Used HPE ProLiant DL360/DL380 Buying Guide 2026

I sold a DL380 Gen10 to a client last year. Three weeks later, they called: the server had taken out their entire VM cluster. Turns out the seller hadn’t replaced the 8-year-old capacitors in the power supply. They paid me $3,200 for a paperweight. You don’t want that outcome.

Here’s the math: a single dead DL360 Gen10 can cost you $10k+ in lost productivity if it’s holding mission-critical apps. Replacement parts like a working DDR4 ECC RDIMM module run $250–$700 each. You’re not just buying metal—you’re buying a gamble. Let’s break down how to tip the odds in your favor.

---

Should You Pay $1,500–$4,000 for a Used DL360/DL380?

First, know the baseline. A Gen10 DL360 in decent condition (2017–2020 manufacture date) costs $1,500–$3,500. A Gen10 Plus with newer Intel Xeon Scalable CPUs runs $2,500–$4,000. But here’s the gotcha: 20% of these servers in the wild have silent disk failures. That’s not HPE’s fault—it’s physics. Hard drives in these boxes fail at 15–25% rates depending on age.

If the seller won’t let you RMA the drive array, walk. A single 1.2TB SAS drive replacement costs $120–$180, but labor to replace it? That’s your IT team’s time.

---

Power Supplies: Don’t Trust the LED

A “green” power supply doesn’t mean it’s healthy. Capacitors in PSUs dry out after 5–7 years. I’ve opened 15% of DL380 Gen10s and found bulging capacitors. Replacement PSUs? $300–$500 each.

Ask the seller for date codes on the PSUs. Look for stamps like “23W” (week 23 of 2023)—anything older than 2019 is a risk. If they can’t provide this, add $400–$600 to your budget for eventual replacement.

---

Memory: ECC Isn’t Foolproof

These servers use DDR4 ECC RDIMMs. Even with error correction, bad memory modules will crash your OS. I’ve seen 30% failure rates in unbranded modules.

Check the part number against HPE’s compatibility list. A working 754519-083 module is worth $200–$350. If the seller says “memory works,” ask for a MemTest86+ log. No log? Assume one module is dead.

---

CPU Sockets: Hidden Tombstones

Gen10 servers with Xeon E5-2600v4 CPUs often have “tombstoned” processors—dead chips that look fine. I’ve seen 10% of used DL360s with CPUs that won’t POST.

Demand a CPU stress test log (using tools like Prime95). If the seller refuses, factor in $800–$1,500 for a replacement Xeon. Also, check the CPU’s manufacture date. A 2016 chip in a 2020 server? That’s a reboxed lemon.

---

RAID Controllers: The Silent Killer

A dead RAID card can wipe your array. HPE’s Smart Array P408i is common in DL380s—but 25% of used units I’ve tested have firmware corruption.

Ask for the controller’s firmware version. It should match HPE’s latest update for the model. If it’s outdated, ask why. Also, verify the battery charge on the cache module—less than 1.8V means data corruption risks. Replacement cache modules cost $150–$250.

---

What to Do Next

1. Inspect physically: Look for corrosion under the chassis, especially near the power supply and drive bays.
2. Check serial numbers: Use HPE’s PartSurfer to verify part dates.
3. Request diagnostics: Ask for a full hpacucli report showing drive health, controller status, and battery levels.
4. Budget for spares: Buy one extra drive and PSU per 10 servers. You’ll need them.

---

**