"Another one bites the dust.." BEEP BEEP BEEP!
😬 "Shit, make that two."

Down from 5TB redundant to 3TB non-redundant and still waiting for my new Reds. No choice but to shut everything down and wait. 😩

RAID is not a backup.

It's supposed to be highly available high speed storage that offers only the best compensation from higher risk of hardware failure when more pieces are involved.
A backup however is always offline, or otherwise disconnected from the high availability storage.

@FlameHazeLinux yeaaaa op please elaborate. On a RAID 6 you can have 5 TB and two drive failures if you have 7 1TB drives but I don’t know if a scenario where you lose storage if part of your RAID fails. If you lose enough of your array you lose storage you’ve lost data.

Do t use green's for raids

No green wds for hardware raid. Seriously they're great sitting idle to serve something occasionally, but the energy saving is not worth it.
Black or blue. Mix brands, mix batches
(feel vindicated for seeing someone else crossing bad hdds with a permanent marker)

Oh OP. So much sadness here.

RAID is not a backup.
Don’t use Greens for RAID
You must have had Data loss if you had storage space decrease

Also, monitor closely what the raid controller is saying, in my experience, it's way more valuable than arbitrary values on smart reports. Even if you're not using the raid capabilities of the hardware controller, like lvm, md, btrfs or zfs. Query it daily, parse the log and have your spares ready.

@nbamaral no. Absolutely don't mix brands. Don't even mix models. But make sure they come from different locations

@Kimmax
Yes, there aren't many brands left anyway.
I don't on servers, but try my best to manage the hdd purchases to avoid joining from the same batch.
At home I mix and match a lot more, but I know what you're saying.

I'm using a mix of Btrfs and Linux software RAID. I built this NAS before I knew what I was doing (and before I could afford Reds) and never got around to redoing it.
Everything on it is either a backup of other devices or not very important (downloads, DVD images, etc.).
I lost some downloads which weren't redundant and some old backups of my VPSs when both mirrored drives failed about a week apart.
Each time a drive in a RAID died, I shut the pair down until I could replace it so I didn't lose anything important.
I've ordered some Reds and will be redoing everything with Btrfs. I'll probably have to live off ramen for the next month, but at least I won't lose my photos.

@AndSoWeCode you're probably talking about RAID 0. RAID 1 is for rendundancy

@zankar no. RAID no matter what, is NOT A BACKUP. Redundancy saves only from very specific, localized hardware failures. Those are much rarer than software or user failures.