Is CrystalDiskInfo a reliable way to tell when a storage device might fail?

PCMasterRays · March 6

Hi LTT forums

I have some concerns over old drives, mainly HDDs. They're a mix of "I found it somewhere" and "I found it somewhere" drives. AKA they're all used.

CrystalDiskInfo says OK on all the drives, but my experience personally with storage devices has always been positive, never had a drive fail in over a decade of being a PC hardware enthusiast, even despite my exclusively second hand drives (I can hear Linus stirring in his sleep from the UK)
I have dates on some of them, one being 2013, the rest are at least that old if not older, for the HDDs anyways.
The SSDs I have are newer, maybe 2015ish, but still, my only personal experience being positive, and the experience I have fixing others PCs only being it's completely f'd, I don't really know what I'm looking for..

I'm aware there's many factors that go into a drive failing, and also out of best practice I really should start replacing the drives, but in my case, I have the important stuff in the cloud, and I do plan to replace the entire rig at some arbitrary point in the future, if a pile of cash falls out the sky to fix my motorcycle, car, PC, relationship, and home server, but for now I would just like to know what I'm looking out for.

Is there something in CrystalDiskInfo, or any software, that might indicate a drive is starting to fail, or might fail in the future, or is it a case of waking up to "womp womp, drive dead"

191x7 · March 6

S.M.A.R.T. is pretty much the only thing we have that comes close to that. But I would not use a single tool, I always check with more, going from the favorite to the least:

Hard Disk Sentinel

Parted Magic (live Linux)

HDTune

CrystalDiskInfo

Seagate SeaTools

WD drive tools

porina · March 6

SMART is meant to report on the current state of the disk, and for some parameters maybe include historic worst. If you get a SMART warning, take note. However there are cases where there can be warning signs that don't always trigger a warning. The biggest two for me are "pending sectors" and "reallocated sectors". These could indicate the surface is going bad.

Before using a disk for important data, I'd take the routine from Unraid.

1, do a full surface read - this can identify any existing unstable sectors and give the drive a chance to repair them on next write

2, do a full surface write - this will freshly write over the surface and can show up previously unknown unstable sectors

3, do a full surface read - this will find any newly identified unstable sectors

Check pending sectors and reallocated sectors both remain at zero, then it is probably ok. Some errors can be self repairable which is why drives don't report them as a major warning, but if they're frequently occurring or getting worse, it is a sign the drive is on the way out.

As always, the safest thing to do is assume any disk can fail any time and take appropriate backups of data you care about.