Before boot: prove the installer image
The v4.0.1 release publishes a SHA-256 for the ISO. A matching digest proves the downloaded bytes match that release artifact; it does not prove the USB write, port, memory, or target drive. In the ThinkPad report, SQUASHFS read errors and a successful older-ISO retry changed the useful hypothesis, but did not prove whether the original download, media write, USB device, or release was at fault.
What this means: Retain the ISO filename, published and calculated SHA-256, download URL, imaging tool and verification result, USB make/model, port, and whether the same media reaches the live environment on another machine. Stop if the digest differs.
Basis: Official Omarchy code or documentation · Omarchy v4.0.1
Before or at GRUB: firmware and bootloader
The v4.0.1 manual requires Secure Boot and/or TPM to be disabled for installation, so their state belongs in every report. But a specific error outranks a generic prerequisite: the Framework Ryzen AI report repeatedly reached a GRUB out-of-memory failure and says Secure Boot, UMA allocation, TPM visibility, and unloading the TPM module did not change it.
What this means: Photograph the firmware version and relevant settings, then retain the literal bootloader error and whether the menu ever appeared. Do not keep cycling unrelated firmware toggles when an exact repeatable error points to a narrower boot stage.
Basis: Model-specific user report · Framework Ryzen AI 9 HX 370 v4.0.1 installer boot report
After the kernel starts: live display path
A restart or freeze before the installer UI can be a graphics path, but the GPU vendor alone is not a diagnosis. One Dell i7-8650U report freezes after kernel load on Intel i915. A separate RTX 3060 desktop report attributes an earlier restart to Nouveau and reports that a temporary nomodeset parameter reached the next phase. Neither report supplies a universal safe default.
What this means: Record whether the bootloader, kernel messages, cursor, or installer UI appeared; list every GPU and which output the monitor uses. Treat a temporary boot parameter as a diagnostic branch and retain the before/after symptom, not as a permanent fix copied across machines.
Basis: Model-specific user report · Intel i915 live-ISO freeze on Dell Latitude E7490
Inside the installer: read, write, and partition errors
Classify the operation that failed before interpreting the message. SQUASHFS read error -5 appeared in the ThinkPad media path. In contrast, ENOSPC during a full-disk install on two Intel Macs occurred while Archinstall erased a ChromeOS Flex GPT; the retained example had ample capacity, and both installs completed after the old partition maps were removed.
What this means: Retain the full installer phase and traceback plus lsblk, filesystem, partition, target model, and capacity evidence. A read-only ISO showing 100% usage and a target-disk ENOSPC are not interchangeable. Never run an erasing workaround until the exact target is independently identified and recoverable.
Basis: Model-specific user report · False ENOSPC during full-disk install over ChromeOS Flex
Installer finished, first boot stops in firmware or Limine
Installation success does not prove a usable firmware entry. On one Asus ProArt with separate Linux and Windows NVMe drives and ESPs, the expected UKI existed on disk but no direct Omarchy firmware entry pointed to it. Limine panicked; a separately created direct entry then booted the installed system.
What this means: Before reinstalling, retain the firmware boot list, each ESP's disk and partition identity, the literal Limine or UEFI error, and whether the UKI and configuration files exist. Keep a read-only capture before changing boot entries.
Basis: Model-specific user report · Missing first-boot UKI entry on a dual-NVMe Asus ProArt
Bootloader succeeds, screen goes black before login
That phase is later than a firmware or missing-entry failure. On one Acer Ryzen 7 8845HS hybrid, repeated comparisons found that splash plus a password-prompting greeter produced the black screen while removing splash changed the result. The report also records the Radeon iGPU as panel owner and the RTX 4050 Max-Q as having no attached display.
What this means: Record the last service or frame before black, whether a TTY is reachable, autologin and greeter state, every DRM device, panel ownership, and the persistent source of kernel parameters. A black screen after Limine is not evidence that the ISO never booted.
Basis: Model-specific user report · Hybrid AMD and NVIDIA black screen after the bootloader
A later hang appears only after graphics changes
One failure can hide the next. In the RTX 3060 desktop report, bypassing the reported live graphics failure exposed later RTL_OCP_GPHY_COND messages and a time-synchronization wait attributed to a Realtek NIC using r8169. The report says a temporary module blacklist reached an offline install; exact NIC PCI ID and independent reproduction are still missing.
What this means: Preserve the ordered symptom sequence. Record the NIC PCI ID, driver, link state, last systemd unit, network availability, and whether the same hang occurs with the adapter disabled. Do not label the whole machine an NVIDIA failure after the boot advances to a network-specific message.
Basis: Model-specific user report · Live ISO and first-boot NVIDIA plus Realtek report
Hardware suspicion: use controlled comparisons
A hardware diagnosis needs controls. In the Z790 report, a service center confirmed the original Intel CPU failed stress testing; after replacement, Ubuntu and other distributions installed, while Omarchy still did not. Multiple USB devices, NVMe targets, memory arrangements, and BIOS defaults had also been tried. That isolates one real CPU fault and a remaining Omarchy-specific failure without pretending they share one cause.
What this means: Change one dimension at a time and record both positive and negative controls: same media on another machine, another distribution on the same hardware, memory test, target-drive swap, firmware defaults, and the exact phase that moves. Cross-distribution success narrows the question; it does not prove Omarchy compatibility.
Basis: Model-specific user report · Install failures before and after an Intel CPU replacement