在 VMM 中管理儲存空間直接存取Manage Storage Spaces Direct in VMM

重要

已不再支援此版本的 Virtual Machine Manager (VMM),建議升級至 VMM 2019This version of Virtual Machine Manager (VMM) has reached the end of support, we recommend you to upgrade to VMM 2019.

本文提供儲存空間直接存取 (S2D) 的概觀,以及如何在 System Center - Virtual Machine Manager (VMM) 網狀架構中進行部署。This article provides an overview of Storage Spaces Direct (S2D), and how it's deployed in the System Center - Virtual Machine Manager (VMM) fabric.

Windows Server 2016 中引進了「儲存空間直接存取」(S2D)。Storage Spaces Direct (S2D) was introduced in Windows Server 2016. 它將實體存放磁碟機分組成虛擬存放集區,以提供虛擬化的存放裝置。It groups physical storage drives into virtual storage pools, to provide virtualized storage. 您可以使用虛擬化存放裝置:With virtualized storage you can:

  • 將多個實體儲存體來源作為單一虛擬實體管理。Manage multiple physical storage sources as a single virtual entity.
  • 取得低成本存放裝置,而不一定要有外部存放裝置。Get inexpensive storage, with and without external storage devices.
  • 將不同類型的存放裝置聚集至單一虛擬存放集區Gather different types of storage into a single virtual storage pool
  • 輕鬆地佈建存放裝置,並新增磁碟機以依需求擴充虛擬化存放裝置。Easily provision storage, and expand virtualized storage on demand by adding new drives.

它如何運作?How does it work?

S2D 會從連結至 Windows Server 叢集中特定節點的存放裝置中,建立存放集區。S2D creates pools of storage from storage that's attached to specific nodes in a Windows Server cluster. 存放裝置可以在節點內部,或是直接連結至單一節點的磁碟裝置。The storage can be internal on the node, or disk devices that are directly attached to a single node. 支援的存放裝置磁碟機包含 NVMe、透過 SATA 或 SAS 連接的 SSD,以及 HDD。Supported storage drives include NVMe, SSD connected via SATA or SAS, and HDD. 進一步瞭解Learn more.

  • 當您在 Windows Server 叢集上啟用 S2D 時,S2D 會自動探索合格的存放裝置,並將它新增至叢集的存放集區。When you enable S2D on a Windows Server cluster, S2D automatically discovers eligible storage, and adds it to a storage pool for the cluster.
  • S2D 也會建立內建伺服器端存放快取,將效能最大化。S2D also creates a built-in server-side storage cache to maximize performance. 最快的磁碟機會用於快取,其餘磁碟機則用於容量。The fastest drives are used for caching, and the remaining drives for capacity. 深入了解快取。Learn more about the cache.
  • 您將從存放集區建立磁碟區。You create volumes from a storage pool. 建立磁碟區會建立虛擬磁碟 (存放空間)、磁碟分割並將它格式化、新增到叢集,然後將它轉換為叢集共用磁碟區 (CSV)。Creating a volume creates the virtual disk (storage space), partitions and formats it, adds it to the cluster, and converts it to a clustered shared volumes (CSV).
  • 您將為磁碟區設定不同的容錯層級,以使用 SMB 3.0 來指定虛擬磁碟分散到集區中實體磁碟的方式。You configure different levels of fault tolerance for a volume, to specify how virtual disks are spread across physical disks in the pool, using SMB 3.0. 您可以設定無復原的磁碟區,或是具有鏡像或同位檢查彈性磁碟區。You can configure a volume with no resiliency, or with mirror or parity resilience. 進一步瞭解Learn more.

交集和非交集部署Converged and non-converged deployment

您可以使用數種方式部署執行 S2D 的叢集︰A cluster running S2D can be deployed in a couple of ways:

  • 超交集部署:Hyper-V 計算及 S2D 存放裝置會在相同的叢集內執行,而且它們之間沒有區隔。Hyper-converged deployment: Hyper-V compute and S2D storage run within the same cluster, with no separation between them. 這會同時調整計算和存放裝置資源。This provides simultaneous scaling of compute and storage resources.
  • Disaggregated deployment:計算資源會在一個 Hyper-V 叢集上執行。Disaggregated deployment: Compute resources run on one Hyper-V cluster. S2D 存放裝置會在不同的叢集上執行。S2D storage runs on a different cluster. 您分開調整叢集,以進行細微的微調管理。You scale the clusters separately for finely tuned management.

超交集部署Hyper-converged deployment

以下是超交集部署的圖例Here's an illustration for hyper-converged deployment

超交集

圖 1:超交集部署Figure 1: Hyper-converged deployment

  • VM 檔案會存放至本機 CSV。VM files are stored on local CSVs.
  • 不使用檔案共用和 SMB。File shares and SMB aren't used.
  • S2D CSV 磁碟區可用之後,以如同任何其他 Hyper-V 部署的佈建方式來佈建它們。After S2D CSV volumes are available, you provision them as you would any other Hyper-V deployment.
  • 您會同時調整 Hyper-V 計算叢集和其 S2D 存放裝置。You scale the Hyper-V compute cluster together with its S2D storage.

分類式部署Disaggregated deployment

以下是分類式部署的圖例Here's an illustration for disaggregated deployment

分類式

圖 2:分類式部署Figure 2: Disaggregated deployment

  • 檔案共用建立在 S2D CSV 上。File shares are created on the S2D CSVs.
  • Hyper-V VM 已設定為將其檔案儲存至向外延展檔案伺服器 (SOFS),並使用 SMB 3.0 存取。Hyper-V VMs are configured to store their files on the scaled-out file server (SOFS), and accessed using SMB 3.0.
  • 您分開調整 Hyper-V 和 SOFS 叢集,以進行細微的微調管理。You scale the Hyper-V and SOFS clusters separately for finely tuned management. 例如,計算節點可能會接近 VM 數目的完整容量,但存放裝置節點可能會有多餘的磁碟和 IOPS 容量,好讓您新增其他計算節點 (僅限計算節點)。For example, compute nodes might be near full capacity for the number of VMs, but storage nodes might have excess disk and IOPS capacity, so you add only additional compute nodes.

後續步驟Next steps

部署超交集 S2D 叢集Deploy a hyper-converged S2D cluster