode gets discovered automatically if the system provides the necessary kernel modules.
-		// Currently, only `nvme` is supported.
+		// This can be either `nvme` or `sdc`.
 		// ---
 		//  type: string
 		//  defaultdesc: the discovered mode
 		//  shortdesc: How volumes are mapped to the local server
-		"powerflex.mode": validate.Optional(validate.IsOneOf("nvme")),
+		"powerflex.mode": validate.Optional(validate.IsOneOf("nvme", "sdc")),
 		// lxdmeta:generate(entities=storage-powerflex; group=pool-conf; key=powerflex.sdt)
 		//
 		// ---
@@ -251,7 +276,7 @@ func (d *powerflex) Validate(config map[string]string) error {
 	// on the other cluster members too. This can be done here since Validate
 	// gets executed on every cluster member when receiving the cluster
 	// notification to finally create the pool.
-	if d.config["powerflex.mode"] == "nvme" && !d.loadNVMeModules() {
+	if d.config["powerflex.mode"] == powerFlexModeNVMe && !d.loadNVMeModules() {
 		return fmt.Errorf("NVMe/TCP is not supported")
 	}