Study guide
14.1 Disk Partitioning, Volumes & Enterprise File Systems (NTFS vs ReFS)
Administering Windows Server Hybrid Core Infrastructure AZ-800
Track your progress as you read
Finished sections get a check mark.
1 Introduction & Exam Overview
About Exam AZ-800: Format, Scoring & the September 2026 Retirement
2 Chapter 1: Active Directory Domain Services Architecture & Deployment
1.1 Deploying & Managing Domain Controllers On-Premises 1.2 Domain Controllers in Azure IaaS & Virtualized DC Cloning 1.3 Read-Only Domain Controllers (RODCs) & Password Replication Policies 1.4 Flexible Single Master Operation (FSMO) Roles & Troubleshooting 1.5 Active Directory Sites, Subnets & Site Link Topology 1.6 AD DS Replication: Topology Generation, Convergence & Diagnostics 1.7 Forest and Domain Trust Relationships
3 Chapter 2: AD DS Security Principals, Service Accounts & Domain Join
2.1 Managing Users, Groups & Objects in Multi-Domain and Multi-Forest Environments 2.2 Service Account Types & Group Managed Service Accounts (gMSA) 2.3 Joining Windows Server to AD DS, Microsoft Entra Domain Services & Microsoft Entra ID 2.4 Creating & Managing Users, Groups and Organizational Units at Scale
4 Chapter 3: Hybrid Identity & Microsoft Entra Integration
3.1 Microsoft Entra Connect Sync Architecture, Installation & Filtering 3.2 SourceAnchor, Synchronization Rules & Microsoft Entra Connect Health 3.3 Microsoft Entra Cloud Sync & Disconnected Multi-Forest Topologies 3.4 Hybrid Authentication Methods & Staged Rollout 3.5 Microsoft Entra Domain Services 3.6 Integrating Microsoft Entra ID, AD DS & Entra Domain Services: Choosing an Architecture
5 Chapter 4: Group Policy in AD DS and Hybrid Environments
4.1 Group Policy Architecture & Processing Order (LSDOU) 4.2 Group Policy Filtering, Loopback Processing & Client-Side Extensions 4.3 Group Policy Preferences, Item-Level Targeting & Central Store 4.4 Group Policy in Microsoft Entra Domain Services & Fine-Grained Password Policies
6 Chapter 5: Remote Server Management in Hybrid Environments
5.1 Windows Admin Center (WAC) Deployment On-Premises and in Azure 5.2 PowerShell Remoting, WinRM & the Kerberos Double-Hop Problem 5.3 Just Enough Administration (JEA) for Constrained PowerShell Endpoints 5.4 OpenSSH on Windows Server & PowerShell Remoting over SSH 5.5 Secure RDP Access, Remote Desktop Gateway & Azure Bastion
7 Chapter 6: Hybrid Management with Azure Arc and Azure Services
6.1 Azure Arc-Enabled Servers Architecture & Onboarding 6.2 Azure Machine Configuration & VM Extensions on Arc Servers 6.3 Azure Update Manager: Assessment, Maintenance Configurations & Dynamic Scoping 6.4 Azure Automation for Hybrid Workloads: Runbooks & Hybrid Runbook Workers
8 Chapter 7: Hyper-V Virtualization & Virtual Machine Configuration
7.1 Hyper-V Type 1 Architecture, Installation & Hardware Prerequisites 7.2 VM Generations, Secure Boot, vTPM, Enhanced Session Mode & Direct VM Access 7.3 Compute, Dynamic Memory & Hardware Acceleration 7.4 VM Resource Controls, CPU Groups & Host Compute Partitioning 7.5 Hyper-V Virtual Hard Disks & Checkpoint Management 7.6 Hyper-V Virtual Switches, NIC Teaming & Switch Embedded Teaming (SET) 7.7 Integration Services, Guest Components & Time Synchronization
9 Chapter 8: Hyper-V High Availability, Security & Containers
8.1 Hyper-V Failover Clustering & Cluster Shared Volumes (CSV) 8.2 Hyper-V Replica & Azure Site Recovery (ASR) 8.3 Guarded Fabric & Shielded Virtual Machines 8.4 Windows Server Container Hosts, Isolation Models & Images 8.5 Managing Container Instances & Linux Containers with WSL 8.6 Windows Container Networking & AKS Enabled by Azure Arc 8.7 Managing Container Images, Tags & Azure Container Registry
10 Chapter 9: Azure IaaS Virtual Machine Administration
9.1 Azure VM Sizing, Series Selection & Managed Disk Tiers 9.2 Azure VM Host Caching, Disk Bursting, Ephemeral OS Disks & Disk Operations 9.3 Azure VM High Availability: Availability Sets, Zones & Proximity Placement Groups 9.4 Virtual Machine Scale Sets, Autoscale & Azure Dedicated Hosts 9.5 Azure VM Networking, NSGs & Connectivity
11 Chapter 10: Hybrid DNS Infrastructure & Name Resolution
10.1 Active Directory-Integrated DNS Architecture, Dynamic Updates & Zone Types 10.2 DNS Records, Aging & Scavenging and the GlobalNames Zone 10.3 DNS Forwarding, Conditional Forwarding & Azure DNS Zone Integration 10.4 Hybrid Name Resolution with Azure DNS Private Resolver 10.5 Windows Server DNS Policies & DNSSEC
12 Chapter 11: IP Addressing, DHCP & IPAM Management
11.1 DHCP Server Role, the DORA Lifecycle & Active Directory Authorization 11.2 DHCP Scopes, Option Hierarchy, Reservations & Relay Agents 11.3 DHCP High Availability & Failover Configurations 11.4 IP Address Management (IPAM) Deployment & Auditing 11.5 Resolving IP Address Issues in Hybrid Environments
13 Chapter 12: Hybrid Network Connectivity, Remote Access & Private Access
12.1 Remote Access Role, VPN Protocols & Site-to-Site VPN to Azure 12.2 Network Policy and Access Services (NPAS), RADIUS & Policy Processing 12.3 Azure Network Adapter & Azure Extended Network 12.4 Web Application Proxy & Microsoft Entra Application Proxy 12.5 Microsoft Entra Private Access & Global Secure Access 12.6 Azure Relay & Hybrid Connections
14 Chapter 13: File Services, DFS, SMB & Azure File Sync
13.1 Windows Server File Share Access, NTFS Permissions & Access-Based Enumeration 13.2 File Server Resource Manager: Quotas, File Screens & Storage Reports 13.3 DFS Namespaces (DFS-N) & DFS Replication (DFS-R) 13.4 BranchCache: Distributed Cache and Hosted Cache Modes 13.5 Modern SMB Protocols, SMB Multichannel & SMB over QUIC 13.6 Azure File Shares, Tiers & Hybrid Permissions 13.7 Azure File Sync, Cloud Tiering & Monitoring 13.8 Migrating DFS Namespaces & DFS Replication to Azure File Sync
15 Chapter 14: Windows Server Storage Infrastructure & High Availability
14.1 Disk Partitioning, Volumes & Enterprise File Systems (NTFS vs ReFS) 14.2 Data Deduplication Architecture, Usage Types & Management 14.3 Storage Spaces & Storage Spaces Direct (S2D) 14.4 Storage Replica & Storage Quality of Service 14.5 iSCSI Target Server, IQNs, CHAP Authentication & Multipath I/O
14.1 Disk Partitioning, Volumes & Enterprise File Systems (NTFS vs ReFS)
Key Takeaways Explain This Topic AI
GPT supports volumes far beyond the 2 TB MBR ceiling and up to 128 partitions, with a backup partition table at the end of the disk; a 10 TB disk initialized as MBR is truncated to 2 TB.
Converting an initialized MBR disk to GPT with Clear-Disk and Initialize-Disk destroys existing data, so the choice matters at initialization time.
Dynamic Disks are deprecated on Windows Server and are unsupported with Failover Clustering and ReFS; Storage Spaces is the supported replacement for spanned, striped, and mirrored volumes.
ReFS adds integrity streams, proactive scrubbing, automatic repair from a mirrored copy, and block cloning, which makes checkpoint merges and fixed-VHDX creation nearly instantaneous.
NTFS remains required for the boot volume, for Data Deduplication on general file shares, and wherever EFS, hard quotas per user, or compression are needed, because ReFS does not implement them.
Last updated: August 2026
Disk Partitioning, Volumes & Enterprise File Systems (NTFS vs ReFS)
Enterprise storage infrastructure in Windows Server forms the bedrock for mission-critical workloads, virtualization fabrics, database engines, and disaster recovery architectures. Designing, configuring, and optimizing storage requires an in-depth understanding of low-level disk partitioning schemes, modern volume management paradigms, and resilient file systems.
This section covers disk initialization (MBR vs. GPT), basic versus dynamic disks, file system selection (NTFS vs. ReFS), and the ReFS virtualization acceleration mechanisms (Block Cloning, Sparse VDL). Data Deduplication — its architecture, usage types, job scheduling, and PowerShell management — is covered in section 14.2.
1. Disk Partitioning & Initialization: MBR vs. GPT
When a physical disk or virtual hard disk (VHD/VHDX) is attached to Windows Server, it must be initialized with a specific partition style before partitions and volumes can be created. Windows Server supports two partitioning architectures: Master Boot Record (MBR) and GUID Partition Table (GPT) . +-----------------------------------------------------------------------------------+
- | DISK PARTITIONING ARCHITECTURES |
- | |
- | MASTER BOOT RECORD (MBR) GUID PARTITION TABLE (GPT) |
- | +--------------------------+ +-------------------------------+ |
- | | LBA 0: Master Boot Record| | LBA 0: Protective MBR | |
- | | - Bootstrap Code | +-------------------------------+ |
- | | - 4 Partition Table (64B)| | LBA 1: Primary GPT Header | |
- | | - Boot Signature (0x55AA)| +-------------------------------+ |
- | +--------------------------+ | LBA 2-33: Partition Array | |
- | | Primary Partition 1 | | (128 Partition Entries) | |
- | +--------------------------+ +-------------------------------+ |
- | | Primary Partition 2 | | User Data Partitions | |
- | +--------------------------+ | (Up to 18 Exabytes) | |
- | | Primary Partition 3 | | | |
- | +--------------------------+ +-------------------------------+ |
- | | Extended Partition | | LBA -33 to -2: Backup Entries | |
- | | - Logical Drive 1 | +-------------------------------+ |
- | | - Logical Drive 2 | | LBA -1: Backup GPT Header | |
- | +--------------------------+ +-------------------------------+ |
- +-----------------------------------------------------------------------------------+
- Architectural Comparison: MBR vs. GPT
- Feature / Specification Master Boot Record (MBR) GUID Partition Table (GPT)
- Firmware Architecture Legacy BIOS / UEFI CSM Native Unified Extensible Firmware Interface (UEFI)
- Maximum Addressable Disk Size 2 Terabytes (2 TB) (with 512-byte sectors) 18 Exabytes (18 EB / 18,000,000 TB)
- Maximum Number of Partitions 4 Primary (or 3 Primary + 1 Extended) 128 Partitions (Windows Server default)
- Logical Block Addressing (LBA) 32-bit sector addressing 64-bit sector addressing
Redundancy & Structural Integrity None (single point of failure at Sector 0) Dual Headers (Primary LBA 1 + Backup at end of disk) + CRC32 Checksums
- Partition Identification 1-byte System ID (e.g., 0x07 for NTFS) 128-bit Globally Unique Identifier (GUID)
- Protective Mechanism None Protective MBR at LBA 0 (prevents legacy disk tools from overwriting GPT)
Modern Workload Support Deprecated for modern enterprise storage Mandatory for disks > 2 TB, Storage Spaces, and Generation 2 VMs
Deep-Dive Analysis
The 2 TB MBR Ceiling: MBR utilizes a 32-bit field to record sector addresses. Assuming standard 512-byte sectors, the maximum addressable space is 2 32 × 512 bytes = 2 , 199 , 023 , 255 , 552 bytes ≈ 2.19 TB 2^{32} \times 512 \text{ bytes} = 2,199,023,255,552 \text{ bytes} \approx 2.19 \text{ TB} 2 32 × 512 bytes = 2 , 199 , 023 , 255 , 552 bytes ≈ 2.19 TB . Any physical drive capacity beyond 2 TB initialized as MBR becomes unusable unallocated space.
GPT Redundancy and Self-Healing: GPT writes a primary partition table at the beginning of the disk (LBA 1 through 33) and a duplicate backup table at the very end of the disk. Every GPT header contains a Cyclic Redundancy Check (CRC32) checksum of itself and the partition array. If corruption occurs in the primary table (e.g., due to media degradation or an interrupted write), Windows Server validates the CRC32 checksum, detects the corruption, and automatically reconstructs the partition layout using the backup GPT structures.
2. Basic Disks vs. Dynamic Disks (Deprecation Notice)
Windows Server categorizes disk storage into two distinct management types: Basic Disks and Dynamic Disks . +-----------------------------------------------------------------------------------+
- | BASIC DISKS VS DYNAMIC DISKS |
- | |
- | BASIC DISKS (SUPPORTED & STANDARD) DYNAMIC DISKS (DEPRECATED) |
- | +------------------------------------+ +-------------------------------+ |
- | | - Uses MBR or GPT Partition Tables | | - Uses LDM (Logical Disk Mgr) | |
- | | - Simple & Primary Partitions | | - Spanned, Striped, Mirrored | |
- | | - Fully supported in Failover | | - Incompatible with S2D/CSV | |
- | | Clustering & Storage Spaces | | - Cannot host BitLocker OS | |
- | | - Native ReFS Support | | - Deprecated since WS 2012 R2 | |
- | +------------------------------------+ +-------------------------------+ |
- +-----------------------------------------------------------------------------------+
- Detailed Differences and Volume Types
Basic Disks:
The standard disk type used by Windows Server.
Relies strictly on standard MBR or GPT partition tables to define simple volumes.
Supports all Windows Server enterprise features: Storage Spaces, Storage Spaces Direct (S2D), Clustered Shared Volumes (CSV), BitLocker Drive Encryption, and ReFS.
Dynamic Disks:
Introduced in Windows 2000 using a proprietary Logical Disk Manager (LDM) database stored in the last 1 MB of the physical disk.
Permitted software-based multi-disk volumes without hardware RAID controllers:
Simple Volume: Portions of a single dynamic disk.
Spanned Volume: Combines unallocated space from 2 to 32 disks into a single non-fault-tolerant volume.
Striped Volume (RAID-0): Interleaves data across 2 to 32 disks for performance without fault tolerance.
Mirrored Volume (RAID-1): Tolerates 1 disk failure using two identical copies.
RAID-5 Volume: Stripes data and parity across 3 or more disks.
Warning
Dynamic Disks Deprecation & Exam Trap: Dynamic Disks are officially deprecated in Windows Server. Dynamic disks do not support Storage Spaces, Storage Spaces Direct (S2D), Clustered Shared Volumes (CSV), or ReFS advanced features. Microsoft recommends converting all legacy Dynamic Disks to Basic Disks and implementing software pooling via Storage Spaces or hardware RAID.
3. Enterprise File Systems: NTFS vs. ReFS
Windows Server provides two primary file systems for enterprise data storage: NTFS (New Technology File System) and ReFS (Resilient File System) . +-----------------------------------------------------------------------------------+
- | FILE SYSTEM FEATURE COMPARISON |
- | |
- | CAPABILITY / FEATURE NTFS ReFS (v3.x) |
- | -------------------------------+-----------------------+------------------------|
- | Maximum Volume Size | 256 TB (64 KB cluster)| 35 Petabytes (35 PB) |
- | Maximum File Size | 256 TB (64 KB cluster)| 35 Petabytes (35 PB) |
- | Boot / OS Volume Support | YES (Mandatory) | NO (Non-bootable only) |
- | File-Level Security (ACLs) | YES | YES |
- | Integrity Streams (Checksums) | NO | YES (Data & Metadata) |
- | Proactive Data Scrubbing | NO | YES (Self-Healing) |
- | Block Cloning (Fast Checkpoint)| NO | YES (Virtualization) |
- | Sparse VDL (Rapid VHDX Init) | NO | YES (Instant Zeroing) |
- | Native File Compression | YES (LZNT1) | NO |
- | Encrypting File System (EFS) | YES (File-level cert) | NO (Use BitLocker) |
- | Disk Quotas | YES (User-based) | NO (Use FSRM Quotas) |
- | Hard Links / Extended Attr. | YES | NO |
- +-----------------------------------------------------------------------------------+
- Resilient File System (ReFS) Core Innovations
ReFS was engineered from the ground up to address massive storage scalability, resilience against data corruption, and hyper-dense virtualization workloads.
Integrity Streams & Proactive Data Scrubbing:
ReFS utilizes 64-bit checksums for all file system metadata. Checksums for user file data can be enabled using Integrity Streams ( Set-FileIntegrity ).
When hosted on a mirrored or parity Storage Space, ReFS continuously monitors data integrity. When a read operation detects a corrupted checksum ("bit rot"), ReFS queries the alternate mirror copy, delivers uncorrupted data to the application, and transparently overwrites the damaged block on the failing drive.
A periodic background task ( Data Scrubber ) scans the entire volume proactively to repair latent corruption before applications request the affected blocks.
Block Cloning (Metadata Copy-on-Write):
In traditional NTFS, creating a Hyper-V VM checkpoint merge (merging a .avhdx differencing disk into a base .vhdx ) requires reading gigabytes of data from one file and physically writing them to another, causing substantial disk I/O latency.
On ReFS, Block Cloning performs an instant metadata operation: it re-maps the logical cluster pointers of the destination file to point directly to the existing physical disk clusters of the source file. The physical data is never duplicated on disk until a subsequent write modifies a shared block (Copy-on-Write). This reduces VM checkpoint merge times from tens of minutes to sub-second operations .
Sparse Valid Data Length (Sparse VDL):
On NTFS, creating a fixed-size 1 TB .vhdx file requires physically writing 1 TB of zeros across the drive, locking disk throughput for several minutes.
On ReFS, Sparse VDL allows the file system to mark the newly allocated space as zeroed in metadata without writing physical zeros to disk, enabling instant creation of fixed-size VHDX files . # Initialize Disk 1 with GPT and create an ReFS volume optimized for Hyper-V
- Initialize-Disk -Number 1 -PartitionStyle GPT
- $partition = New-Partition -DiskNumber 1 -UseMaximumSize -AssignDriveLetter
Format-Volume -Partition $partition -FileSystem ReFS -NewFileSystemLabel 'HyperV_Storage' -AllocationUnitSize 64KB
- # Enable Integrity Streams on a specific folder or virtual machine directory
- Set-FileIntegrity -FileName 'D:\Hyper-V\Virtual Hard Disks' -Enable $True