If you manage a data center, a server room, or even just a homelab, you already know the soundtrack: the persistent, jet-engine roar of server cooling fans. While a certain amount of noise is expected, it can become a major headache when server fans ramp up to 100% and refuse to spin down—especially when the server isn't even running hot.
If you are running HPE ProLiant servers (like the MicroServer Gen11) and have ever swapped in third-party, non-HPE hard drives, you might be very familiar with this exact scenario.
Recently, an insightful discussion over on the HPE Community Forums shed light on why this happens and, more importantly, how one clever IT professional managed to solve it.
⚠️ A Quick Word Before We Dive In
Before we get to the technical details, we want to emphasize that we do not endorse the use of non-HPE drives in HPE servers, and we do not sell non-HPE drives for use in HPE servers. We strongly encourage people running HPE servers to always use authentic HPE parts to ensure optimal compatibility, reliable warranty support, and correct thermal performance.
The History of the Issue
The "100% fan speed" quirk isn't a new phenomenon. In fact, IT professionals have been reporting this behavior across various ProLiant generations going back as far as 2011.
Typically, HPE's Integrated Lights-Out (iLO) management engine relies on thermal sensors to dictate how fast the chassis fans should spin. If the sensors report a high temperature, iLO spins the fans up to protect the expensive hardware. However, when standard third-party hard drives are installed, many users have noticed the fans screaming at maximum RPM, even when the drives are practically idling and cool to the touch.
For years, this left many system administrators frustrated, wasting power and enduring unnecessary noise without a clear understanding of why it was happening.
The Root Cause
A community member named pfavr decided to dig into the software stack to find out exactly why this was happening on their MicroServer Gen11 equipped with third-party 24TB drives.
Here is what they discovered: The Agentless Management Service (AMS)—specifically the cpqide agent—reports drive temperatures to iLO. However, when it detects a non-HPE drive, it artificially inflates the reported temperature by approximately +25 °C.
iLO then reads this inflated temperature data, genuinely believes the drives are overheating, and triggers the fans to spin at maximum speed as a programmed safety response.
The Open-Source Solution
Because the root cause lies in the telemetry handoff between AMS and iLO, pfavr developed an elegant, open-source workaround to intercept and correct the data.
Instead of dealing with complex and risky binary patching, they wrote a script that corrects the temperature readings before iLO sees them. By feeding iLO the true, uninflated temperature of the drives, the system realizes the hardware is safe and allows the fans to idle normally (dropping back down to a whisper-quiet 20% in the author's case).
The fix is designed to be clean, easily reversible, and simple to deploy for system administrators looking to reclaim their quiet server rooms and reduce wasted energy.
Get the Fix on GitHubWrapping Up
While sticking to vendor-approved hardware is always the safest bet for enterprise support and guaranteed compatibility, the reality is that many IT pros and homelab enthusiasts experiment with third-party components. It's always fascinating to see the IT community come together to reverse-engineer quirks, write custom patches, and share their knowledge with the rest of the world.
If you've been battling the jet-engine roar of a ProLiant server and you're comfortable tinkering under the hood, this clever workaround might just be the peace and quiet you’ve been looking for!
