Your phone buzzes at 6:47 AM. It’s an alert from the monitoring system: server room temperature 28°C and climbing. The setpoint is 22°C. You pull up the BMS remotely — both AC units showing offline. By the time you drive to site, it’ll be 35°C in there. Your servers are heading for thermal shutdown, and every minute counts.
Server room AC failure is a genuine emergency. Unlike a warm office — which is uncomfortable but manageable — a hot server room causes hardware damage, data loss, and business-critical downtime. Servers start throttling performance at 35°C and begin shutting down to protect themselves around 40°C. This post covers what to do right now, what probably caused the failure, and how to prevent it happening again.
Our engineers see this fault regularly across Canary Wharf. A emergency breakdown service contract catches this kind of issue early, well before it escalates into a costly PPM.
Immediate Steps — Do These Now
Check the electrical supply first. Before assuming both AC units have mechanically failed simultaneously — which is unlikely — check the distribution board serving the server room cooling. A tripped breaker is the single most common reason for both units going down at once. If a breaker has tripped, reset it once. If it trips again immediately, you have an electrical fault — stop resetting and call an electrician.
Open doors for temporary airflow. If the server room has no cooling and the temperature is climbing, open the doors to allow air circulation from the surrounding space. It’s not proper cooling, but it slows the temperature rise. Point portable fans into the room if you have them. You’re buying time, not solving the problem.
Shut down non-critical servers. Every running server generates heat. If you can identify non-essential systems — development environments, test servers, backup jobs, staging platforms — shut them down immediately. Reducing the heat load inside the room directly slows the temperature climb. Focus on keeping production systems alive.
Check the BMS or unit controllers for a reset. Some AC systems lock out after a fault and need a manual reset to restart. Check the unit controllers on each indoor unit or the BMS interface. A power cycle on the controller — switching it off at the isolator, waiting 30 seconds, switching it back on — can sometimes clear a transient fault and get the unit restarting.
Call your AC contractor now. Not after lunch. Not after the morning meeting. If you have a maintenance contract with an emergency response SLA, activate it immediately. Server room temperatures above 35°C cause real damage. You need an engineer on site within hours.
Why Both Units Failed Simultaneously
Most server rooms have redundant cooling — two or more units where any single unit can handle the full heat load alone. So why are both down? The answer is almost always one of these.
Power transient tripping both units. A voltage spike, a supply dip, or a momentary power cut trips both units off their electrical supply simultaneously. Both units share the same distribution board, and when the supply glitches, everything goes. The fix might be as simple as resetting a breaker, but the root cause — a faulty contactor, a compressor drawing excessive starting current, or an upstream supply issue — needs investigating to prevent recurrence.
Condensate pump blocked causing safety shutoff. Precision AC units generate condensate, and most server room installations use a condensate pump to remove it. These pumps have a float switch that shuts down the AC unit if the condensate tray fills up — preventing water damage to the server room floor. If both units have condensate pumps of the same age, they tend to block within weeks of each other. We see this pattern constantly. Same pumps, same environment, same lifespan, same failure window.
Dirty filters on both units. If nobody has changed the filters in six months, both units are labouring with restricted airflow. They run hotter, draw more current, and eventually trip on high pressure or thermal overload. Two neglected units deteriorate at the same rate and fail around the same time. It’s not coincidence — it’s predictable neglect.
One compressor fails, the second unit overloads. If one unit’s compressor dies, the remaining unit has to handle 100% of the cooling load. During peak hours — or if the surviving unit is already in marginal condition — it runs flat out, overheats, and trips on overload protection. Now you have two down units: one with a dead compressor, one with a tripped overload. What started as a single failure cascades into total cooling loss.
The Redundancy You Need
Industry standard for server room cooling is N+1 redundancy. If you need one unit’s worth of cooling capacity to handle the heat load, you install two. Each unit should independently handle the full load. The second unit isn’t backup — it’s your insurance policy for exactly this scenario.
This problem often goes hand in hand with an AC system not cooling the building. If you are noticing both symptoms, it is worth getting an engineer out before the system fails completely. We have also covered Daikin U4 communication faults in a separate guide.
If your server room has two units but both need to run simultaneously to maintain temperature, you don’t have redundancy. You have two units splitting a load that’s too big for either one alone. The next single failure becomes a total cooling loss. We audit server room cooling setups regularly and can tell you within an hour whether your redundancy is genuine or on paper only.
Preventing the Next Emergency
Quarterly filter changes. Annual condensate pump cleaning and testing. Compressor current measurement to catch deterioration early. Refrigerant pressure checks. These are the basic PPM tasks that prevent the faults described above. The things that kill server room AC are almost always preventable with routine maintenance that costs a fraction of an emergency call-out.
Environmental monitoring with instant alerts is non-negotiable. If your server room doesn’t have temperature sensors with email or SMS alerts, fit them today. A sensor that costs less than a round of drinks gives you the early warning that turns a disaster into an inconvenience.
ADK’s Server Room Emergency Response
We offer a 2-hour emergency SLA for server room clients across London. Our engineers carry common spares — contactors, capacitors, condensate pumps, fan motors, filters — and can diagnose and repair most faults on the first visit. For compressor failures requiring parts on order, we deploy temporary portable cooling units to bridge the gap and keep your servers running.
If your server room AC has failed and temperatures are rising, call ADK now on 0207 801 0808. This is what we do.