NUT.CONF(5)

💖 NUT.CONF(5) 💖 — like, the NUT Manual, but make it make sense

okay so bestie, welcome to nut.conf — the glow-up file that tries to standardize all the chaotic config files floating around different installs. we're talking /etc/default/nut on Debian-based systems and /etc/sysconfig/ups on RedHat-based systems. it's giving "can we please all agree on one thing" energy. 🎀

An init (or service method) script provided by an OS distribution should source this file in order to determine which components have to be started; it may also be consulted by nutshutdown integration hook script and others. so yeah, this file is basically the group chat moderator for your UPS setup. 💅

Blank lines are ignored. Lines with a hash (#) character at the first position of the line are ignored, too. They can be used to add comments. so go ahead, leave your little notes, we love a documented vibe. ✨

---

🚨 IMPORTANT NOTES 🚨 (read these, fr fr)

• This file is intended to be sourced by shell scripts as well as by service management frameworks like systemd on Linux:
  • There is no guaranteed export VAR=VAL syntax. so don't come crying to me when your export doesn't export. 😭
  • No guaranteed expansion of variables like VAR1="$VAR2-something" — only verbatim assignments. like, it's not gonna do the math for you, bestie. 💀
  • You may need to export VAR when sourcing it into init-scripts or other scripts, for eventual propagation of certain settings to NUT programs. Not-exported variables can only be consumed by the script which "sourced" the file (and may choose to export them independently). so if you want your fave settings to travel, you gotta send them out properly. 🫶
• You MUST NOT use spaces around the equal sign! like, MODE = standalone is a no. MODE=standalone is the vibe. periodt. ✋
• Practical support for this file and its settings currently varies between different OS packages and NUT sample scripts, but should converge over time. so it's a little messy rn, but we're working on it. 🌈
• Contents of this file should be pure ASCII (character codes not in range would be ignored with a warning message). so no fancy unicode emoji in the config file itself, save those for the group chat. 🙈

Refer to the EXAMPLE section for illustrations. (yes, there's a whole example down there, keep scrolling babe.)

---

✨ DIRECTIVES ✨ (the actual settings, slay)

• MODE — Required. Recognized values are none, standalone, netserver and netclient. Defaults to none. this is like the main character of the file, you gotta pick one. 👑
  • none — Indicates that NUT should not get started automatically, possibly because it is not configured or that an Integrated Power Management or some external system, is used to startup the NUT components. basically "don't start nothin', we're chillin'." 😴
  • standalone — Addresses a local only configuration, with 1 UPS protecting the local system. This implies to start the 3 NUT layers (driver, upsd and upsmon), with the related configuration files. This mode can also address UPS redundancy. so it's like a solo mission but with backup. 💪
  • netserver — Like the standalone configuration, but also possibly need one or more specific LISTEN directive(s) in upsd.conf. Since this MODE is open to the network, a special care should be applied to security concerns. so like, you're serving the network now, bestie — lock your doors. 🔐
  • netclient — When only upsmon is required, possibly because there are other hosts that are more closely attached to the UPS, the MODE should be set to netclient. you're just here for the updates, not the whole party. 👀
• ALLOW_NO_DEVICE — Optional, defaults to false. Set this to true to allow starting the upsd NUT data server service even if ups.conf has no device sections configured at the moment. This environment variable overrides the built-in "false" flag value in the upsd program, and an optional same-named default flag that can be set in upsd.conf. If you want a data server always running and responding on the network, even if it initially has nothing to serve (may be live-reloaded later, when devices become configured), this option is for you. like, it's the "open for business even if the shelves are empty" setting. 🏪
• ALLOW_NOT_ALL_LISTENERS — Optional, defaults to false. Set this to true to allow starting the upsd NUT data server even if not all LISTEN directives can be honoured at the moment. This environment variable overrides the built-in "false" flag in the upsd program, and an optional same-named default flag that can be set in upsd.conf. If you want a data server always running, even if it would potentially not serve all clients on every uptime, this option is for you (note you would have to restart upsd to pick up the LISTENed IP address if it appears later). Probably configuring LISTEN * is a better choice in such cases. so like, it's the "we'll figure it out as we go" energy, but maybe just use the wildcard instead. 🤷‍♀️
• UPSD_OPTIONS — Optional. Set upsd specific options. See upsd(8) for more details. It is ignored when MODE above indicates that no upsd should be running. so don't bother setting this if upsd isn't invited to the party. 🎉
• UPSMON_OPTIONS — Optional. Set upsmon specific options. See upsmon(8) for more details. It is ignored when MODE above indicates that no upsmon should be running. same vibe as above — only relevant if upsmon is actually doing its thing. 🕺
• POWEROFF_WAIT — Optional. At the end of an emergency system halt, the upsmon primary will signal the UPS to switch off. This may fail for a number of reasons. Most notably is the case that mains power returns during the shutdown process. See the section "Power races" in /usr/share/doc/nut/FAQ.txt.gz. The system will wait this long for the UPS to cut power, and then reboot. It should be long enough to exhaust the batteries, in case line power continues to be unavailable. On the other hand, it should not be so long that the system remains offline for an unreasonable amount of time if line power has returned. See sleep(1) for compatible time syntax. If you specify the time in seconds, use the "s" suffix. so it's like setting a timeout for the UPS to actually cut the power — too short and it might not work, too long and you're waiting forever. 😬

⚠️ Warning ⚠️ This workaround might be dangerous under some circumstances. Please read http://bugs.debian.org/358696 for more details. like, fr fr, read the bug report before you go messing with this. 💀

• POWEROFF_QUIET — Optional, defaults to false. This setting controls if the NUT shutdown integration scripts or service units would emit messages about their activity (or lack thereof). By default they may be verbose, to aid in post-mortem troubleshooting via logs or console captures. Set to true to avoid that trove of information, if you consider it noise. so if you're not about that chatty log life, set this to true and enjoy the silence. 🤫
• NUT_DEBUG_LEVEL — Optional, defaults to 0. This setting controls the default debugging message verbosity passed to NUT daemons. As an environment variable, its priority sits between that of DEBUG_MIN setting of a driver and the command-line options. so it's like the middle child of debug settings — not the most important, not the least. 🧚
• NUT_DEBUG_PID — Optionally add current process ID to tags with debug-level identifiers. This may be useful when many NUT daemons write to the same console or log file, such as in containers/plugins for Home Assistant, storage appliances... so when everyone's talking at once, you can tell who said what. 🗣️
• NUT_DEBUG_PROCNAME — Optionally log process name (or however the program chose to identify itself). This may be useful when multiple NUT daemons log into the same file or console, without syslog to prefix the name into each line (e.g. in tests, single init script systems like Home Assistant, etc.) so again, name tags for your logs. 🏷️
• NUT_DEBUG_SYSLOG — Optional, unset by default. Normally NUT can (attempt to) use the syslog or Event Log (WIN32), but the environment variable NUT_DEBUG_SYSLOG allows to bypass it, and perhaps keep the daemons logging to stderr (useful e.g. in NUT Integration Test suite to not pollute the OS logs, or in systemd where stderr and syslog both go into the same journal). Recognized values:
  • stderr — Disabled and background() keeps stderr attached
  • none — Disabled and background() detaches stderr as usual
  • default — Not disabled
  • unset/other — Not disabled

so if you're like "please don't touch the system logs," you got options. ✨

• NUT_IGNORE_CHECKPROCNAME — Optional, defaults to false. Normally NUT can (attempt to) verify that the program file name matches the name associated with a running process, when using PID files to send signals. The NUT_IGNORE_CHECKPROCNAME boolean toggle allows to quickly skip such verification, in case it causes problems (e.g. NUT programs were renamed and do not match built-in expectations). This environment variable can also be optionally set in init-scripts or service methods for upsd, upsmon and NUT drivers/upsdrvctl. so if you renamed stuff and NUT is like "who dis?", this is your escape hatch. 🚪
• NUT_QUIET_INIT_UPSNOTIFY — Optional flag to prevent daemons which can notify service management frameworks (such as systemd) about passing their lifecycle milestones, to not report loudly if they could NOT do so (e.g. running on a system without a framework, or misconfigured so they could not report and the OS could eventually restart the false-positively identified "unresponsive" service. Currently such reports, done by default, help troubleshoot service start-up and highlight that NUT sources (or package build) did not take advantage of tighter OS service management framework integration (if one exists, so that developers could focus on adding that). Reasons to set this flag could include platforms without such a framework and not expecting one, although nagging your favourite OS or contributing development to make it better is also a way. so like, if you don't have the framework and don't want the noise, set this. but also, maybe help make it better? just saying. 💜

---

🌸 EXAMPLE 🌸 (yes, actual code, copy responsibly)

# /etc/nut/nut.conf.  See nut.conf(5)

MODE=none

UPSD_OPTIONS=""

UPSMON_OPTIONS=""

# POWEROFF_WAIT=15m

look at that clean little config. we love a minimal moment. 🎀

---

🦋 INTEGRATION 🦋

An init script, such as /etc/init.d/nut, is expected to source this file in order to determine which components have to be started. so yeah, the init script reads this file like it's the group chat pinned message — that's where the instructions live. 💌

---

(part 1 done, but we're not signing off — more coming, bestie!)

✨ SEE ALSO ✨

okay so like, this page doesn't live alone bestie, it's literally in a whole friend group 💅 go read the rest of the squad:

• ups.conf(5) — where you actually define your UPS devices 🎀
• upsd.conf(5) — the daemon config, the server-side icon ✨
• upsd.users(5) — users and their permissions, gatekeeping done right 🔥
• upsmon.conf(5) — the monitoring side of things, main character energy 👑

🌈 Internet resources

the NUT (Network UPS Tools) home page is right here, no cap:

https://www.networkupstools.org/historic/v2.8.5/

---

and that's the whole page, bestie — you ate that 🫶 this one's from Network UPS Tools 2.8.5, dated 06/21/2026 (yes that's the source/version and date footer, consider it your little receipt) 💖✨🎀