one thing to consider with a raid is the drive resonance. if you don’t have proper hardware to ensure the drives are properly isolated you could shave months or years off their lifespan.
always use the same make/model and data density when building a raid. you should also buy them either all new or used that were pulled from another raid.
also more is better. raids are designed to fail. buy extra drives to replace active drives. my personal rule is to buy 50-75% more than required. so if the raid will have four drives, I buy six or seven.
also, when one drive fails others will inevitably follow due to manufacturing defects or resonance changes. replace your backup stock when affordable.
I do agree though. of you’re selfhosting, a raid array is the best place to start. though, I’m personally partial to a raid-10 over a 5.
ShredderFeeder@shredderfood.net
cmnybo@discuss.tchncs.de 5 days ago
RAID isn’t an alternative to a backup. If a second drive fails while rebuilding the RAID-5 array, all of the data will be lost.
Dudewitbow@lemmy.zip 5 days ago
its usually for redudancy but prople tend to use the wrong term for it.
unless the data is off site, its not truely a backup
Two is one and one is none. No matter if it’s offsite or on top of your main server. Also 3-2-1 is an industry standard for a reason. Plus unless you test that you can actually restore your backups they don’t really exist (also known as Schrödinger Backups).
nexttech@lemmy.world 5 days ago
people* 🤓 👆
RAID is for continuous of operations, backup is for restoration after a failure.
I have both, but run Raid-6 on my array due to the size of the drives. As has been observed here, if one drive fails and then another drive fails during the rebuild, you’ve lost everything.
Then to add to that I’ve got backups of everything I can’t replace. Plex library will come back automatically (though it will take a while). But my documents, digital filing cabinet, pictures, videos of my kids, those are all protected by raid, backup, AND off-site replication.