When "Pro" Means "Please Replace Often"

A supposedly “Pro” Samsung 990 NVMe turned my home server into a biweekly reboot ritual. Meanwhile, a forgotten SK Hynix laptop SSD ran flawlessly for five years. Here’s how a boring Goodram and a FireCuda 530R restored my faith in hardware.
When "Pro" Means "Please Replace Often"
Samsung 990 pro

(or reboot, whichever happens first.)

I’ve been running a few small servers at home for years. Nothing fancy - just a few Docker containers here and there. You know, the usual "it’s not production but it kinda is" setup.

For the longest time, my main server ran on a completely generic OEM SK Hynix NVMe. I pulled straight out of a laptop. I don't remember when and from where, probably HP Envy laptop. No fancy label on SSD, no heatsinks, no "gaming-grade" marketing. I'm not sure if it even has a promo page. Just a quiet, boring piece of silicon that was actually working on for nearly five years of 24/7 uptime without a single hiccup.

Then, one day, I thought: hey, maybe it’s time to upgrade. I mean, I did rebuilt my server - finally threw away the ancient RAID card in favor of ZFS, rearranged some drives. So I though if I have to reinstall everything anyway - might as well to upgrade the SSD and dedicate the only NVMe slot to docker containers apps data.

BEHOLD! The Samsung 990 Pro - the flagship of consumer NVMe drives.
PCIe 4.0, absurd speeds, "pro" in name. The whole package. Should be good.

And for about two weeks, it was great.

Then "the greatness" became read-only. And then - complete controller self-defense panic.

I’m not even exaggerating. Every couple of weeks, the controller would just give up on life and flip the disk into "nah, you can only read now" mode. Docker apps froze, logs screamed, and my faith in Samsung's QA department was gone.

Checked SMART values. Tried mounting it back somehow. Tried updating firmware - I was running the latest.

Nothing helped, except the full power-off and reboot. The moment it had to actually work - write logs, handle containers, do something other than idle benchmarks - it would die again. Roughly every two weeks.

So, naturally, I did what any responsible person would: shredded partitions and sent it back.

Whether it still boots now, I honestly don’t know or care.

In its place now beats the heart of a true legend: Seagate FireCuda 530R. I've been running the Seagate FireCuda 530 (without "R") on another PC for a few years now -
no crashes, no firmware crises, even handles absolutely degenerate-level swap abuse. No "pro" label trying to compensate for something. I mean, yeah, it has heavy gaming marketing, but specs are absurdly good.

The idea is - it's the kind of drive that actually deserves to be left running for months without supervision.

And for the linux OS drive? It’s currently spinning happily on Goodram something-something SATA SSD.

It's SATA. It's boring. It's blue... I think. I'm not even sure how it's called... No, I'm not roasting it - it works. It works flawlessly. And I value that.

Sometimes, boring is the ultimate upgrade.

Subscribe to Technikatsu newsletter and stay updated.

Don't miss anything. Get all the latest posts delivered straight to your inbox. It's free!
Great! Check your inbox and click the link to confirm your subscription.
Error! Please enter a valid email address!