OPINIONATED OMARCHY
TROUBLESHOOTING CORPUS
790
RECORDS
790
AUDITED
0
UNCHECKED
2388
SOURCES
LIVE
display-monitors high common DISPUTE

Recover a monitor that stays black because it was switched off at boot

omarchyarchendeavouroscachyoshyprlandwaylanddesktop
CORRECTED An auditor found the fix wrong or incomplete and rewrote it.
AUDIT NOTE

Checked against Hyprland v0.56.2 source, the kernel DRM sysfs source, the installed omarchy 4.0.2-1 package on this workstation, and the Arch KMS wiki. Four defects. First, the danger names systemd-boot /boot/loader/entries/*.conf and GRUB /etc/default/grub, neither of which exists on a stock Omarchy 4 install. Confirmed on this machine that Omarchy 4 boots a UKI through Limine: /etc/limine-entry-tool.d/omarchy-defaults.conf and omarchy-uki.conf are owned by omarchy-settings 4.0.2-1 and carry KERNEL_CMDLINE[default]+=, and /usr/lib/limine/limine-common-functions lines 99 to 129 load /usr/share/limine-entry-tool.d/*.conf, then /etc/limine-entry-tool.d/*.conf, then /etc/default/limine at highest priority. /usr/bin/limine-update runs limine-install --no-efi-register and limine-mkinitcpio, which is what rebuilds the UKI. Second, the cause claims Hyprland never retries. It retries three times at one second intervals: CMonitor::scheduleModeRetry in src/output/Monitor.cpp with MAX_MODE_RETRIES = 3, then gives up permanently. More importantly the record omits that Omarchy 4 already recovers this by itself. omarchy-hyprland-monitor-watch is launched from /usr/share/omarchy/default/hypr/autostart.lua, and its recover_modeless function polls omarchy-hyprland-monitor-modeless and issues hyprctl reload on a backoff capped at 60 seconds until a mode appears. That process is live here as PID 2316. Third, the off then on then detect sysfs sequence is wrong and dangerous. status_store in drivers/gpu/drm/drm_sysfs.c maps off to DRM_FORCE_OFF, on to DRM_FORCE_ON and detect to clearing the force, and re-probes whenever the force changed or is zero, so only the final detect does useful work while the first two leave a connector forced until something writes detect back. That same function emits no hotplug event, which is exactly why the record's hyprctl reload is needed and why powering a panel on changes nothing on its own. Fourth, the record gave no route in when the black monitor is the only one, which is the whole point of the problem. On duplication with monitor-config-ignored-name-or-mode: only the cause sentence overlaps. This record uniquely holds the detection test, the DRM re-probe, the kernel video= pin and the Omarchy watcher. The sibling holds mode and scale matching for rules that are accepted but ignored. Merging is an operator call, but they are not the same failure. omarchy-hw-recover-internal-monitor is NOT the recovery for this and is not cited here, correctly: read in full, it only removes ~/.local/state/omarchy/toggles/hypr/internal-monitor-disable.lua when omarchy-hw-external-monitors reports no external panel, driven by the oneshot user unit omarchy-recover-internal-monitor.service before graphical-session-pre.target. It has nothing to do with a modeless output. Confirmed live on this machine: the four DRM connectors are card1-DP-1, card1-DP-2, card1-DP-3 and card1-HDMI-A-1, only HDMI-A-1 connected, and hyprctl monitors all -j carries the name, width, height and disabled fields the detection and verify commands read. NOT exercised: nothing was written, no reload or re-probe was run, and no monitor was powered off, so the recovery sequence itself is reasoned from source rather than performed. I also could not list /boot to confirm the Limine fallback entry, because it is a vfat ESP mounted dmask=0077 and I was not to use sudo.

The cause above was rewritten on 2026-09-13 to match this note; the fix was corrected by the audit itself.
RISK

Editing kernel parameters touches the bootloader, and a mistake there can stop the machine booting. On Omarchy 4 that is Limine with a UKI: the drop-ins are /etc/limine-entry-tool.d/*.conf and /etc/default/limine, and sudo limine-update rebuilds /boot/EFI/Linux/omarchy_linux.efi. There is no /boot/loader/entries and no /etc/default/grub on a stock Omarchy 4 install, so advice naming those paths is written for a different system. Omarchy's /etc/limine-entry-tool.d/omarchy-uki.conf sets ENABLE_LIMINE_FALLBACK=yes and a BOOT_ORDER that includes fallback and Snapper snapshot entries, so the boot menu should offer something older, but keep a live USB handy rather than relying on that. Separately, writing off or on to a connector's /sys/class/drm/*/status forces that connector's state until you write detect back or reboot. Aim it at the wrong connector and you blank a monitor that was working.

Symptom

A monitor that was switched off, or on a different input, when you booted stays black afterwards even once you power it on. hyprctl monitors all -j lists it as enabled with a width and height of 0. Pressing the monitor's own power button changes nothing, because that fires no hotplug event. Unplugging and replugging the cable often does recover it, since that does fire a hotplug, but on a laptop dock or a panel you cannot reach that is not an option.

Cause

A display that is switched off, or on another input, answers the probe with a partial EDID that carries no video modes. Hyprland brings the output up anyway with no mode, so it reports 0x0 and stays black. It does retry, but only three times at one second intervals (CMonitor::scheduleModeRetry, MAX_MODE_RETRIES = 3, src/output/Monitor.cpp on v0.56.2), and then gives up for good. Powering the panel on afterwards fires no DRM hotplug event, so nothing asks the compositor to look again, and a reload is the only thing that re-reads the connector. Omarchy 4 fills that gap in userspace: omarchy-hyprland-monitor-watch runs a recover_modeless loop that polls omarchy-hyprland-monitor-modeless and issues hyprctl reload on an exponential backoff capped at 60 seconds until the output reports a mode.

Fix

**On Omarchy 4 this usually recovers itself.** omarchy-hyprland-monitor-watch is started from /usr/share/omarchy/default/hypr/autostart.lua and runs for the life of the session. Power the monitor on, or select the right input, and wait up to a minute. Check the watcher is alive first, from a working screen or over ssh:


pgrep -af omarchy-hyprland-monitor-watch

**Detect the state.** This is the exact test Omarchy's own omarchy-hyprland-monitor-modeless runs, and it works on any Hyprland:


hyprctl monitors all -j | jq -c '.[] | select(.disabled != true and (.width == 0 or .height == 0)) | {name,width,height}'

**If the black monitor is your only screen, you need another way in.** The session is running fine and only that output has no mode, so ssh from another machine is the reliable route. Ctrl+Alt+F2 is not: the console draws on the same dead connector, so the TTY is black too.

**Manual recovery on any Hyprland 0.55+ with a Lua config.** Power the monitor on, then reload. A reload is what re-reads the connector, and it needs no root:


hyprctl reload

**If a reload is not enough, force a DRM re-probe.** This needs root. Write detect and nothing else: per status_store in drivers/gpu/drm/drm_sysfs.c, detect clears any forced state and re-probes, while off sets DRM_FORCE_OFF and on sets DRM_FORCE_ON and either one sticks until something writes detect back. List your real connector names first:


for p in /sys/class/drm/*/status; do echo "$p: $(cat "$p")"; done

echo detect | sudo tee /sys/class/drm/card1-HDMI-A-1/status   # use your own connector
hyprctl reload

The write re-probes but sends no hotplug event, so the hyprctl reload is still needed.

**Pin an explicit mode** in ~/.config/hypr/monitors.lua so Hyprland does not depend on the EDID. When the named resolution is in no driver mode list, Hyprland falls through to trying it as a custom DRM mode, which the driver may still reject:


hl.monitor({ output = "HDMI-A-1", mode = "1920x1080@60", position = "auto", scale = 1 })

**Last resort: force the mode from the kernel command line**, so it is set before the compositor starts:


video=HDMI-A-1:1920x1080@60

*Omarchy 4 (Limine with a UKI).* Add a drop-in rather than editing a packaged file. /etc/limine-entry-tool.d/*.conf and /etc/default/limine are both read by /usr/lib/limine/limine-common-functions, the latter at highest priority:


printf 'KERNEL_CMDLINE[default]+=" video=HDMI-A-1:1920x1080@60"\n' \
  | sudo tee /etc/limine-entry-tool.d/99-force-monitor-mode.conf
sudo limine-update

limine-update runs limine-install --no-efi-register and limine-mkinitcpio, which rebuilds the UKI at /boot/EFI/Linux/omarchy_linux.efi with the new command line. Confirm afterwards with cat /proc/cmdline on the next boot.

*Plain Arch.* systemd-boot: append it to the options line in /boot/loader/entries/*.conf. GRUB: append it to GRUB_CMDLINE_LINUX_DEFAULT in /etc/default/grub, then run sudo grub-mkconfig -o /boot/grub/grub.cfg.

Verify

hyprctl monitors all -j | jq -c '.[] | {name,width,height}' reports a non-zero width and height for the output, and the screen shows a picture. On Omarchy 4, omarchy-hyprland-monitor-modeless exits 1 when no enabled output is modeless, 0 while one still is, and 2 when the compositor cannot answer.

Sources

monitor-powered-off-at-boot-black-0x0