Setup Guide

UPS Installation and Shutdown Software

Updated 5 min read

Installing a UPS is straightforward. What separates a working installation from an expensive space heater is the shutdown configuration — and it is the step most often left for later and never done. A UPS with no automatic shutdown does not prevent a hard power loss on an unattended system. It only delays it.

Before you start

  • Confirm the input circuit and plug type — larger units often need dedicated electrical work
  • Check the rack's weight rating; UPS batteries are heavy
  • Have the network details ready if fitting a management card
  • Know how long a full graceful shutdown actually takes on your slowest system
  • Plan the outage window — connecting the load means powering things down
Electrical work is an electrician's job

Higher-capacity UPS units frequently require a dedicated circuit at a higher voltage, with a specific receptacle. That is licensed electrical work in most jurisdictions and not something to improvise with an adapter. Confirm the requirement from the datasheet before the unit arrives — waiting on an electrician is the most common reason a UPS sits boxed for a fortnight.

Physical installation

  1. Get help with the lift Rack UPS units are heavy, and the weight is concentrated in the batteries. Many models let you remove the battery modules first, which makes the chassis manageable. Do that rather than lifting the whole unit.
  2. Mount it at the bottom of the rack Weight low keeps the rack stable. Check the rack's rated load, especially in older or half-populated cabinets.
  3. Install battery modules after mounting Insert them once the chassis is secured, following the marked orientation.
  4. Leave clearance for airflow UPS units generate heat, and online double-conversion units generate it continuously because they are always converting. Follow the clearance specified and do not box them in.
  5. Connect input power and let it charge Give it the manufacturer's specified initial charge time before relying on it. A UPS is not protecting anything until the batteries are charged.
  6. Connect the load, distinguishing battery from surge-only outlets Anything that must survive an outage goes on battery outlets. Laser printers and non-critical loads go on surge-only or elsewhere entirely.
  7. Label every outlet and cable Which device, and whether it is battery-backed. This takes five minutes and saves an overload six months later.
  8. Record the install date on the unit Batteries are consumables with a service life measured from installation. Writing the date on the chassis is the simplest possible asset record.
Do not daisy-chain UPS units or power strips

Plugging one UPS into another causes both to misread the waveform and behave unpredictably. Plugging a power strip into a UPS outlet makes it trivially easy to exceed the rating without anyone noticing. If you need more outlets, use a rack PDU rated for the job and count the load properly.

Fit and configure the management card

A network management card gives the UPS its own network presence, so it can be monitored, alert on faults, and trigger shutdowns without depending on a connected host.

  1. Install the card in the accessory slot and connect it to the network. Many are hot-swappable, but check for your model.
  2. Give it a reserved IP address. Same reasoning as any infrastructure device — a monitoring interface that moves is a monitoring interface nobody can find.
  3. Change the default credentials immediately. A UPS management card can shut down your servers. Treat it as the privileged device it is.
  4. Update the card firmware. These have had real security issues; keep them current and on a management network where possible.
  5. Configure email or SNMP alerts for on-battery, low-battery, overload, and battery-fault events. The battery-fault alert is the one that tells you about a dead battery before an outage does.
  6. Set the date and time, so the event log is usable when you need to reconstruct what happened.

Configure automatic shutdown

This is the step that turns runtime into protection. Without it, the UPS carries the load until the battery is exhausted and then everything loses power hard — the exact outcome you bought it to prevent.

Method Suits
Agent software on each host A small number of physical machines. Each runs the vendor agent, talks to the UPS, and shuts itself down on the configured trigger.
Network management card to multiple hosts Several machines sharing one UPS. The card notifies each host over the network. No USB tether required.
Virtualization integration Virtual hosts. Shuts guests down in order, then the host. Sequencing matters here — a storage array powered off before the guests using it causes exactly the corruption you were avoiding.
NAS and appliance built-in support Many storage appliances have native UPS support over USB or network. Check before buying third-party software.
  1. Decide the trigger Usually a combination: on battery for longer than a set period, or battery below a set percentage. Time-based triggers are easier to reason about; capacity-based ones adapt as batteries age.
  2. Set the shutdown order deliberately Guests before hosts, application servers before databases, and anything depending on shared storage before the storage itself.
  3. Allow more time than the shutdown takes If a clean shutdown takes eight minutes, do not trigger it with eight minutes of battery left. Leave real margin.
  4. Decide the restart behaviour Should equipment power back on automatically when utility power returns? Usually yes for infrastructure, and it needs the servers' own BIOS power-on settings to agree.
  5. Test it properly Pull the input power during a planned window and watch the whole sequence run. This is the only way to know it works.
An untested shutdown configuration is an assumption

Shutdown sequences fail for ordinary reasons — an agent that stopped after an update, a credential that expired, a host added later and never configured. The only way to know is to test with a real power interruption during a planned window, and to retest after any significant change. A configuration that has never been exercised should not be counted as protection.

Before you call it finished

Check Why
Run a self-test from the panel or card Confirms the batteries take the load at all.
Pull input power in a planned window The real test. Watch the load transfer and the shutdown sequence run.
Confirm every host shut down cleanly Check logs, not just that the lights went out.
Confirm equipment restarts when power returns Including BIOS power-on settings on each server.
Verify alerts actually arrive Send a test alert to a monitored mailbox, not to someone's personal inbox.
Confirm the load is under the watt rating Read it from the panel with everything running.
Record install date, model, and battery part number Batteries are consumables and you will need this in three years.

Official manufacturer resources

Shutdown agents, management card firmware, and virtualization integrations are product-specific. Confirm supported versions against your operating systems and hypervisor.

Planning an install?

Tell us the model, what it will feed, and what has to shut down cleanly. We will flag the input circuit requirement early and help you get the shutdown sequence right the first time.

Still stuck? Talk to someone who works on this hardware daily.

Tell us the model, what you have already tried, and what you are seeing. We will tell you whether it is a setting, a consumable, or a part — and we will say so if you do not need to buy anything.