This post documents the stress of having to replace all my HDDs on UNRAID after they began failing one after the other… all within a matter of days. They were 3.5 years old, all bought together, and used 24/7 — so, yeah, they retired in sync.
It Started With One
I bought a new HDD, thinking I’d expand my array. But before I could even do that, one of my drives started failing. It was lucky timing—that new disk was about to arrive and went straight in as a replacement.
But then, another drive failed. Then another. In a few days, all my drives were showing SMART issues. That slow burn quickly became a wildfire.
UNRAID didn’t warn me about SMART issues
I expected UNRAID to warn me about the issues, but it didn’t. I only caught it after a forced reboot failed, and my BIOS showed SMART errors on one of the drives. That’s when I jumped into the web UI and manually checked SMART reports.
Failing drives showed:
Reallocated sector countmarked FAILING NOWReported uncorrected errorsin the thousands- Multiple failed
Short SMART Tests
I had to go into the web UI, check each disk manually, and see what was going wrong.
How to manually run a SMART test using Unraid UI
- Go to the Main tab
- Click on a disk in the array
- Open the Self-test tab
- Select Short or Extended and click Start
- Review the results under the same tab after the test completes
Replacing the Drives
Before swapping anything, I had to identify which disk was failing.
- Go to the Main tab in UNRAID
- Click on the suspected disk
- Open the “Identity” or “Device Info” tab
- Check the Serial Number
Once I matched the serial to the physical drive, I went ahead and replaced it.
Steps I followed:
- Shut down the server
- Physically replace the bad HDD
- Booted the server back up
- Assigned the new disk in the same slot
- Let UNRAID rebuild it
Repeat for each failing disk.
Eventually, all three HDDs were replaced.
🧠 Lessons Learned
Here’s what I wish I had done differently and what I’d recommend:
Schedule SMART Tests
Unraid doesn’t run these on its own.
- Find a way to schedule SMART tests and notify errors.
If you buy drivers at once, expect them to fail together
Stagger replacements every year to avoid a full-blown panic 3.5 years later.
It’s hard because we usually build our servers at once, so if you have the option, reuse existing HDDs that you may have around to be able to stagger the buying.
I’m in the same situation again; ALL HDDs are bought close together.
Making HDDs last longer
To make your HDDs last longer, you can increase your cache and allow UNRAID to spin down the drivers when not in use
- Write about my Spin down setup