eRacks Systems Tech Blog
Open Source Experts Since 1999
Your Hardware, Your Choice: Lock-In from Ink Cartridges to Server Drives
Printers taught a lot of us what lock-in feels like, and the business model behind it is simple: sell the printer cheap, then make the money back on the ink, cartridge after cartridge, for as long as you own it. The printer is the loss leader; the supplies are the business. HP’s CEO, Enrique Lores, said as much on CNBC in January 2024: “Every time a customer buys a printer, it’s an investment for us. We are investing in that customer, and if that customer doesn’t print enough or doesn’t use our supplies, it’s a bad investment” (as quoted by TechSpot).
The way to protect that investment is to make sure the cheaper cartridges don’t work. A third-party cartridge runs for a while, then a firmware update stops it. HP’s own support document says a cartridge with a non-HP chip “may function for a period of time, but then cease to function” (HP, “Dynamic Security Enabled Printers”), and HP pushed another round of that firmware on January 29, 2026 (Ars Technica).
TVs run the same play with a different follow-on. As Bruce Perens asked on LinkedIn last month: “Why can you buy a 65 inch 4K TV for $250 and a smaller and less capable computer monitor can cost $1000? Because most of the income from the TV is fees for recording your TV viewing preferences” (LinkedIn). With the printer, the ink pays for the hardware. With the smart TV, your viewing habits do.
We’ll never buy or recommend another HP printer, and this is why.
What the follow-on is worth
Supplies were about 65 percent of HP’s printing revenue last year (HP’s fiscal 2025 annual report). HP’s CFO described the value of “locking that person” into a longer relationship as a 20 percent uplift on the customer (as quoted by TechSpot). HP+ printers, once activated, require original HP ink “for the lifetime of the printer” (HP). Branded inkjet ink runs $9 to $11 per milliliter (Forbes, November 1). When we priced toner for our own new printer this week, the maker’s high-yield set cost about 6.7 times a well-reviewed aftermarket set.
We don’t mind a brand premium. If a maker’s own supplies are better, people will pay for them. What we object to is firmware that removes the customer’s choice: subsidizing the hardware, then policing the supplies to win the money back. Compete; don’t restrict or mandate.
The security argument, checked
HP says cartridges can carry viruses. Ars Technica traced that to one lab-only bug, found through HP’s own bug bounty program, published and patched in 2022, with no known attacks in the wild. The lockout itself started in 2016, six years earlier (Ars Technica). If a cartridge chip can attack a printer, the fix is in the printer’s firmware, not in locking out the customer.
Regulators have mostly punished the secrecy, not the lock. Australia’s consumer regulator had HP compensate customers in 2018 for not disclosing it (ACCC), and Italy fined HP 10 million euros in 2020 (AGCM). Courts in the Netherlands and the US have so far let the practice stand. So the real check is what buyers choose.
The same playbook in servers, where it costs real money
- Branded drives. In April 2025 Synology said its new Plus NAS models would fully support only its own drives or certified ones. After pushback it restored third-party hard drives on those desktop models in October 2025, but its current policy still limits unlisted M.2 SSDs on every model, and unlisted hard drives on its rack and enterprise models, to migration only (Synology).
- CPUs fused to one vendor. AMD’s Platform Secure Boot (PSB) can permanently bind an EPYC or Threadripper PRO processor to one system maker’s firmware key. Dell’s support page says enabling it “will permanently fuse the CPU to only work on Dell platforms” (Dell), and Lenovo did the same with every Threadripper PRO 5000 in its P620 workstation (Tom’s Hardware). A used CPU pulled from one of those machines may never run in another board.
- Fans for “foreign” cards. Dell’s 13th-generation PowerEdge servers raise fan speed by default when they see a third-party PCIe card. Dell documents how to turn that off (Dell).
- Coded network optics. Cisco switches can shut a port down for a third-party SFP module (the small plug-in transceiver for fiber or copper) until you enable an “unsupported transceiver” command (Cisco).
- Paying to switch on what you own. Intel sold activation of features already built into 4th and 5th generation Xeon processors under “Intel On Demand.” It has since been wound down and is not on Xeon 6 (HotHardware).
- Licenses that limit where hardware runs. NVIDIA’s GeForce driver license says the software “is not licensed for datacenter deployment” (NVIDIA), and its virtual GPU features are licensed per user, per year.
Where buyers won
Synology backed off within six months for its desktop models. Intel’s On Demand is gone from Xeon 6. And Oregon, Colorado and Minnesota now limit “parts pairing,” where a maker refuses a working replacement part because it didn’t come from them. Pushback works.
How to spot lock-in before you buy
- Will it run third-party drives with full features: health reports, firmware updates, new storage pools?
- Do the fans ramp up, or warnings appear, when you add a card or drive of another brand? Can you turn that off?
- Are BIOS and firmware updates free, or tied to a support contract?
- Is the CPU fused to the system maker? Will it run in another board if the motherboard dies?
- Do the switch ports take standard optics?
- Are features already in the silicon sold as paid unlocks?
- Does a software license limit where the hardware can run?
- Can you roll firmware back if an update breaks something?
Not every answer has to be “no lock-in” for a purchase to make sense. But you should know the answers before you buy, not after the first update.
How we build
The printer and smart-TV model hides the real price of the hardware in what comes after it: the cartridges, the subscription, the data. We price the other way around. Our systems are priced for what they are, with no subsidy to win back later through locked supplies or subscriptions, so there is nothing to lock. Our home page has said it for years: no vendor lock-in, standard off-the-shelf components, your choice of operating system. We build on standard boards, drives, memory and cards you can buy anywhere, and on open-source operating systems. And adding or changing components doesn’t void our warranty. It’s your machine.
Questions about a build, or about lock-in in something you already own? Write to us.
joe November 7th, 2026
Posted In: News, Technology
Tags: AMD PSB, eRacks, NAS, open hardware, printers, right to repair, Servers, vendor lock-in
Refurbished, Renewed, and Burned In: What We Do to a Used GPU Before It Ships
This week a customer’s IT team asked us for “more details on your refurb process”. One of them had already asked an AI assistant, which summarized our process from our own website: a 72-hour burn-in, fans checked, thermal pads refreshed. It got most of it right. Here is the full answer, from us.
Why refurbished at all
Our new systems ship with new parts, bought through authorized US distribution and validated-new sources, and that stays the default. Some customers ask for something else on purpose. The AMD Threadripper 3000 series is a good platform that distributors no longer carry. The GeForce RTX 3090 is the last GeForce card with NVLink (NVIDIA’s bridge that links two cards’ memory), which matters for some training work, and a pair of them costs a fraction of a current professional card. For builds like these the parts can only come refurbished or renewed, and we are glad to build them.
The words cover a lot of ground, though. “Refurbished” can mean a part that was properly reworked, or a used part that was wiped down and relisted. Our components page describes the other end of that market: “clean pull” lots with a short broker warranty, and server CPUs locked to their first owner’s firmware. We don’t buy from that market. Our refurbished cards come from refurbishers who back them with a warranty, and then every refurbished part goes through our own process.
What we do to a refurbished part
- Inspection. Every part is checked when it arrives, and anything that doesn’t meet our standards goes back. Not every “refurbished” or “renewed” part passes; we have rejected and returned some.
- Rework on the cards. We check the fans and replace them as needed, and we always replace the thermal compound or pads (the material that carries heat from the chip to the cooler). Worn connectors, plugs or jacks get replaced, and visibly marginal solder joints get resoldered. Then the card is cleaned and reassembled.
- Assembly. Everything is double- and triple-checked, neat, clean and properly cabled, and every card has to be recognized by its driver before we go further.
- Operating system and test suite. The OS and drivers go on, with our own utilities and our testing and burn-in suite.
- Burn-in. 72 hours of testing under full load, CPU, memory and every GPU together, including overnight runs and several multi-hour burn-ins, with temperatures logged and the driver logs watched for errors. If a part fails, we replace it and restart the burn-in from zero.
- The AI stack. On an AI server the stack goes on last, and the system is benchmarked with real models and tested before it ships.
The 72 hours are the same burn-in every eRacks system gets. The 24-hour minimum on our components page is typically for rush orders.
What it means for you
When a quote includes a refurbished or renewed part, it says so, line by line. Refurbished GPUs carry a 1-year warranty. And the system they go into gets the same assembly, the same checks and the same 72-hour burn-in as a system built entirely from new parts.
Looking at a previous-generation build, a pair of NVLinked cards, or an AI server on a budget? Browse our AI rackmount servers, or write to us at info@eracks.com and tell us what you need.
joe October 29th, 2026
Posted In: News, Technology
Tags: AI server, burn-in, eRacks, refurbished GPU, renewed hardware, RTX 3090, Threadripper
eRacks/GAMER: A Linux Gaming PC That Boots Straight Into Steam
We have built Linux machines since 1999, and for most of that time gaming was the one thing people told us Linux couldn’t do. That changed. Valve’s Proton (a compatibility layer built into Steam that runs Windows games on Linux) and the Steam Deck turned Linux into a real gaming platform. So we built a proper desktop gaming model, with all the bells and whistles: eRacks/GAMER.
What it is
eRacks/GAMER is a tower built to order, starting at $4,395 with an AMD Ryzen 7 9700X, an AMD Radeon RX 9070 XT with 16GB, 32GB of DDR5-6000 RGB memory and a 2TB WD Black NVMe drive (fast solid-state storage). It sits in a panoramic tempered-glass case, black or white, with three ARGB (addressable RGB) fans and a 360mm ARGB liquid cooler, on an AM5 X870 motherboard with Wi-Fi 7 and 2.5Gb Ethernet, powered by an 850W ATX 3.1 supply.
From there the configurator goes where you want it: Ryzen X3D processors up to the 16-core Ryzen 9 9950X3D2 (X3D chips carry AMD 3D V-Cache, extra on-chip cache that AMD aims at games), GeForce RTX 5070, 5070 Ti and 5080 cards, 64GB of memory, 4TB of storage, and the lighting set the way you like it: a synced rainbow, one color, or off.
It boots into Steam
GAMER ships with Bazzite, a Fedora-based gaming system that boots into Steam Gaming Mode, the full-screen, controller-friendly interface from the Steam Deck. Steam comes installed with Proton, along with Heroic for Epic and GOG libraries, Lutris, GameMode and MangoHud (an on-screen frame-rate and temperature overlay). Sign in with your own Steam account and your library is there.
Prefer something else? Ubuntu with Steam is in the menu, and so are Fedora, Nobara, CachyOS, Pop!_OS and Mint. Every one gets the same gaming stack, installed and tested before it ships.
Why not SteamOS?
We asked ourselves the same thing. SteamOS is Valve’s own, and Valve supports it on its own hardware plus one Lenovo handheld: its FAQ lists the Steam Deck, Steam Machine, Steam Frame and Lenovo Legion Go S as the only devices “Powered by SteamOS”. On desktop PCs, AMD graphics are in beta and NVIDIA isn’t supported, and Valve’s own SteamOS page says not to treat it as a replacement for a desktop operating system. Bazzite gives you a Steam Gaming Mode like SteamOS’s, plus NVIDIA drivers and a full desktop. If you want SteamOS anyway, it is an option on Radeon builds.
The games that lock Linux out
Proton runs Windows games on Linux. The exception is a handful of online games whose publishers have decided Linux players aren’t welcome: their anti-cheat software refuses to run on Linux, and you’re shut out for no reason except your choice of operating system. The community list at areweanticheatyet.com tracks them. As of today it shows Valorant and Fortnite as denied, and recent Call of Duty titles as broken.
We think that’s the wrong call. We’d rather send our customers to the studios that support Linux players, and there are plenty of great games that welcome you. We also understand that some games are compelling enough that people play them anyway, whatever their publisher’s appetite for control. If that’s you, it’s your machine: tell us what you want, Windows dual boot included, and we’ll build it that way.
On the couch, on the TV
Plug GAMER into the living-room TV and Steam’s Big Picture mode turns it into a console. Or leave it at your desk and stream to the TV with Steam Remote Play: Valve’s Steam Link app runs on Apple TV, Android TV and Amazon Fire TV. Add up to four Xbox, DualSense or 8BitDo controllers in the configurator, or Valve’s new Steam Controller by quote (Valve sells it through a reservation queue).
Streaming your games instead? Add a 4K capture card, a USB microphone, and Discord and OBS Studio preinstalled.
A word on prices
Graphics cards cost more than they did a year ago. On October 5, 2025 we priced the GeForce RTX 5070 Ti at $799 at Newegg; this morning the lowest new, in-stock 5070 Ti there was $1,169.99. The Radeon RX 9070 XT went from $799 in April to $879.99 today. That is one reason the RX 9070 XT is our default: it is the better match for Linux and SteamOS, and today it costs about $290 less than the 5070 Ti. The 5070 Ti is one click away if you want it.
For the longer version, with sources, see our Linux gaming PC guide. Ready to build one? Configure eRacks/GAMER, or write to us at info@eracks.com.
joe October 16th, 2026
Posted In: News, Technology
Tags: Bazzite, eRacks/GAMER, gaming PC, Linux gaming, Proton, Steam, SteamOS
How We Came to KVM: Tucson 2007, Napa 2008, and Why We Built eVirt
In our last post on virtualization I promised the story of how we came to KVM, the Kernel-based Virtual Machine. Here it is. It starts in a resort outside Tucson in August 2007, at the first KVM Forum, and it ends with the small piece of software our own servers still run on.
Tucson, August 2007
The first KVM Forum ran from August 29 to 31, 2007, at the Loews Ventana Canyon in Tucson. KVM had been in the mainline Linux kernel for about six months by then: it arrived in version 2.6.20 in February, as a driver for the hardware virtualization features in Intel and AMD processors. Avi Kivity, who started it at a small company called Qumranet, opened the forum with a talk titled “KVM, One Year On”.
I went expecting to learn more about the Linux kernel and about virtualization. What I found was a technology that was going to change where virtualization went. The default answer in 2007 was Xen, a separate hypervisor that ran underneath Linux. By my count at the time it was roughly ten times the code, and its host side lived outside the kernel. KVM took the opposite approach: Linux itself becomes the hypervisor. The scheduler, the memory manager and the device drivers are the ones the kernel already has, so every improvement to Linux is an improvement to the hypervisor for free. Sitting in that room, it was obvious to me that this would replace the virtualization engines of the day. It largely has.
Two moments that stuck
The first was meeting Avi Kivity and having drinks with the whole Qumranet crew, including Benny Schnaider, the CEO, who closed the forum with the final keynote. I was so impressed with Avi: his intuitive grasp of everything he was doing, and the mental bandwidth he had for all of it at once.
The second was in a session with Rusty Russell, the kernel developer who wrote ipchains and then iptables, the Linux packet filter behind countless firewalls. His talk in Tucson was about lguest, his deliberately minimal hypervisor, but the remark I remember was about iptables. He said, more or less, that iptables was never intended as an end-user product. It was a toolkit for building one. That changed how I looked at the whole stack. The kernel gives you good tools. The product layer on top should be small, opinionated and yours, and it should not try to re-wrap every tool underneath it.
Napa, June 2008
The second forum moved to the Napa Valley, June 11 to 13, 2008, and the agenda showed how fast things were moving: Mac OS X running in KVM, memory sharing between virtual machines, passing physical devices straight through to a guest. Three months later, in September 2008, Red Hat bought Qumranet for $107 million. KVM was no longer an interesting experiment. It was the future of Linux virtualization, and by then we were already running it.
What libvirt got wrong
The obvious way to manage KVM then, and still the way most guides assume, was libvirt. It did not work for us. In 2008 it was full of stubs and unimplemented features, and it rewrote the surface area for getting anything done, so that using it was more work than going directly to the QEMU commands underneath. QEMU is the machine emulator that KVM accelerates; it already had a perfectly good command line and monitor.
Then there was the XML. Every machine became an XML document, and XML as a configuration format is a disaster of invalid states: documents that parse but do not mean anything, settings that conflict, and no good way to see what the machine will actually do until you run it. The industry has largely moved on to simpler configuration languages since. To be fair, libvirt has picked up a few redeeming features over the years. In my view it is still not worth it, which is Rusty’s point exactly: a tool that tries to be the product ends up harder to use than the toolkit it wraps.
The first eVirt
So we built the opposite. The first eVirt did one thing: it launched virtual machines from a simple configuration file with well chosen defaults, talking to QEMU directly. It took a tiny fraction of the code of anything built on libvirt, and we have used it in production ever since. The current version, the fourth, keeps the same idea: a short YAML file per machine (CPU, memory, disks, bridge, fixed address), an Ubuntu cloud image seeded with cloud-init, and one command to build or rebuild it. Our own web, mail and lab servers are defined that way, and I described it in more detail in the last post.
What it taught us
Nearly two decades later, the lessons from those two forums are still how we build. Put the hard parts in the kernel, where they get the most eyes and the longest support. Keep the layer on top thin enough that you can read it, replace it, or skip it. And treat the management tools as a choice, not a commitment: a virtual machine disk is a file, and a good host lets you change your mind.
That is why our default virtualization host is still Ubuntu Server with KVM, and why we install whatever management layer a customer asks for: libvirt with Cockpit or virt-manager, Proxmox, Incus, IncusOS or XCP-ng. If you would rather run what we run, we install eVirt on a consulting basis.
Planning a virtualization build, or moving off a platform that has started sending license invoices? Write to us at info@eracks.com, or start with a quote request.
joe October 8th, 2026
Posted In: News, servers, Technology
Tags: eVirt, KVM, KVM Forum, libvirt, Linux kernel, QEMU, virtualization
The Apps Were Never the Box: Why We Build Open NAS Systems
An XDA Developers writer published a piece last week that we could have written ourselves. He kept his Synology NAS (network-attached storage: a box of drives on your network that serves files and apps) and walked away from almost every app Synology ships with it. His photos moved to Immich, his music to Navidrome, his movies to Plex and Jellyfin, his documents to Paperless. All of them run in Docker, a way to package an app with everything it needs so it runs the same on any Linux machine.
His reasons are specific. Synology Photos, in his words, is “pretty dated”: no major new features in a while, and missing the GPS tagging, import workflows and shared libraries Immich has, with a mobile backup he found less reliable. Synology’s music app is “not very good at all”. And Synology discontinued its own video streaming app, which leaves a basic DLNA server as the only first-party option.
We have been saying this for years. It is worth saying why.
A NAS is two things
Every NAS is storage plus the apps that use it. Appliance vendors sell the two as one product, and for a while that was a fair trade: the apps were the reason to buy the box. Docker ended that. The open-source apps now move faster than the vendors’ own, which is the writer’s first reason, and because they are containers, they do not care which hardware they run on.
The XDA writer makes the point himself. His apps are “as independent as possible”, so if he moves from Synology to another brand, or to his own server, Immich comes with him. A proprietary app cannot make that promise, and neither can a vendor that decides which of its own apps survive the next release.
Once the apps are open, the box is just storage. And storage should be open too.
What open storage looks like
This is how we build NAS systems, and have for a long time:
- ZFS by default. ZFS is an open-source file system that checksums every block, so it detects and repairs silent corruption instead of serving it back to you. Our NAS line defaults to RAIDZ2, dual parity: any two drives can fail without losing data.
- The drives talk to ZFS directly. We offer an HBA (host bus adapter, the card the drives plug into) in IT mode, which passes each drive straight through instead of hiding it behind a hardware RAID controller. ZFS sees the real disks and their health.
- Standard NAS drives. Our current default is the 24TB Seagate IronWolf Pro, a CMR drive (conventional magnetic recording, the kind that behaves well in a rebuild), bought in the channel like any other part. If a drive fails in year four, you buy another one anywhere.
- Your choice of operating system. Ubuntu LTS by default, or Debian, TrueNAS Community Edition, OpenMediaVault and others. The data outlives the choice: a ZFS pool exported from one machine imports on another running a current OpenZFS.
- Sizes that fit the job. From NAS4, a 1U four-bay unit, to NAS100, 102 bays in 4U, plus the QUIET3 for a home or an office where a rackmount fan would be too loud.
The same apps, without the vendor in the middle
Immich, Jellyfin, Plex, Navidrome and Paperless run on any of these systems exactly as they run on the XDA writer’s Synology. The difference is what sits underneath: an operating system you can update on your schedule, a file system you can move, and hardware with no vendor deciding which parts or apps you may use. Name the apps you want and we install and test them before the system ships, the same way we set up TrueNAS with Jellyfin on a quiet media server this month.
To be fair to Synology: if you want an appliance you never think about, and you are happy with its apps, it is a reasonable buy. The writer still uses his as the backbone of his setup. But if you are already running your own apps, as he is, part of what the appliance price pays for is software you have stopped using.
Where to start
The full NAS line is configurable online, and our July post covers how we re-verified every default drive and part against this year’s shortages. The XDA article is here.
Tell us what you store and which apps you run, and we will size it. Write to info@eracks.com, or start with a quote request.
joe September 24th, 2026
Tags: Docker, Immich, Jellyfin, NAS, Navidrome, Open Source, self-hosting, Synology, TrueNAS, ZFS