Seven years ago, I embarked on a journey by building my first home server. This endeavor significantly accelerated my software development workflow and brought a new level of enjoyment to my work. Since then, I’ve become increasingly involved in the home server community, expanding my setup with a custom storage server, another development server, and even a dedicated firewall.
However, as my homelab grew, a less desirable side effect emerged: my office became overrun with a chaotic tangle of wires. Initially, I was in denial, dismissing it as a “normal” amount of cabling. But upon closer inspection, it was undeniable – the wire situation had become quite extensive.
My office, upon closer inspection, kind of had a lot of wires.
Many home server enthusiasts opt for server racks to organize their equipment. Yet, I resisted the idea for a long time. I didn’t see myself as someone “serious” enough for a full rack. To me, racks were for data centers and hardcore IT professionals, not someone with a VM server here and a storage server there. Purchasing a rack felt like admitting I was more than just a casual user; it meant embracing the “homelab weirdo” label.
Eventually, the mounting wire clutter forced my hand. I conceded and invested in a server rack. Looking back, it was one of the best decisions I made for my homelab. It not only tamed the sprawling mess of wires but also made my servers significantly more accessible and pleasant to work with.
I. Direct to the Rack: No Time to Waste
If you’re eager to bypass the background and dive straight into the server rack details, jump directly to my final setup below.
II. Table of Contents
III. What Exactly is a Homelab?
The term “homelab” has gained considerable traction over the past decade within the tech community.
Simply put, a homelab is a dedicated space within your home designed for experimenting with IT hardware and software typically found in professional office environments or data centers. It serves as a sandbox for honing technical skills, exploring new technologies, and running personal projects that demand server infrastructure. Whether you’re a software developer, IT professional, or tech enthusiast, a homelab provides a safe and controlled environment to learn, test, and innovate without impacting your primary workstation or professional networks.
IV. Why Invest in a Server Rack for Your Home?
For those unfamiliar with server setups, the idea of housing multiple servers at home, let alone constructing a dedicated rack for them, might seem excessive. However, the benefits of a homelab server rack are numerous and cater to various needs.
Here are my primary motivations for building a server rack:
-
Streamlined Software Development: A dedicated server for virtual machines (VMs) isolates my development environment. Rebooting or upgrading my workstation no longer disrupts ongoing server processes. I can effortlessly create and discard experimental VMs without affecting other projects or risking system instability. This agility is crucial for rapid prototyping and testing different configurations.
-
Centralized and Scalable Storage: Instead of relying on fragmented storage across multiple devices, a network-attached storage (NAS) server offers a unified repository accessible to all devices on my network. This not only simplifies data management but also provides ample storage capacity for growing digital needs. My storage server utilizes ZFS, a robust file system that incorporates data integrity features, significantly minimizing the risk of data loss, even in the event of hard drive failures.
-
Enhanced Network Control and Security: Consumer-grade routers often come with limitations and buggy firmware. Building my own router using open-source software like OPNsense provides granular control over my network traffic, security settings, and advanced features. This DIY approach allows for customized firewall rules, VPN configurations, and intrusion detection systems, resulting in a more secure and reliable network compared to off-the-shelf solutions.
V. Why This DIY Server Rack Guide?
V.1. From a Beginner, For Beginners
Despite years of homelab experimentation, this server rack build was my first. This guide is intentionally crafted from a beginner’s perspective, documenting the process as I learned and navigated the challenges.
Many existing articles on server racks are written by seasoned experts, often glossing over crucial initial considerations. They frequently omit explanations of component selection rationale or the evaluation of alternatives, as their decision-making has become second nature after numerous rack builds.
Being a first-time server rack builder, I am free from the “curse of knowledge.” This guide meticulously details my thought process, from initial planning to final assembly, providing insights and considerations relevant to those embarking on their first DIY server rack project.
V.2. Unbiased Recommendations
This guide is entirely independent and free from commercial influences. I am not sponsored, nor have I received any complimentary products for this review. There are no affiliate links that could create a conflict of interest.
While affiliate marketing is a common practice in the homelab blogging sphere, it can inadvertently incentivize authors to promote specific products or brands, potentially compromising objectivity. My primary motivation for writing this guide is to share my experience, contribute to the homelab community, and receive feedback that helps refine my own approach.
It’s worth noting that my rack includes a TinyPilot, a hardware device that I created. However, this fact does not influence the recommendations for other components. I will always transparently disclose my connection to TinyPilot whenever it is mentioned.
VI. Selecting the Right Server Rack: A Step-by-Step Approach
Choosing a server rack might seem like the logical first step in building your homelab infrastructure. However, it’s more of an iterative process intertwined with component selection. You can’t definitively choose a rack without knowing the equipment it will house, yet understanding rack types influences the components you might consider.
Here’s the methodology I employed to select my server rack:
-
Initial Rack Exploration: Begin by browsing various server rack options to gain a general understanding of pricing, available features, and size dimensions. Websites of major rack manufacturers and online retailers are excellent starting points.
-
Preliminary Component List: Create a rough inventory of the hardware you intend to house within the rack. This list should include servers, network switches, UPS, and any other rack-mountable or shelf-mounted devices.
-
Rack Unit and Depth Calculation: Determine the total rack units (RU) required by summing the RU height of each component. Also, consider the depth requirements of your deepest components to ensure compatibility with the rack’s depth specifications. Account for future expansion by adding buffer space.
-
Rack Shortlisting: Narrow down your rack choices to models that meet your calculated RU height and depth requirements, while also considering features like post design (2-post or 4-post) and mobility (wheels).
-
Iterative Refinement: Revisit steps 2-4 as needed. As you refine your component list or discover new rack options, reassess your requirements and adjust your selection accordingly. This iterative process ensures that your final rack choice aligns perfectly with your homelab needs.
VI.1. Rack Unit (RU) Capacity: How Much Height Do You Need?
Server rack capacity is measured in rack units (RUs), with 1 RU equaling 1.75 inches in height. Racks are commonly designated by their RU height, such as an 18U rack, which provides 18 RUs or 31.5 inches (18 x 1.75″) of vertical mounting space.
Server racks are sized in rack units where each rack unit is 1.75″.
Rack-mountable equipment is designed in multiples of RUs. Network switches are often 1U, UPS units are typically 2U or 3U, and servers commonly range from 1U to 4U or more, depending on their configuration and features.
It’s crucial to accurately estimate your RU needs. Underestimating might leave you without enough space for future expansions, while overestimating could result in an unnecessarily large and space-consuming rack. As you select components, meticulously tally their RU heights and factor in anticipated future additions to your homelab.
VI.2. Rack Depth: Matching Rack to Component Depth
Server racks come in varying depths to accommodate different types of equipment. Enterprise-grade server racks are often designed to handle servers up to 50 inches long, built for maximum component density and cooling capacity in data centers.
However, for a home environment, space is often a premium. My own office is relatively compact, and I wanted to avoid a server rack dominating the entire room. Therefore, I opted to prioritize shallower components designed for front-mounting only, minimizing the overall depth footprint.
I focused on racks with a minimum depth of 19 inches. This depth provided sufficient space for rack shelves and front-mounted server chassis without protruding excessively into the room. Consider your available space and the depth of your planned equipment when determining your ideal rack depth.
VI.3. Two-Post vs. Four-Post Racks: Stability and Mounting Options
Server racks are available in two primary configurations: two-post and four-post. Four-post racks offer mounting points at both the front and rear, providing superior stability, especially for heavier and deeper equipment.
If you plan to utilize full-depth servers or heavy equipment, a four-post rack is essential to ensure secure mounting and prevent tipping. Two-post racks, while taking up less floor space, are generally more suitable for lighter, shallower equipment and may require careful weight distribution.
Although my components were primarily front-mounted and relatively lightweight, I opted for a four-post rack for the added stability. The slightly larger footprint was a worthwhile trade-off for the increased robustness and peace of mind.
VI.4. Mobility: The Convenience of Wheels
Mobility is a often overlooked yet highly practical feature for a home server rack. Racks equipped with wheels, or casters, allow for easy repositioning of the entire setup, greatly simplifying tasks like cleaning behind the rack or accessing rear connections.
For me, wheels were a critical requirement. The ability to effortlessly move the rack for cleaning and maintenance was a significant advantage. Consider whether mobility is important for your setup based on your space constraints and maintenance needs.
VI.5. Server Rack Candidates: Evaluating Options
StarTech is a well-regarded brand in the server rack market, known for its quality and comprehensive product line. Their website is user-friendly and provides detailed specifications, making them a natural choice for comparing different rack options. I focused my initial search on StarTech racks.
| Brand | Model | Min Depth | Posts | Wheels | Height | Price |
|---|---|---|---|---|---|---|
| StarTech | 4POSTRACK18U | 22″ | 4 | Yes | 18U | $316 |
| StarTech | 4POSTRACK15U | 22″ | 4 | Yes | 15U | $301 |
| StarTech | 2POSTRACK16 | 12″ | 2 | No | 16U | $165 |
While a 15U rack seemed initially sufficient for my current needs, the minimal price difference between the 15U and 18U models made the 18U a more future-proof choice. The extra 3U of space provided valuable headroom for future expansion without a significant cost increase.
VI.6. Server Rack Review: StarTech 4POSTRACK18U 18U Rack
- Grade: A
The StarTech 4POSTRACK18U rack has proven to be an excellent choice. Its construction is robust, providing a stable platform for all my equipment. The integrated wheels are smooth and make moving the fully loaded rack surprisingly easy.
Assembly was relatively straightforward, taking approximately 2.5 hours from start to finish. While the lack of labeled parts was a minor inconvenience, the clear instructions and part shapes made identification manageable.
The rack’s depth adjustability is a useful feature. However, the shallowest depth setting (22 inches) introduces a minor design flaw. Some screw holes become inaccessible at this depth. I addressed this by temporarily extending the depth, securing the inaccessible screws, and then readjusting to the desired shallow depth. This workaround, while slightly inconvenient, did not detract significantly from the overall quality and functionality of the rack.
VII. Network Switch Selection: Balancing Speed, Features, and Budget
Choosing a network switch proved to be the most challenging decision in the entire rack building process. Network switches can quickly escalate in price, especially as you move towards higher speeds and more advanced features. I wanted to strike a balance between performance, features, and budget, avoiding overspending on unnecessary capabilities while ensuring sufficient capacity for my homelab network.
VII.1. Network Speed Requirements: GbE, 2.5GbE, or 10GbE?
For rack-mounted switches, the common speed tiers are:
- 1 Gigabit Ethernet (GbE)
- 2.5 Gigabit Ethernet (2.5GbE)
- 10 Gigabit Ethernet (10GbE)
For over a decade, 1 Gbps Ethernet has been the standard in my home network, and it generally sufficed for internet browsing and general file sharing. The primary bottleneck was typically my internet service provider (ISP) speed, not my internal network.
However, as my data storage needs grew and I implemented a NAS server, the limitations of 1 Gbps became apparent. Transferring large files to and from my storage server became time-consuming, highlighting the need for a network upgrade. Performance benchmarks on my storage server clearly indicated that the 1 Gbps switch was now the bottleneck.
While 10 Gbps might seem like an excessive jump from 1 Gbps, research into 2.5 Gbps hardware revealed concerns about reliability and compatibility. Anecdotal evidence and online discussions suggested that 2.5 Gbps technology could be somewhat “flaky.” The consensus seemed to lean towards making the leap directly to 10 Gbps, as the complexities and potential headaches of 2.5 Gbps might outweigh the incremental speed gain over 1 Gbps.
Despite initial reservations about 10 Gbps being overkill, I ultimately decided to pursue a 10 Gbps capable switch, anticipating future bandwidth demands and the potential for smoother upgrades down the line.
Important Consideration: When evaluating 10 Gbps switches, carefully examine the port specifications. Some switches advertised as “10 Gbps” might only offer 10 Gbps speeds on a limited number of ports, with the remaining ports operating at 1 Gbps. Ensure the switch meets your 10 Gbps port requirements.
VII.2. Managed vs. Unmanaged Network Switches: Configuration Control
Network switches are broadly categorized into managed and unmanaged types.
-
Managed Switches: Offer extensive configuration options, allowing for granular control over network traffic, security, and features. A primary benefit of managed switches is the ability to create virtual LANs (VLANs). VLANs logically segment your network, enhancing security by isolating traffic between different device groups. For example, you could create a separate VLAN for IoT devices, preventing them from directly communicating with your primary computers and servers.
-
Unmanaged Switches: Function as plug-and-play devices with no configuration required. They simply forward network traffic between connected devices. While simpler to set up, they lack advanced features and security controls. In an unmanaged switch environment, all devices on the network can communicate freely with each other.
Initially, I leaned towards an unmanaged switch, preferring simplicity and avoiding the perceived complexity of managed switch configuration. I had not previously utilized managed switches and aimed for a straightforward setup.
However, the available 10 Gbps switches that met my other criteria were predominantly managed switches. Reluctantly, I opted for a managed switch. To my surprise, delving into VLAN configuration proved to be an engaging and beneficial learning experience. The ability to segment my network and enhance security with VLANs has become a valuable asset, and I now plan to implement VLANs more extensively throughout my homelab.
VII.3. Power over Ethernet (PoE): Simplifying Device Power
Power over Ethernet (PoE) is a technology that allows certain low-power devices to receive both power and data connectivity through a single Ethernet cable. This eliminates the need for separate power adapters, simplifying cabling and installation, particularly for devices located in hard-to-reach areas.
For example, my Ruckus R310 WiFi access point supports PoE. Using a PoE-enabled switch, it only requires a single Ethernet cable for both network connectivity and power.
My Ruckus R310 WiFi access point supports PoE, so it only needs a single Ethernet cable for power and data.
To utilize PoE devices, you need a PoE-enabled network switch. PoE switches provide dedicated power circuitry on specific ports, delivering power to compatible devices.
The tradeoff with PoE switches is that they typically consume more power themselves and are generally more expensive than non-PoE switches. If you don’t have PoE-compatible devices, a PoE switch would be an unnecessary expense and potentially contribute to higher energy consumption. However, if you utilize PoE devices like wireless access points or IP cameras, a PoE switch can significantly simplify cabling and installation.
VII.4. Port Density: Current and Future Needs
Determining the required number of ports on your network switch involves assessing both your current wired devices and anticipating future expansion. At a minimum, the switch must have enough ports to accommodate all your existing wired network devices.
The more complex question is how many extra ports to factor in for future growth. While you can always add more switches later, doing so consumes additional rack space (1U per switch) and adds to cable clutter. Ideally, you want a single switch with sufficient ports to handle current and near-future needs.
Consider your homelab growth plans over the next few years. If you anticipate adding more servers, network devices, or wired clients, it’s prudent to choose a switch with higher port density to avoid needing to replace or supplement it soon after initial setup.
I currently have eight devices requiring Ethernet connections. To accommodate potential future additions and provide ample headroom, I targeted switches with at least 16 ports, aiming for a 24-port model if feasible.
VII.5. Network Switch Candidates: Comparing Features and Prices
| Brand | Model | Ports | Speed | Managed | PoE | Price |
|---|---|---|---|---|---|---|
| TP-Link | TL-SG3428X | 24 | 4x10Gbps 24x1Gbps | Yes | No | $299.00 |
| Microtik | CRS328-24P-4S+RM | 28 | 4×10 Gbps SFP+ 24x1Gbps | Yes | Yes | $490.50 |
| TP-Link | Unnamed Chinese Model | 18 | 2×10 Gbps SFP+ 16 x 2.5 Gbps | No | No | $499.90 |
| TP-Link | T1600G-28TS | 24 | 4×10 Gbps SFP 24x1Gbps | Yes | No | $299.00 |
| TP-Link | T1600G-28PS | 24 | 4×10 Gbps SFP 24x1Gbps | Yes | Yes | $295.99 |
| TP-Link | T1700G-28TQ | 24 | 4×10 Gbps SFP 24x1Gbps | Yes | No | $958.40 |
I have experimented with Microtik products in the past and appreciate their feature richness and independent company ethos. However, I found their 90s-style web interface (RouterOS) to be unintuitive and cumbersome to navigate.
I want to like Microtik, but I can’t get over their weird 90s-style admin UI
Conversely, I’ve had consistently positive experiences with unmanaged TP-Link switches, finding them reliable and easy to use. This positive brand experience inclined me towards TP-Link for my managed switch needs.
I briefly considered a 16-port 2.5 Gbps TP-Link switch. However, its availability only through Chinese retailers and lack of US safety certifications raised concerns about safety and compliance. I decided against this option due to potential risks.
The TP-Link T1600G-28PS was attractive as it offered PoE in addition to the features of the TL-SG3428X. However, user reviews reported loud fan noise, which was a significant concern for my home office environment. Ultimately, I chose the TL-SG3428X, prioritizing silent operation and planning to add a separate, quieter PoE switch for my limited PoE device needs.
VII.6. Network Switch Review: TP-Link TL-SG3428X
- Grade: B-
Overall, the TP-Link TL-SG3428X switch is a solid performer. Its silent operation is a major advantage, particularly in a home office environment. Reliability has been excellent, with no unexpected downtime or performance issues.
However, the TP-Link web-based administrative interface is less user-friendly than desired. While functional, it lacks the intuitiveness and visual clarity of some competitors’ interfaces. Configuring VLANs, for example, required a significant amount of trial and error and referencing online guides. My experience debugging VLANs on the TP-Link switch highlights the interface’s shortcomings. Comparing it to QNAP’s VLAN control interface reveals a stark contrast in user experience and ease of configuration.
This page in the TP-Link web UI shows which ports are members of the Guest VLAN, but it always takes me a few minutes to remember how to interpret the screen.
VII.7. Network Switch Review: Netgear GS116LP 16-Port Unmanaged PoE Switch
- Grade: A
Initially, I intended to use a small, 5-port PoE switch placed on a shelf to power my limited number of PoE devices. However, as my workplace prepared to relocate our office, I repurposed a Netgear GS116LP switch from the office setup for my homelab rack.
As an unmanaged switch, the Netgear GS116LP is straightforward and reliable. It efficiently powers my PoE devices, was simple to install, and operates silently. For basic PoE functionality without complex configuration, it’s an excellent choice.
In retrospect, seeking a single managed switch that incorporated PoE ports might have been a more integrated solution, reducing the need for two separate switches. The benefits of a single PoE-enabled managed switch are further discussed later in this guide.
VIII. 10G Network Interface Cards (NICs): Unlocking 10 Gbps Speeds
Simply having a 10 Gbps switch is only half the battle. To fully realize 10 Gbps network speeds, each device intended to operate at 10 Gbps must be equipped with a 10 Gigabit Network Interface Card (NIC). Standard 1 Gbps NICs will function with a 10 Gbps switch, but their speed will be capped at 1 Gbps.
Sourcing 10G NICs for home use presented unexpected challenges. The primary market for 10G NICs is enterprise data centers, resulting in high prices for new cards, often approaching $1000 each. Used NICs offered a more budget-friendly alternative, typically under $100. However, finding compatible used NICs required extensive forum research and verifying compatibility reports from other users.
I successfully installed a used 10G NIC in my Windows desktop after some initial troubleshooting. However, attempts to install 10G NICs in my TrueNAS storage server were unsuccessful. I tested three different NIC models:
-
Mellanox ConnectX-3 CX311A
- Initially, the card was not recognized in my Windows desktop. No activity lights, and no detection in Device Manager.
- A forum post suggested trying a different PCI slot. Skeptical, I tried it, and surprisingly, it resolved the issue on my desktop.
- However, the TrueNAS server failed to recognize this NIC in any PCI slot.
-
Chelsio T520-LL-CR
-
Chelsio Dual Port T520-CR
My best guess is that the consumer-grade ASUS motherboard in my TrueNAS server has limited compatibility with these enterprise-class 10G NICs. Enterprise NICs may have specific hardware or firmware requirements not fully met by consumer-grade motherboards.
Future plans include building a new storage server with a motherboard more explicitly compatible with enterprise 10G NICs, allowing me to utilize the three 10G NICs I’ve accumulated.
Currently, the only 10 Gbps link in my network is between my Windows desktop and the managed switch. While this provides a high-speed connection for specific tasks, the full potential of the 10 Gbps switch remains partially unrealized until more devices are upgraded with 10G NICs.
IX. Uninterruptible Power Supply (UPS): Protecting Against Power Outages
During my time living in Manhattan, power outages were a recurring inconvenience, averaging about five per year. While most were brief, they were long enough to cause unexpected system shutdowns and potential data loss.
To mitigate the impact of power outages, I invested in a battery backup system, also known as an Uninterruptible Power Supply (UPS). My chosen UPS was an APC BR1500G, which has served reliably for six years.
In short power interruptions, the UPS provides seamless power, preventing system downtime. For extended outages, the battery backup offers sufficient time to gracefully shut down servers and workstations, safeguarding against data corruption and hardware damage.
The primary drawback of the UPS was the added cable complexity. With devices distributed across my office, connecting everything to the UPS resulted in a network of bulky power cables, detracting from the desired cable management improvements.
IX.1. Graceful Shutdown Time: Battery Capacity Considerations
For prolonged power outages, the UPS battery runtime dictates the time available to perform a graceful shutdown of connected systems before battery depletion. The required runtime depends on the UPS battery capacity and the combined power consumption of the devices it supports.
While I could have meticulously measured the wattage of each device using a Kill A Watt meter to precisely calculate power draw, I opted for a more pragmatic estimation based on my previous UPS experience.
My APC UPS, with an 865W battery, reported a 12-minute runtime while powering my desktop, VM server, storage server, firewall, and network switch. Based on this, I determined that an 800W minimum battery capacity would be adequate for my rack setup.
IX.2. UPS Alerting Capabilities: Automated Shutdown Notifications
After setting up my rack, a colleague pointed out that modern UPS systems often possess network alerting capabilities. These UPS units can communicate with devices on the local network, sending shutdown signals when a power outage occurs, enabling automated graceful shutdowns.
For my home environment, the added complexity of configuring automated shutdowns via UPS alerts seemed unnecessary. However, in environments with critical systems or frequent power disruptions, this feature can be highly valuable. Consider your tolerance for manual shutdowns and the frequency of power outages in your area when evaluating UPS alerting features.
IX.3. UPS Candidates: Comparing Rack-Mountable Options
| Brand | Model | Power | Outlets | Price |
|---|---|---|---|---|
| CyberPower | CP1500PFCRM2U | 1000 W | 8 | $335 |
| Tripp Lite | SMART1500LCD | 900 W | 8 | $298 |
| CyberPower | CPS1500AVR | 950 W | 8 | $460 |
IX.4. UPS Review: CyberPower CP1500PFCRM2U
- Grade: A
The CyberPower CP1500PFCRM2U UPS is a well-designed and user-friendly unit. The LCD display provides valuable real-time metrics on power consumption, battery status, and runtime. The display can also be turned off to minimize rack lighting, which is a thoughtful detail for home environments.
The UPS reports a 30-minute battery runtime when powering my VM server, storage server, firewall, and network switch, with a combined typical power draw of 200W. This runtime provides ample time for graceful shutdowns during extended outages.
CyberPower offers PowerPanel Business, a free UPS management software suite. Connecting a computer to the UPS via USB enables monitoring and configuration through the PowerPanel interface. While a Raspberry Pi would be an ideal platform for running PowerPanel as a dedicated UPS monitoring appliance, CyberPower unfortunately does not offer ARM-compatible versions of the software.
I rate CyberPower’s UPS management software as “okay.”
While I briefly explored PowerPanel, I found it somewhat clunky and ultimately unnecessary for my needs. The physical controls on the UPS itself provide sufficient functionality for basic operation.
PowerPanel does offer the capability to execute custom scripts upon power loss, potentially enabling automated shutdowns of rack-mounted devices. However, for my current setup, the manual shutdown process is manageable and preferred over the complexity of software-based automation.
Update (2024-04-07): Several readers recommended the Network UPS Tools (NUT) project, an open-source alternative to proprietary UPS vendor software. NUT offers broader compatibility and greater flexibility for integrating UPS monitoring and management into homelab environments. While I haven’t tested NUT yet, it appears to be a promising option for those seeking open-source UPS management solutions.
A significant positive attribute of the CyberPower UPS is its silent operation. While silent operation might be expected for battery backups, not all UPS units are equally quiet.
IX.5. UPS Review: Tripp Lite SMART1500LCD
- Grade: D
My initial UPS purchase for the server rack was the Tripp Lite SMART1500LCD. However, this unit proved to be excessively noisy, making it unsuitable for a home office environment.
I was surprised to discover that UPS units could generate noticeable noise during normal operation. My previous APC UPS was completely silent except during battery failover events.
The Tripp Lite UPS, in contrast, was the loudest component in my entire rack, and audibly intrusive throughout my house. The constant fan noise resembled a hairdryer running continuously in my office, easily audible even a floor away in my wife’s office.
Initially, I questioned if I had received a defective unit. Surely, a UPS designed for home or office use wouldn’t be intentionally engineered to be so loud?
I contacted Tripp Lite customer support and provided a video demonstrating the noise. Their response confirmed that the noise level was within the expected operating parameters of the unit and considered “normal.”
Despite attempting to acclimate to the noise, it proved too distracting for sustained use. After two days, I decided to return the Tripp Lite UPS.
To my surprise, Newegg’s return policy for the UPS was “replacement only.” My previous positive return experiences with Newegg led me to overlook checking the specific return policy beforehand. It appears they have stricter policies for heavier items like UPS units.
Fortunately, Newegg customer service, upon polite request, granted a full refund, reinforcing my continued preference for Newegg as a reliable retailer.
X. Power Strip Selection: Distributing Power Efficiently
Even with a UPS providing multiple power outlets, a dedicated power strip within the rack remains a valuable addition. Power strips offer extra outlets for non-essential devices that don’t require battery backup during power outages.
For example, I house a small IoT device in my rack that monitors my solar panel performance. This device is non-critical and can tolerate brief power interruptions. In fact, during a power outage, I prefer it to shut down to conserve limited UPS battery capacity for essential servers and network equipment.
X.1. Power Strip Candidates: Comparing Features and Outlets
Power strips are relatively straightforward components, and extensive shopping around wasn’t a priority. I primarily considered two rack-mountable power strip options.
| Brand | Model | Outlets | Price |
|---|---|---|---|
| Tripp Lite | RS-1215-RA | 12 | $78 |
| CyberPower | CPS1215RMS | 12 | $60 |
X.2. Power Strip Review: Tripp Lite RS-1215-RA
- Grade: B+
The Tripp Lite RS-1215-RA power strip has performed reliably. The rear-facing outlets are adequately spaced, accommodating bulky “wall wart” style power adapters without blocking adjacent outlets.
While I haven’t regularly used the front-facing outlets, they provide convenient access for temporary device connections or testing purposes.
X.3. Power Strip Review: CyberPower CPS1215RMS
- Grade: C
I purchased the CyberPower CPS1215RMS power strip several years ago for my work office rack. The primary drawback is the closely spaced outlets. Many office devices utilize larger power bricks, which often obstruct neighboring outlets on this power strip.
XI. Rack Shelf Selection: Accommodating Non-Rackmount Equipment
I had several pre-existing components from my previous homelab setup that were not rack-mountable. To integrate these devices into the rack, I needed rack shelves. I estimated needing at least 2U of shelf space.
XI.1. Rack Shelf Candidates: Comparing Brands and Prices
| Brand | Model | Price |
|---|---|---|
| Pyle | PLRSTN62U 19″ 2U | $64 for two |
| StarTech | CABSHELFV 2U 16″ | $88 for two |
XI.2. Rack Shelf Review: Pyle PLRSTN62U Rack Shelves
- Grade: A
Pyle was an unfamiliar brand to me, but their rack shelves, discovered online, have proven to be a surprisingly good value.
I keep non-mountable components on two 2U Pyle rack shelves.
These shelves are easy to install, reasonably priced, and feature a raised lip around the perimeter, preventing equipment from sliding off accidentally.
XI.3. Rack Shelf Review: StarTech CABSHELFV 2U Rack Shelves
- Grade: D
Initially, I opted for StarTech shelves, given their strong reputation in the server equipment market. However, upon installation, I encountered a peculiar design choice. The StarTech shelves incorporate a downward-facing lip that extends into the rack unit below.
StarTech shelves have a downward-facing lip that messes up the rest of the rack layout.
This downward lip effectively makes a “2U” shelf consume 3U of rack space if you intend to maintain consistent rack unit spacing. Alternatively, you could shift all subsequent equipment down by 0.5U, disrupting the vertical alignment.
The purpose of this downward lip was unclear. If it were upward-facing, it would logically serve as a retaining lip to prevent items from sliding off. However, the downward orientation seemed to offer no structural benefit or practical purpose.
Update (2024-04-09): Helpful readers pointed out that the downward lip likely enhances the shelf’s structural rigidity, preventing sagging in the shelf’s center under load. While this may be the intended purpose, the design still compromises rack unit spacing and overall equipment layout.
Searching online reviews, I found other users mentioning this design quirk, but seemingly accepting it without major concern. Comments often had a tone of “it takes up a bit more than 2U, whatever.”
Reviewer acknowledges that StarTech’s 2U rack shelf takes up 3U of space, still rates it 4 out of 5.
I found this design compromise unacceptable – a 2U shelf occupying 3U of valuable rack space. I promptly returned the StarTech shelves and replaced them with the Pyle shelves, which adhere to standard 2U dimensions.
XII. Patch Panel Selection: Organizing Network Cable Fronts
While browsing homelab setups online, patch panels frequently appeared as a standard component. However, their purpose remained unclear until I began building my own rack.
XII.1. Understanding Patch Panels: Cable Management and Organization
Initially, patch panels seemed perplexing. Just a row of empty slots? What’s the point?
The concept clicked into place during rack assembly. A patch panel is a cable management device that creates a demarcation point for network cable runs. Instead of directly connecting every device to the network switch with long cables running across the front of the rack, you use short patch cables to connect the switch to the patch panel. Then, longer network cables run from the back of the patch panel to your devices.
Patch panels keep the front of your rack clean by routing network cables to the rear of the rack.
This approach significantly cleans up the front of the rack, hiding long cable runs behind the panel and creating a more organized and professional appearance. Patch panels also simplify cable management and troubleshooting, as you can easily identify and reroute connections at the patch panel level without accessing the switch directly.
Recommendation: Position a patch panel adjacent to each network switch in your rack for optimal cable management and organization.
XII.2. Patch Panel Candidates: Comparing Port Density and Features
| Brand | Model | Price |
|---|---|---|
| NewYork Cables | 24-Port 1U | $19 |
| Tripp Lite | 16-Port 1U | $13 |
XII.3. Patch Panel Review: NewYork Cables 24-Port 1U Patch Panel
- Grade: B+
Patch panels are relatively simple devices, primarily consisting of metal and plastic. The NewYork Cables patch panel is well-constructed and feels sturdy. Installation into the rack was straightforward.
One advertised feature of the NewYork Cables patch panel was a rear cable management bar, intended to support Ethernet cables. However, in my setup, this bar proved ineffective. It was positioned too close to the Ethernet ports to provide meaningful support, and the cables themselves didn’t seem to require additional support.
I thought the rear bar of the patch panel would help support my Ethernet cables, but they turned out not to need it.
My minor critique is the port labeling system. It uses paper slips under plastic covers, reminiscent of old landline phone speed dial labels. Some patch panels utilize small whiteboard strips, which allow for erasable labeling and greater flexibility. I prefer the whiteboard strip style for easier label updates.
The labels on the NewYork patch panel are like the little speed dial labels that landline phones had in the 90s.
XII.4. Patch Panel Review: Tripp Lite 16-port 1U Patch Panel
- Grade: A
Similar to the NewYork Cables patch panel, the Tripp Lite model is functional and well-made. Patch panels are generally unremarkable devices.
However, the Tripp Lite patch panel utilizes small whiteboard labels, which I find significantly more convenient than paper slips. I had standard whiteboard markers initially, but their tips were too broad for the small label areas. Purchasing ultra fine-tip whiteboard markers resolved this, providing precise and erasable labeling.
The Tripp Lite patch panel features tiny labels you can write on with ultra fine-tip whiteboard markers.
XIII. Raspberry Pi Rack Mount Selection: Integrating Single-Board Computers
As a frequent user of Raspberry Pi single-board computers for both professional and hobby projects, a rack mount for Raspberry Pis seemed like a fun and practical addition to my homelab rack. I opted for the first decent-looking Raspberry Pi rack mount I encountered without extensive comparison shopping.
| Brand | Model | Price |
|---|---|---|
| UCTRONICS | Ultimate Rack with PoE Functionality | $190 |
XIII.1. Raspberry Pi Rack Mount Review: UCTRONICS Ultimate Rack
- Grade: C+
The UCTRONICS Ultimate Rack Raspberry Pi rack mount is an adequate, but not exceptional, product.
For the price, it offers reasonable value. Considering that individual PoE HATs for Raspberry Pi 4 typically cost around $20, the inclusion of four PoE HATs in this rack mount already accounts for a significant portion of the total cost. Additionally, the rack mount includes microSD extenders, HDMI extenders, OLED screens, and cooling fans for each Pi slot.
However, the build quality is mediocre. Component fit is not precise, resulting in noticeable gaps around the HDMI and microSD ports.
The UCTRONICS Pi rack mount had significant gaps around the HDMI and microSD ports.
The HDMI ports are weakly secured to the mount. Plugging in an HDMI cable causes noticeable connector bending and strain, raising concerns about potential long-term durability.
While PoE capability is beneficial, it generates additional heat. The integrated fans are intended to mitigate heat buildup, but they produce a constant, high-pitched whirring noise. I opted to disable the fans to eliminate the noise, accepting potential CPU throttling or shutdown under heavy load as a tradeoff, which is acceptable for my hobby projects.
Assembly instructions are poorly written and confusing. Step three abruptly jumps to simultaneously assembling five remaining components, lacking clear guidance and making assembly more challenging than necessary.
The UCTRONICS Pi rack mount instructions rapidly increase in difficulty.
XIV. Ethernet Cable Selection: Patch Cables and Length Considerations
Transitioning to a server rack setup often necessitates purchasing new Ethernet cables, especially if your previous setup was less organized. For patch panels, short Ethernet cables (6-12 inches), often called “patch cables,” are essential to connect the patch panel ports to the adjacent network switch ports.
You will likely need a mix of patch cable lengths. The distance between corresponding ports on the switch and patch panel can vary depending on the port layout of each component and their vertical positioning within the rack. For instance, on my rack, port 16 on the switch is only 1.5 inches from port 16 on the patch panel, while port 1 on the switch is 6 inches from its corresponding patch panel port.
Consider that the distance between ports on your switch and patch panel may vary depending on the port layout of each component.
I purchased 6-inch, 12-inch, and 3-foot Ethernet cables in a ratio of approximately 5:2:1 to accommodate the varying port distances in my rack.
Some enthusiasts use color-coded Ethernet cables to visually differentiate network segments or device types. I opted for standard blue and black Ethernet cables for a clean and professional appearance, prioritizing uniformity over color-coded organization.
XV. Fiber Cable Selection: 10 Gbps Connectivity Options
XV.1. Ethernet, DAC, or Fiber: Choosing the Right 10 Gbps Medium
For 1 Gbps networks, standard RJ45 Ethernet cables are the straightforward choice. However, exceeding 1 Gbps requires considering alternative cabling options: Ethernet or fiber.
Ethernet cabling for 10 Gbps (10GBase-T) utilizes RJ45 connectors and Cat6a or higher-rated copper cables. Fiber optic cabling for 10 Gbps (10GBase-SR/LR) utilizes fiber optic cables and connectors like LC or SFP+ Direct Attach Copper (DAC) cables.
While Ethernet cabling offers familiarity, fiber optic cabling becomes increasingly attractive at 10 Gbps and beyond due to its greater bandwidth capacity, longer transmission distances, and immunity to electromagnetic interference.
Fiber optic networking introduces a slightly more complex cabling paradigm. Fiber network devices typically use SFP or SFP+ ports, which do not directly accept fiber cables. Instead, you need transceivers to convert the SFP/SFP+ port to a compatible fiber connector type.
For my 10 Gbps network, both my switch and NICs used SFP+ ports. Therefore, the connection path involved:
- SFP+ port on the network switch
- SFP+ to connector type transceiver
- Connector type cable
- SFP+ to connector type transceiver
- SFP+ port on the 10G NIC
To bridge the SFP+ ports, I considered three primary connector types:
- RJ45 (Ethernet)
- LC (Fiber)
- DAC (Direct Attach Copper)
Since my rack setup included patch panels, the chosen connector type needed to be compatible with patch panel integration. Patch panels are readily available for Ethernet (RJ45) and fiber (LC) connectors, but I found no readily apparent patch panel solutions for DAC cables. It’s possible DAC patch panels exist or alternative integration methods exist, but they were not immediately obvious.
This consideration narrowed my options to RJ45 (Ethernet) or LC (Fiber) for my 10 Gbps connections.
XV.2. Ethernet vs. Fiber: Performance, Cost, and Practicality
Comparing RJ45 (Ethernet) and LC (Fiber) options for 10 Gbps, I found no significant practical performance differences for my homelab environment. LC fiber cables are slightly thinner and visually cleaner, which is a minor aesthetic advantage. However, transitioning to fiber meant introducing a different cable type alongside my existing Ethernet infrastructure.
The most surprising difference was in cost. SFP+ to RJ45 transceivers were considerably more expensive than SFP+ to LC fiber transceivers. Conversely, Ethernet cables are generally less expensive than comparable length fiber optic cables.
However, when factoring in the total cost of transceivers, cables, and patch panel components, fiber optic cabling proved to be significantly more economical:
| Component | Ethernet price | Fiber price |
|---|---|---|
| Three transceivers (for switch, desktop, storage server) | $150 | $60 |
| One 16’ cable (desktop to switch) | $9 | $15 |
| One 3’ cable (storage server to switch) | $7 | $10 |
| Two 7″ patch cables | $0* | $30 |
| Four patch keys | $0* | $19 |
| Total | $163 | $134 |
* These Ethernet patch cables and patch keys were already factored into the cost for the remaining 1 Gbps ports on the switch, so they are effectively “free” for this comparison.
Important Note: When using fiber optic cabling, ensure compatibility between transceivers and fiber cables in terms of “mode.” Single-mode fiber and multi-mode fiber are incompatible. For 10 Gbps over shorter distances typical in a homelab, multi-mode fiber is generally recommended. Multi-mode fiber is less expensive than single-mode and adequately supports 10 Gbps speeds for homelab distances. Verify that your transceivers and fiber cables are both multi-mode (typically OM3 or OM4).
My final rack setup section lists all the fiber cables and transceivers I purchased for my 10 Gbps network links.
XVI. Pre-Existing Components: Integrating Existing Hardware
XVI.1. Router: Qotom Q355G4 with OPNsense
My home network router is based on a compact Qotom Q355G4 mini PC running the open-source OPNsense firewall distribution. As the Qotom unit is not rack-mountable, it resides on a dedicated rack shelf within the server rack.
My OPNsense firewall running on Qotom Q355G4 mini PC
XVI.2. WiFi Access Point: Ruckus R310
My WiFi access point, a Ruckus R310, is not physically located in the server rack but connects to the PoE switch within the rack. The Ruckus R310 is a high-performance access point that supports multiple SSIDs with VLAN tagging, enabling network segmentation for guest WiFi and other segregated networks.
My Ruckus R310 WiFi access point
XVI.3. Out-of-Band Management: TinyPilot Voyager 2a PoE
Full Disclosure: TinyPilot is a product I created and sell.
For out-of-band management, I utilize TinyPilot Voyager 2a PoE. While primary server access is via SSH or web interfaces, TinyPilot provides essential hardware-level access for tasks like OS reinstallation, BIOS configuration changes, and network troubleshooting.
Instead of physically connecting a keyboard and monitor to the rack, TinyPilot Voyager 2a enables remote KVM (Keyboard, Video, Mouse) access through a web browser.
I use TinyPilot to get physical-level access to my homelab servers through the browser.
XVI.4. Software Testing: Dell Optiplex 7040 Mini PC
For TinyPilot development and testing, a Dell Optiplex 7040 mini PC serves as a dedicated test device. Its compact size and readily available hardware make it ideal for software testing, allowing for frequent reboots and OS reimaging without disrupting my primary development environment.
XVII. Rack Component Arrangement: Layout and Best Practices
Once components were selected, the next step was planning the physical arrangement within the rack. While no universally standardized layout exists, logical principles and best practices guide component placement.
To visualize and plan the layout, I used a spreadsheet, color-coding components by type. This spreadsheet also aided in determining the necessary rack unit height and overall rack size.
I considered different rack layouts by just swapping elements in a spreadsheet.
XVII.1. Bottom-Heavy Weight Distribution: Stability First
A fundamental principle in rack layout is placing heavier components at the bottom. This lowers the center of gravity, maximizing rack stability and reducing the risk of tipping, especially when the rack is moved or accessed. A toppled rack can cause significant equipment damage and potential injuries.
The UPS, being the heaviest component at 27 lbs, was logically positioned at the bottom of my rack. Patch panels and network switches are relatively lightweight, allowing them to be placed higher in the rack without compromising stability.
XVII.2. Front Connection Proximity: Cable Management Optimization
Clustering components with front-facing connections, like patch panels and network switches, is crucial for efficient cable management. Placing them adjacent to each other minimizes patch cable lengths and avoids cables needing to run across other components, contributing to a cleaner and more organized front panel.
XVII.3. Rear Cable Management: Less Critical in Home Environments
Some guides emphasize minimizing power cable lengths within the rack. However, in a home environment, the difference between a 2-foot vs. 4-foot power cable is negligible. Cable length optimization becomes more critical in large data centers with hundreds or thousands of racks, where even small cable length reductions accumulate into significant material and space savings. For a home rack, rear cable lengths are less of a concern.
XVIII. My Final Rack Setup: Components and Satisfaction
| Component | Choice | Price | Satisfaction |
|---|---|---|---|
| Server rack | StarTech 4POSTRACK18U | $316 | A |
| Network switch (managed) | TP-Link TL-SG3428X | $299 | C+ |
| Network switch (PoE, unmanaged) | Netgear GS116LP | $139 | A |
| UPS | CyberPower CP1500PFCRM2U | $335 | A+ |
| Power strip | Tripp Lite RS-1215-RA | $78 | B+ |
| Rack shelves | Pyle PLRSTN62U 19″ 2U | $64 | A |
| Patch panel (24-port) | NewYork Cables 1U | $19 | B+ |
| Patch panel (16-port) | Tripp Lite 1U | $13 | A |
| Raspberry Pi rack mount | UCTRONICS Ultimate Rack with PoE Functionality | $190 | C+ |
| Total | $1,453 |
Smaller components and accessories:
XIX. Future Rack Enhancements: Next Steps and Ideas
XIX.1. Rack-Mounted Server Chassis
A notable omission from my current rack is a dedicated rack-mounted server chassis. My existing VM and storage servers are still in their pre-rack tower cases. Migrating them to rack-mounted chassis is planned for future upgrades, but was deferred to keep the initial rack build manageable in scope.
XIX.2. Rack Top “Hat” or Enclosure
Currently, the top of my rack is open and unused. I’m searching for a compatible “hat” or enclosure for the rack top. Ideally, this would be a secure-fitting top piece that provides a flat surface for storing lightweight items or accessories on top of the rack, maximizing vertical space utilization. If you have suggestions for rack top solutions, please share them.
XX. Avoiding Common Mistakes: Lessons Learned
XX.1. Pre-Test the UPS Before Rack Mounting
The UPS was the most physically challenging component to install in the rack due to its weight and bulk. Mounting a heavy UPS is awkward and requires careful maneuvering. Avoid the frustration of mounting a UPS only to discover it’s defective or excessively noisy. Always test the UPS functionality and noise level before rack installation.
XX.2. Noise-Conscious UPS Selection
UPS noise levels vary significantly. Some are virtually silent, while others generate constant fan noise. If the rack will be located in a noise-sensitive area like a home office, prioritize UPS models with low noise ratings or user reviews confirming quiet operation. Checking UPS reviews for noise complaints is crucial.
XX.3. Verify Return Policies
Before purchasing components, especially heavier items like UPS units, carefully review the retailer’s return policy. Newegg’s “replacement only” policy on the Tripp Lite UPS was unexpected, as my previous experiences had been with easy returns for refunds. Proactive policy checks prevent surprises and ensure flexibility if a component proves unsuitable. Checking return policies is a vital step in the purchasing process.
XX.4. Single PoE-Enabled Managed Switch: A More Integrated Approach
Currently, my rack uses two separate switches: a managed switch for core networking and an unmanaged PoE switch for PoE devices. This adds complexity and consumes extra rack space. Get a PoE-enabled switch if you have any PoE components. In retrospect, searching more diligently for a single managed switch that also offered PoE and silent operation would have been a more integrated and streamlined solution. Ideally, a fanless managed switch with PoE ports and 10 Gbps uplinks would be optimal.
XX.5. Cage Nut Installation: Effortless and Pain-Free
Cage nuts, used to secure equipment to rack rails, are designed for easy, tool-less installation. If you find yourself struggling or experiencing pain while installing cage nuts, you are likely using incorrect technique. Cage nuts aren’t supposed to hurt. Correctly installed, cage nuts clip into the rack rail from behind, eliminating the need to hold them while inserting screws.
How I was incorrectly installing cage nuts
Tip: If hardware installation is physically painful or requires excessive force, reassess your technique. Server equipment is designed for relatively easy installation by IT professionals of varying physical fitness levels.
The correct way to install cage nuts is to let them clip in from behind the hole in the rack post.
XX.6. Patch Key Orientation: Correct Insertion for Secure Connections
Patch panel keys, used to create RJ45 ports in patch panels, have a specific orientation for proper installation. Incorrect installation can lead to loose connections and patch keys dislodging when cables are removed. Don’t install patch keys backward.
My initial attempts involved incorrect patch key installation, resulting in keys easily popping out when disconnecting Ethernet cables.
Subsequent attempts involved installing keys from the rear, which was slightly better but still incorrect.
Embarrassingly, I thought this was how RJ45 patch keys were supposed to work for about six months.
Correct patch key installation involves aligning the key with the patch panel slot and pushing until a click is heard, indicating the key’s locking tab has engaged. The front face of the key should be roughly flush with the patch panel surface.
Patch keys should be flush with the face of the patch panel, and their tabs click into place in the rear.
XX.7. PCI Slot Compatibility for 10G NICs
If a motherboard fails to detect a newly installed 10G NIC, try a different PCI slot. In my case, a Mellanox 10G NIC was not detected in one PCI slot on my desktop motherboard, but worked perfectly in a seemingly identical adjacent PCI slot. PCI slot compatibility issues, while uncommon, can occur.
XX.8. Fiber Mode Compatibility: Multi-mode vs. Single-mode
Mixing single-mode and multi-mode fiber optic cables and transceivers is a common error that can lead to intermittent network connectivity or complete link failure. Don’t mix SFP+ multimode and single mode fiber cables.
My initial 10 Gbps setup experienced intermittent connection drops due to inadvertently using single-mode patch cables with multi-mode transceivers and fiber cables.
My network connection went into a reset loop every 24 hours because I accidentally used multimode patch cables in a single mode fiber system.
Always verify that all fiber optic components (transceivers, cables, patch cables) are compatible in terms of fiber mode (multi-mode or single-mode). For homelab 10 Gbps setups, multi-mode fiber is generally recommended for cost and performance reasons.
XX.9. Consider Used Equipment: Budget-Conscious Homelabbing
While this guide primarily focuses on new equipment, exploring the used equipment market can significantly reduce rack building costs. Consider used equipment. Marketplaces like eBay, Facebook Marketplace, and Craigslist often offer used server racks, switches, UPS units, and other homelab components at substantial discounts compared to new equipment. While used equipment requires more research and potentially involves some risk, the cost savings can be considerable for budget-conscious homelab builders.
XXI. Life with a Server Rack: Organization and Functionality
I am highly satisfied with my new server rack. It has transformed my previously disorganized office setup into a streamlined and functional homelab environment. The rack has eliminated the wire clutter and provided a central, organized location for all my server and networking equipment.
The aesthetic improvement is also notable. The rack presents a more professional and less cluttered appearance to visitors, replacing the “weird slob with cables everywhere” look with a more “quirky nerd” persona.
The physical proximity of all devices within the rack has also enhanced the usability of TinyPilot (again, disclosure: TinyPilot is a product I created). Previously, TinyPilot was kept on the floor, requiring cumbersome cable swaps and repositioning for each server access. Now, with TinyPilot rack-mounted and adjacent to all servers, accessing any device for low-level management is quick and seamless. This improved accessibility has been invaluable for tasks like installing NixOS on a Raspberry Pi and upgrading my VM server to Proxmox.
Building a server rack was a worthwhile investment that has significantly improved the organization, functionality, and aesthetics of my homelab. For anyone serious about home servers and networking, a DIY server rack is a transformative upgrade.
