Proxmox UPS
We recently had a troubleshoot a Proxmox clustering issue. It appears the problem was linked to the communication between the Proxmox node and the UPS. This customer was not hosting Proxmox into a datacenter environment. They have their own computer room, with a physical rack.
There is a cluster of 3 proxmox node. Each power supply of each node is connected to an UPS (battery). The UPS is also connected to the node with a USB cable, to allow a proper shutdown of the node in case of prolonged electricity cut.
The right configuration with Proxmox NUT
# lsusb
Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
Bus 001 Device 002: ID 0665:5161 Cypress Semiconductor USB to Serial
# cat /etc/nut/ups.conf
[ups]
driver = "nutdrv_qx"
port = "auto"
vendorid = "0665"
productid = "5161"
bus = "001"
device = "002"
busport = "003"
desc = "UPS_Xanto"dsn=4.4.3, status=deferre>
Feb 15 07:56:58 pve02 nut-server[1911]: Data for UPS [ups] is stale - check driver
Feb 15 07:56:58 pve02 upsd[1911]: Data for UPS [ups] is stale - check driver
Feb 15 07:56:59 pve02 nut-monitor[1915]: Poll UPS [ups@localhost] failed - Data stale
Feb 15 07:56:59 pve02 nut-monitor[1915]: Communications with UPS ups@localhost lost
Feb 15 07:56:59 pve02 nut-monitor[1915]: UPS [ups@localhost] was last known to be calibrating and currently is not communicating, assuming dead
Feb 15 07:56:59 pve02 nut-monitor[1915]: Executing automatic power-fail shutdown
Feb 15 07:56:59 pve02 upsd[1911]: Client ups@127.0.0.1 set FSD on UPS [ups]
Feb 15 07:56:59 pve02 nut-server[1911]: Client ups@127.0.0.1 set FSD on UPS [ups]
Feb 15 07:56:59 pve02 nut-monitor[1915]: Auto logout and shutdown proceeding
Feb 15 07:56:59 pve02 nut-monitor[97858]: Network UPS Tools upsmon 2.8.1
Feb 15 07:56:59 pve02 nut-monitor[97863]: Network UPS Tools upsmon 2.8.1
Feb 15 07:57:04 pve02 nut-monitor[1915]: Network UPS Tools upsmon 2.8.1
Feb 15 07:57:04 pve02 systemd-logind[1126]: Creating /run/nologin, blocking further logins...
Feb 15 07:57:04 pve02 systemd-logind[1126]: System is powering down.
Feb 15 07:57:04 pve02 shutdown[97879]: Shutdown scheduled for Sun 2026-02-15 07:57:04 CET, use 'shutdown -c' to cancel.
Feb 15 07:57:04 pve02 nut-monitor[1914]: Network UPS Tools upsmon 2.8.1
Feb 15 07:57:04 pve02 systemd[1]: nut-monitor.service: Deactivated successfully.
Feb 15 07:57:04 pve02 systemd[1]: Stopped nut-monitor.service - Network UPS Tools - power device monitor and shutdown controller.
Feb 15 07:57:04 pve02 systemd[1]: nut-monitor.service: Consumed 8.480s CPU time, 3.7M memory peak.It appears to be a communication issue, cause there was no electrical issue at that time.
The solution
The following changes allowed to fix the problem:
We modified the following files:
/etc/nut/upsmon.conf: DEADTIME 60
/etc/nut/upsd.conf: MAXAGE 60
Restart nut and the problem will be fixed.
7 Benefits of Using an UPS with Proxmox Nodes
Running Proxmox VE in a homelab or production environment means you are hosting virtual machines and containers that expect stability. Power interruptions—whether brief outages or longer failures—can seriously impact data integrity and service availability. Pairing your Proxmox nodes with an Uninterruptible Power Supply (UPS), especially when integrated via NUT (Network UPS Tools), provides several important benefits.
We would recommend hosting Proxmox VE into a datacenter environment, which offers a strong and reliable electrical network. We offer such services for datacenter in Belgium.
Nevertheless, if you have specific requirements of having servers on-site (e.g. NVR or SCADA sites,..), here are the benefits of using UPS.
1. Protection Against Sudden Power Loss
The most obvious benefit of an UPS is protection from unexpected power cuts. Without an UPS, a sudden outage forces Proxmox nodes to shut down instantly, which can lead to:
- Corrupted virtual disks
- Damaged filesystems inside VMs or containers
- Inconsistent cluster state in multi-node setups
An UPS supplies temporary power, allowing your Proxmox hosts to keep running long enough to shut down cleanly.
2. Graceful Shutdown of Proxmox Hosts and VMs
When integrated with Proxmox using NUT, an UPS enables automated and orderly shutdowns. Instead of an abrupt stop:
- Virtual machines and containers are shut down cleanly
- Proxmox nodes power off safely once battery levels reach a threshold
- Cluster-wide shutdowns can be coordinated from a single UPS
This dramatically reduces the risk of data loss and avoids lengthy recovery operations after power is restored.
3. Improved Data Integrity
Virtualized workloads often include databases, file servers, and applications that perform frequent disk writes. Abrupt power loss during write operations can corrupt data or leave applications in an inconsistent state.
Using an UPS ensures that:
- Disk writes can complete properly
- Journaling filesystems remain consistent
- Backup jobs and replication tasks are not interrupted mid-process
This is especially important for ZFS-backed Proxmox storage, where clean shutdowns help maintain pool health.
4. Increased Hardware Longevity
Power fluctuations, brownouts, and frequent hard power-offs can shorten the lifespan of hardware components such as power supplies, disks, and SSDs.
An UPS helps by:
- Filtering voltage spikes and drops
- Reducing stress on power supplies
- Preventing repeated forced shutdowns
Over time, this contributes to more reliable hardware and fewer unexpected failures.
5. Better Availability for Short Outages
Not all power issues are long outages. Many are brief interruptions lasting only seconds or minutes.
With an UPS:
- Proxmox nodes continue running during short outages
- VMs and containers remain available to users and services
- No reboot or recovery is needed once power returns
This is particularly valuable for home labs and small servers where short power cuts are common.
6. Centralized Power Management with NUT
When combined with NUT, a single UPS can manage multiple Proxmox nodes over the network. This allows:
- One UPS to safely shut down an entire Proxmox cluster
- Consistent power policies across all nodes
- Monitoring of battery status and power events from one place
This setup is cost-effective and scalable, especially for clusters without redundant power feeds.
7. Reduced Recovery Time After Outages
Without an UPS, recovering from a power outage may involve filesystem checks, VM repairs, or even restoring from backups. With an UPS-managed shutdown:
- Proxmox boots cleanly
- VMs start normally
- Services return faster and with less manual intervention
This saves time and reduces operational stress after power incidents.