Server Hardware Installation and Management — CompTIA Server+ (SK0-005) GED Study Notes

Chapter 1

Server Hardware Installation and Management

Course: Comptia Server Sk0 005
Subject: Comptiaserver Sk0 005
Chapter: Server Hardware Installation And Management

1. RAID Levels and Types

RAID (Redundant Array of Independent Disks) combines multiple physical drives into a single logical volume to improve performance, redundancy, or both. RAID can be implemented via hardware controllers or software (e.g., mdadm in Linux or Storage Spaces in Windows).

a. RAID 0 (Striping)

Data is striped across two or more disks.
No redundancy – if one disk fails, all data is lost.
Excellent performance for read/write operations.
100% usable capacity.

Minimum disks: 2 Fault tolerance: None Storage efficiency: 100% Performance: High (best for speed)

Use cases:

Temporary data like video rendering, scratch disks.
Gaming rigs or lab setups where speed is critical and data loss is acceptable.

Real-life example: A workstation used for 4K video editing uses two NVMe SSDs in RAID 0 for high throughput. All critical projects are backed up separately.

b. RAID 1 (Mirroring)

Data is mirrored (duplicated) on two disks.
Provides redundancy – system keeps running if one drive fails.
Read performance is good; write speed slightly slower due to mirroring.
50% usable capacity.

Minimum disks: 2 Fault tolerance: 1 disk can fail Storage efficiency: 50% Performance: Good reads, slower writes

Use cases:

Boot drives, critical application volumes.
Systems where data availability is essential.

Real-life example: Enterprise servers use RAID 1 for the OS volume to maintain uptime even if a boot drive fails.

c. RAID 5 (Striping with Distributed Parity)

Data and parity information are striped across all drives.
Can tolerate a single drive failure.
Balanced performance, storage efficiency, and redundancy.
Write performance is slower due to parity calculations.

Minimum disks: 3 Fault tolerance: 1 disk Storage efficiency: (N-1)/N Performance: Good reads, slower writes

Use cases:

General-purpose file servers.
Environments with read-heavy workloads and moderate fault tolerance needs.

Drawbacks:

Rebuild time is slow with large drives.
Risk of total data loss if a second disk fails during rebuild.

Real-life example: A small office uses RAID 5 with 4 drives in a Synology NAS to balance capacity, performance, and fault tolerance for shared files.

d. RAID 6 (Striping with Dual Parity)

Similar to RAID 5, but uses two parity blocks per stripe.
Can survive up to two simultaneous drive failures.
Write performance is slower than RAID 5 due to the second parity calculation.
Excellent for large arrays with big drives where rebuild time is long and risks are higher.

Minimum disks: 4 Fault tolerance: 2 disk failures Storage efficiency: (N-2)/N Performance: Good reads, slower writes than RAID 5

Use cases:

Archival systems or backup servers with large SATA disks.
Data centers with petabyte-scale storage where dual-drive failure is statistically likely during rebuild.

Real-life example: A media company uses a RAID 6 array with 12 x 10TB drives for video archives. Even with 2 failed disks, data remains accessible.

e. RAID 10 (RAID 1 + RAID 0)

Combines mirroring (RAID 1) and striping (RAID 0).
Provides high performance and high redundancy.
Can survive multiple disk failures, but only if failures are in different mirrored pairs.
No parity overhead and fast rebuilds.

Minimum disks: 4 Fault tolerance: 1 disk per mirror pair Storage efficiency: 50% Performance: Excellent read/write speed

Use cases:

Databases, OLTP systems, VM storage.
Environments demanding low latency and fast recovery.

Real-life example: A financial services company runs PostgreSQL on RAID 10 SSDs for fast IOPS and high uptime, critical for transaction processing.

f. JBOD (Just a Bunch of Disks)

Not a RAID level – drives are independent or spanned into a single logical volume.
No redundancy and no performance improvement.
Simple and flexible, often used with software-defined storage or LVM/ZFS.

Minimum disks: 1 Fault tolerance: None Storage efficiency: 100% Performance: Depends on configuration (usually basic)

Use cases:

Media storage, temporary volumes, lab environments.
Custom ZFS deployments that handle redundancy at the filesystem level.

Real-life example: A developer lab uses JBOD to test ZFS pools across multiple low-cost disks, letting ZFS manage parity and checksumming.

g. Hardware vs. Software RAID

Hardware RAID

Uses a dedicated RAID controller with onboard CPU, RAM cache, and often battery backup unit (BBU).
Minimal load on the host system CPU.
Supports advanced features like RAID-level migration, caching, and monitoring.

Pros:

Faster performance, better fault handling, independent of OS.
Preferred in enterprise-grade servers (e.g., Dell PERC, HPE Smart Array).

Cons:

Higher cost, proprietary firmware, recovery complexity if the controller fails.

Software RAID

Managed at the OS level (e.g., mdadm in Linux, Storage Spaces in Windows, ZFS pools).
Uses host CPU and memory.
Easier to migrate across systems and more flexible.

Pros:

Cost-effective, portable, great for non-critical or homelab setups.
Some advanced file systems (like ZFS) provide built-in software RAID with data integrity features.

Cons:

Slight performance overhead, dependent on OS stability.

Real-life example: A VMware ESXi host uses RAID 10 with a hardware RAID controller that has battery-backed write cache to store VMs. This ensures high performance and protection from power loss.

2. Capacity Planning

Purpose: Ensure storage infrastructure can handle current workloads and future growth without performance degradation or data loss.

Key Considerations:

Growth Estimation:

Base estimates on historical data growth trends (e.g., 20–30% annually).

Include business expansion, new applications, and retention policies.

Example: A legal firm with increasing document digitization may see 40% yearly growth due to scanned files and e-discovery requirements.

Base estimates on historical data growth trends (e.g., 20–30% annually).
Include business expansion, new applications, and retention policies.
Example: A legal firm with increasing document digitization may see 40% yearly growth due to scanned files and e-discovery requirements.
RAID Overhead:

RAID levels like 5, 6, and 10 reduce usable storage:

RAID 5: N-1 drives usable

RAID 6: N-2 drives usable

RAID 10: 50% usable space

Always calculate effective usable space after RAID and filesystem overhead.

Example: A 10-drive RAID 6 array using 8TB drives gives ~64TB raw but only ~48TB usable after RAID and formatting overhead.

RAID levels like 5, 6, and 10 reduce usable storage:

RAID 5: N-1 drives usable

RAID 6: N-2 drives usable

RAID 10: 50% usable space

RAID 5: N-1 drives usable
RAID 6: N-2 drives usable
RAID 10: 50% usable space
Always calculate effective usable space after RAID and filesystem overhead.
Example: A 10-drive RAID 6 array using 8TB drives gives ~64TB raw but only ~48TB usable after RAID and formatting overhead.
Snapshots and Backups:

Dedicate capacity for versioning, snapshots, and backup retention.

Estimate based on frequency and retention policy (e.g., daily snapshots for 7 days, weekly for 4 weeks).

Cloud environments (e.g., AWS EBS snapshots) also incur storage costs.

Dedicate capacity for versioning, snapshots, and backup retention.
Estimate based on frequency and retention policy (e.g., daily snapshots for 7 days, weekly for 4 weeks).
Cloud environments (e.g., AWS EBS snapshots) also incur storage costs.
Performance Needs:

IOPS (Input/Output Operations Per Second): Crucial for databases, VMs, and OLTP systems.

Throughput (MBps/GBps): Important for large file transfers, video editing, or backup operations.

Match storage performance to application demand using baseline performance tools.

IOPS (Input/Output Operations Per Second): Crucial for databases, VMs, and OLTP systems.
Throughput (MBps/GBps): Important for large file transfers, video editing, or backup operations.
Match storage performance to application demand using baseline performance tools.
Application Requirements:

Temporary file growth (e.g., /tmp, SQL tempdb).

Log retention (especially verbose logs like web/app access logs).

Example: A MySQL server with slow-query logging and binary logs can quickly consume additional space.

Temporary file growth (e.g., /tmp, SQL tempdb).
Log retention (especially verbose logs like web/app access logs).
Example: A MySQL server with slow-query logging and binary logs can quickly consume additional space.

Monitoring & Tools:

Linux:

iostat: Disk I/O performance stats

iotop: Live view of process disk usage

df, du: Storage usage

iostat: Disk I/O performance stats
iotop: Live view of process disk usage
df, du: Storage usage
Windows:

Performance Monitor (PerfMon): Tracks disk latency, IOPS, queue depth

Performance Monitor (PerfMon): Tracks disk latency, IOPS, queue depth
Vendor-Specific Dashboards:

NetApp Active IQ, HPE InfoSight, Dell OpenManage: Predictive analytics, health, and growth projections

NetApp Active IQ, HPE InfoSight, Dell OpenManage: Predictive analytics, health, and growth projections

Best Practices:

Add 20–30% overhead to capacity forecasts to account for:

Unexpected growth

Snapshots or failed backups

Migration staging or temp working space

Unexpected growth
Snapshots or failed backups
Migration staging or temp working space
Tier storage: Use fast disks (SSD/NVMe) for high-demand apps and bulk storage (HDD) for archival data.

3. Hard Drive Media Types

Understanding media types is crucial for designing a storage solution based on performance, cost, and workload.

a. Solid State Drive (SSD)

No moving parts, resulting in:

Faster access (low latency)

Higher IOPS

Better shock resistance

Faster access (low latency)
Higher IOPS
Better shock resistance
Ideal for high-transaction environments like VMs, databases, and search indexes.

Types:

SATA SSD:

6Gbps speed limit (approx. 550MB/s max throughput)

Cost-effective

Great for boot, general-purpose apps

Example: Crucial MX500, Samsung 860 series

6Gbps speed limit (approx. 550MB/s max throughput)
Cost-effective
Great for boot, general-purpose apps
Example: Crucial MX500, Samsung 860 series
NVMe SSD:

Connects via PCIe lanes (x4 or x8)

Extremely high performance: up to 7GB/s+ read speeds

Requires motherboard/controller support

Real-World Example: Samsung PM9A3 NVMe with 3 DWPD used in OLTP database clusters for rapid transactions.

Connects via PCIe lanes (x4 or x8)
Extremely high performance: up to 7GB/s+ read speeds
Requires motherboard/controller support
Real-World Example: Samsung PM9A3 NVMe with 3 DWPD used in OLTP database clusters for rapid transactions.

i. SSD Wear Factors:

Write Endurance:

Measured in DWPD (Drive Writes Per Day) or TBW (Total Bytes Written)

SSDs degrade over time due to flash memory cell limitations

Read-Intensive (RI):

Lower endurance (e.g., boot drives)

Example: Intel D3-S4610 (for read-heavy environments)

Write-Intensive (WI):

Higher endurance SSDs

Used in log servers, write-heavy databases

Example: Intel P4610, Micron 7400 PRO

Measured in DWPD (Drive Writes Per Day) or TBW (Total Bytes Written)
SSDs degrade over time due to flash memory cell limitations
Read-Intensive (RI):

Lower endurance (e.g., boot drives)

Example: Intel D3-S4610 (for read-heavy environments)

Lower endurance (e.g., boot drives)
Example: Intel D3-S4610 (for read-heavy environments)
Write-Intensive (WI):

Higher endurance SSDs

Used in log servers, write-heavy databases

Example: Intel P4610, Micron 7400 PRO

Higher endurance SSDs
Used in log servers, write-heavy databases
Example: Intel P4610, Micron 7400 PRO
Wear Leveling and Over-Provisioning:

Techniques used by SSD firmware to distribute writes evenly and reserve spare space for degraded blocks.

Techniques used by SSD firmware to distribute writes evenly and reserve spare space for degraded blocks.

b. Hard Disk Drive (HDD)

Mechanical storage with spinning platters
Slower but more cost-efficient for storing large volumes of data

i. Rotational Speed (RPM):

15,000 RPM:

High-performance SAS drives (legacy, being replaced by SSDs)

Used in transactional databases, now mostly deprecated

High-performance SAS drives (legacy, being replaced by SSDs)
Used in transactional databases, now mostly deprecated
10,000 RPM:

Moderate performance SAS disks

Still viable in tiered hybrid environments

Moderate performance SAS disks
Still viable in tiered hybrid environments
7,200 RPM:

Common SATA drives for bulk data

Use in backup systems, archival, surveillance storage

Common SATA drives for bulk data
Use in backup systems, archival, surveillance storage

Use Case:

7.2K RPM HDDs deployed in a Veeam backup repository for storing multi-terabyte daily backups.

c. Hybrid Drives (SSHD)

Combines a small amount of SSD cache with traditional HDD storage
Faster than HDDs for repeated reads (e.g., OS boot), but still slower than full SSD
Not recommended for modern enterprise deployments
Mostly used in consumer desktops or branch office file servers

Summary Comparison:

4. Interface Types

Understanding interface types is critical for selecting the right storage or expansion technology based on performance, reliability, and scalability needs.

a. Serial Attached SCSI (SAS)

Enterprise-grade storage interface based on the SCSI protocol.
Point-to-point and full-duplex (simultaneous read/write).
Common speeds:

SAS-2: 6Gbps

SAS-3: 12Gbps

SAS-4: 24Gbps (emerging)

SAS-2: 6Gbps
SAS-3: 12Gbps
SAS-4: 24Gbps (emerging)

Key Features:

Hot-swap support – Drives can be replaced without system shutdown.
Advanced error detection and correction, command queuing.
Dual-port capability allows high availability paths in redundant setups.
Backward-compatible with SATA drives (but limited to SATA performance when mixed).

Use Cases:

High-IOPS environments such as:

Virtualization hosts

Database clusters (e.g., Oracle RAC, SQL Server)

SAN backends with high availability (e.g., Dell EMC PowerVault, HPE MSA)

Virtualization hosts
Database clusters (e.g., Oracle RAC, SQL Server)
SAN backends with high availability (e.g., Dell EMC PowerVault, HPE MSA)

Real-World Example:

A Dell PowerEdge R740xd server using 12Gbps SAS backplane to host a RAID 10 array of 10K RPM SAS drives for transactional workloads.

b. Serial ATA (SATA)

Consumer-level and lower-tier enterprise interface.
Designed for cost efficiency rather than performance.

Characteristics:

Maximum speed: 6Gbps (SATA III)
Half-duplex: Can either read or write, not both simultaneously.
Longer command queuing and higher latency than SAS.
Lower MTBF (Mean Time Between Failures) and endurance in heavy-write environments.

Benefits:

Cheaper per GB.
Wide availability of drives in large capacities (1TB–18TB+).

Use Cases:

Cold data, backups, archives
Entry-level NAS or file servers
Budget-conscious VMs or home labs

Real-World Example:

Synology NAS with 4x 8TB SATA drives configured in RAID 5 for archival photo storage in a media company.

c. Peripheral Component Interconnect Express (PCIe)

High-speed interface used primarily for:

NVMe SSDs

RAID cards

Host Bus Adapters (HBAs)

GPUs and NICs in servers

NVMe SSDs
RAID cards
Host Bus Adapters (HBAs)
GPUs and NICs in servers

PCIe Generations:

| Gen | Bandwidth (x1) | Bandwidth (x4) | Bandwidth (x16) | | 3.0 | 1 GB/s | 4 GB/s | 16 GB/s | | 4.0 | 2 GB/s | 8 GB/s | 32 GB/s | | 5.0 | 4 GB/s | 16 GB/s | 64 GB/s |

Each generation doubles bandwidth over the previous.
PCIe is full-duplex, point-to-point, and hot-plug capable (in some systems).
Lane configuration matters (x1, x4, x8, x16) – Determines max bandwidth for devices.

Use Cases:

NVMe drives for ultra-fast read/write (e.g., VMware datastores, OLTP databases)
HBAs for SAN connectivity (Fibre Channel, SAS expanders)
High-performance NICs (25GbE/100GbE) and GPUs (AI/ML acceleration)

Real-World Example:

A VMware ESXi host using PCIe 4.0 x8 NVMe SSDs (Samsung PM1735) and a 100GbE NIC on separate PCIe lanes for simultaneous fast I/O and network throughput in a hyper-converged infrastructure.

d. External SATA (eSATA)

External interface version of the SATA standard.
Provides SATA-level performance (up to 6Gbps) to external drives.
Supports hot-swapping, depending on host/controller.
Requires dedicated eSATA ports (not interchangeable with standard SATA ports).

Limitations:

Lacks bus power (unlike USB), so external drives need a separate power supply.
Limited device ecosystem and less flexible than USB.
No plug-and-play support like USB; not ideal for casual users.

Modern Relevance:

Largely deprecated and replaced by USB 3.x and Thunderbolt due to better speed, power delivery, and device support.

Use Cases (historical):

External disk enclosures for direct-attached storage (DAS).
Used in small office/home NAS or backup setups.

Real-World Example:

An older backup workstation with an eSATA port used to connect a 2TB external drive for weekly imaging using Acronis or Clonezilla.

e. Universal Serial Bus (USB)

Ubiquitous standard for peripheral connectivity and storage devices.
Supports plug-and-play, hot-swapping, and power delivery.

USB Versions and Speeds:

Use Cases:

External backup drives, flash drives, tape drives (LTO over USB).
Offline data transfer and recovery (e.g., ISO loading, recovery media).
Boot media creation for OS installs (e.g., Windows PE, Linux live distros).

Real-World Example:

A disaster recovery technician uses a USB 3.2 NVMe enclosure (e.g., Samsung T7) for bare-metal restores using Veeam or Clonezilla at client sites.

f. Secure Digital (SD)

Flash memory card standard, primarily used in portable and embedded systems.

Variants:

Comes in standard, mini, and micro form factors.

Use Cases:

Raspberry Pi boot devices, logging storage in IoT/edge systems, and digital signage.
Occasional use in air-gapped or forensic systems for quick software deployment.

Real-World Example:

A Raspberry Pi firewall appliance boots from a high-end SDXC card running a hardened Linux distro (e.g., pfSense, Raspbian with iptables).

5. Shared Storage

Shared storage is a centralized data repository accessible by multiple hosts over a network or fabric. It plays a critical role in:

High availability (HA) clusters
Live VM migration (e.g., VMware vMotion, Hyper-V Live Migration)
Centralized backups and archiving
Data consistency and collaboration across distributed systems

Types of shared storage include NAS, SAN, object storage, and software-defined storage (SDS).

a. Network Attached Storage (NAS)

NAS is a file-level storage solution connected to a network (usually via Ethernet).
Serves files over IP using protocols like NFS, SMB/CIFS, or FTP.
Appears to clients as shared folders or mapped drives.
Managed via dedicated NAS OS or software layer (e.g., Synology DSM, TrueNAS, QNAP QTS).

Key Characteristics:

Easier to deploy than SAN.
Optimized for file sharing and unstructured data.
Integrates with LDAP/Active Directory for access control.
Supports snapshots, replication, deduplication, and cloud sync.

i. NFS (Network File System)

Native to Unix/Linux systems.
Allows remote directory mounting over IP.
Supports user/group permissions, file locking, and caching.
Versions: NFSv3 (stateless), NFSv4 (stateful, supports ACLs and encryption).

Use Case:

Linux servers mount NFS exports for home directories, VM storage, or media streaming.
Backup targets in hybrid environments.

Real-World Example:

A VMware ESXi cluster mounts a NetApp FAS storage array via NFSv3 to host VM datastores, enabling vMotion and DRS features.

ii. CIFS/SMB (Common Internet File System / Server Message Block)

Windows-native protocol for file and printer sharing.
CIFS is the older SMB dialect (SMB1); modern versions are:

SMB2 – Reduced chatter, improved performance

SMB3 – Adds encryption, integrity signing, and multichannel support

SMB2 – Reduced chatter, improved performance
SMB3 – Adds encryption, integrity signing, and multichannel support
Fully integrated with Windows ACLs, AD authentication, and NTFS permissions.
Used in mixed OS environments with Linux via Samba.

Use Case:

Centralized document storage, group shares, and printer sharing in Windows domains.
Mapped drives for corporate users.

Real-World Example:

A Synology NAS exports a SMB3 share used by Windows 11 endpoints for storing HR files, while the same unit provides NFS shares for Linux-based backup appliances.

Advantages of NAS

Simplicity: Easy to deploy and manage with web-based interfaces.
Cost-effective: No need for FC switches or HBAs like in SANs.
Flexibility: Supports mixed OS environments and diverse workloads.

Limitations

File-level access may not suit block-oriented applications (e.g., databases, VM disks).
Network performance can be a bottleneck without 10GbE or LACP bonding.
Requires proper tuning for high concurrency (e.g., NFS thread counts, SMB caching).

b. Storage Area Network (SAN)

A SAN provides block-level access to shared storage devices across a dedicated network or fabric. Unlike NAS, which operates at the file level, SAN volumes appear to the operating system as local disks, enabling high-performance storage suitable for databases, virtualization, and transactional workloads.

Key Features:

Offers high throughput, low latency, and scalability.
Enables clustering, HA (high availability), and centralized storage management.
Often paired with multipathing software (e.g., DM-Multipath, VMware NMP) for redundancy and load balancing.

i. iSCSI (Internet Small Computer Systems Interface)

A TCP/IP-based SAN protocol that transmits SCSI commands over standard Ethernet.
Enables block-level storage access without requiring specialized Fibre Channel hardware.
Implemented in software (e.g., Linux open-iscsi) or hardware iSCSI HBAs.

Deployment Considerations:

Requires dedicated VLANs or network segmentation to isolate storage traffic.
Use QoS to prioritize iSCSI packets and avoid network congestion.
Often used in 1GbE, 10GbE, or faster Ethernet environments.

Pros:

Cost-effective and easier to deploy than Fibre Channel.
Ideal for small to medium enterprise environments.

Real-Life Example:

A VMware vSphere cluster uses Dell EqualLogic iSCSI SAN over a dedicated 10GbE VLAN to host shared VMFS datastores, allowing vMotion and HA functionality.

ii. Fibre Channel (FC)

A dedicated high-speed network technology used for storage traffic.
Uses Fibre Channel switches (e.g., Brocade, Cisco MDS) and HBAs (Host Bus Adapters).
Designed for low-latency, high-reliability, and deterministic performance.

Speed Tiers:

8 Gbps (legacy)
16 Gbps (commonly deployed)
32 Gbps (modern data centers)
64 Gbps (emerging in high-end infrastructures)

Topologies:

Point-to-point
FC-AL (arbitrated loop, legacy)
Fabric (most common; via FC switches)

Pros:

Superior performance under heavy loads.
Extremely low error rates and jitter.

Real-Life Example:

An Oracle RAC cluster in a financial institution connects to an EMC VMAX SAN via dual-path 32Gbps FC. FC zoning ensures isolation between application workloads.

iii. Fibre Channel over Ethernet (FCoE)

FCoE encapsulates Fibre Channel frames inside Ethernet frames.
Runs over lossless Ethernet enabled by Data Center Bridging (DCB) features.
Allows consolidation of storage and data traffic on the same physical infrastructure.

Requirements:

DCB-capable switches (e.g., Cisco Nexus)
FCoE-capable NICs (Converged Network Adapters - CNAs)

Pros:

Reduces cabling and port requirements in top-of-rack switches.
Maintains FC features (e.g., zoning, WWNs) while leveraging Ethernet infrastructure.

Limitations:

Still requires strict QoS and flow control.
Not as widely adopted as traditional FC or iSCSI due to complexity.

Real-Life Example:

A Hyper-converged cluster in a government agency uses Cisco UCS servers with FCoE CNAs connecting to NetApp AFF arrays, reducing rack clutter while maintaining FC-class storage reliability.

SAN Best Practices

Use multipathing (MPIO) to avoid single points of failure.
Monitor storage latency and queue depths to prevent performance degradation.
For iSCSI, isolate traffic using dedicated NICs, jumbo frames, and VLANs.
Implement zoning (FC) or CHAP authentication (iSCSI) for security.

Comparison Summary: