Why Does a UPS Shutdown Signal Reach the Host but Not Its Virtual Machines?

Eva Wong

IceWhale author

Eva Wong is the Technical Writer and resident tinkerer at ZimaSpace. A lifelong geek with a passion for homelabs and open-source software, she specializes in translating complex technical concepts into accessible, hands-on guides. Eva believes that self-hosting should be fun, not intimidating. Through her tutorials, she empowers the community to demystify hardware setups, from building their first NAS to mastering Docker containers.

Why Does a UPS Shutdown Signal Reach the Host but Not Its Virtual Machines? - Zima Store Online

A UPS shutdown signal can reach the virtualization host without reaching its VMs because power detection and guest shutdown orchestration are separate steps.

First prove the host receives the UPS event, then test whether every important VM responds to a normal hypervisor shutdown. Inspect the low-battery action, guest-agent or ACPI behavior, stop ordering, and runtime margin. The safe end state is guests off first, host off last, with enough battery left for delays.

Prove Which Component Receives the UPS Event

Trigger a supported test event or inspect logs from a recent battery transition and record whether NUT, apcupsd, SNMP monitoring, or another daemon on the host sees it.

A practical Proxmox UPS shutdown flow separates UPS detection on a Proxmox host from the later actions required to stop workloads and power down the machine.

If the host never receives the event, stay at UPS communication. If it does, do not assume guest shutdown is automatic; inspect the next orchestration step.

Test Guest Shutdown Without the UPS

From the hypervisor, request an ordinary graceful shutdown of each critical VM and measure how long it takes. Check guest-agent state, ACPI behavior, and applications that delay shutdown.

A NUT with Proxmox deployment shows why NUT can be the event source while Proxmox still needs a defined response path for the virtualized workloads it hosts.

Fix guests that ignore a normal hypervisor shutdown before testing battery emergencies. UPS automation cannot make an unresponsive guest suddenly shut down cleanly.

Make the UPS Action Call the Virtualization Layer

Inspect the script or service that handles low battery. A direct host poweroff can bypass normal VM stop ordering if it does not invoke the hypervisor’s guest shutdown process.

Another Proxmox-oriented Proxmox NUT orchestration demonstrates the need to connect NUT state to deliberate system actions rather than treating the UPS daemon itself as a guest manager.

Use the platform’s supported shutdown orchestration or a narrow hook that requests guest stops first. Avoid custom kill commands that bypass application shutdown inside the VM.

Featured

ZimaCube 2 Personal Cloud Home NAS

ZimaCube2

Budget Time for Slow Guests and the Host

Measure total guest-stop time, host shutdown time, UPS runtime under load, and the low-battery threshold. Leave margin for one guest taking longer than usual.

Purpose-built Proxmox UPS shutdown service exists because orderly Proxmox shutdown needs policy and timing, not merely proof that an SNMP or USB signal reached the host.

Set bounded guest timeouts and an escalation order so a stuck noncritical VM cannot exhaust the battery. Critical storage or database guests may deserve earlier shutdown.

Run One Controlled End-to-End Power Test

During a maintenance window, simulate or trigger the supported UPS condition, watch guest shutdown order, confirm the host powers off last, then restore mains and verify clean starts.

Home-lab guidance on power-outage homelab resilience reinforces that resilience depends on testing the entire outage sequence rather than collecting UPS telemetry alone. The related ZimaSpace guide on VM backup readiness adds the restore and backup boundary.

The workflow is complete only when every required VM shuts down cleanly before the host, the battery retains safety margin, and the same sequence works again after configuration changes.

Frequently Asked Questions

Do VMs need their own NUT client when the host already has one?

Not always. The host can orchestrate guest shutdown through the hypervisor, but that guest-stop step must be configured, tested, and allowed enough time before host power-off.

What happens if a VM refuses to shut down?

Use a bounded timeout and escalation policy. One stuck guest should not consume the entire UPS battery until the host loses power abruptly.

How should I test UPS shutdown safely?

Use a maintenance window, healthy backups, a controlled low-battery or simulated event where supported, and verify guest stop order, host shutdown, and clean recovery after power returns.