Red-pill me on RAID

Red-pill me on RAID.
Do we still use it, does anyone care?

>Red-pill me

its still used for servers for data redundancy, for home systems just get an ssd

When you use RAID:
When you have many disks, and need to keep your filesystem online 24/7. The important parts here are throughput, redundancy, and online recovery. You buy very expensive cards to do this, when you can't afford to have your systems offline whilst you restore from a backup due to a hardware failure.

When you don't use RAID:
Every other time.

Note that redundancy is not a backup. Backups are separate, typically offline, and scheduled. 1:1. Not parity bits.

I'm running a Proxmox server with 4tb (2tb) in raid 1 for backups, web server, media server ect. Should I be using a different raid or am I good?

Redpill is a legit thing to say despite/pol/niggers constantly using it.

No it isn't.

>despite/pol/niggers constantly using it

Why lower yourself to that level then?

my laptop has a pretty nice raid setup

>Do we still use it, does anyone care?
of course we still use it. disks still die.

With 2 disks a mirror setup is basically all you can get, higher raid levels require at least 3 disks. It's probably enough for a regular person.

It is a staple for my job. Currently trying to convince a co-worker to move to RAID 10 instead of RAID5 plus spare

>backups
>>backups
>>>backups
>>backups
>backups

it is.
>against hdd failure

no. it is redundancy.

What if a hard drive fails in a way that writes corrupted bits to crucial parts of the filesystem?

If your backup is connected to your primary, and is instantaneous, it is not a backup.

>hard drive fails in a way that writes corrupted bits to crucial parts of the filesystem
by itself? nothing. should not affect the other discs. if the OS fucks it up its not a hdd failure

what about a ransomware?

not a hdd failure
do you not understand what a hdd failure is or did you not understand that I only said its a backup against hdd failure?

>by itself? nothing. should not affect the other discs.

RAID has no built in health checking. If your drive shits itself in a way that is not immediately obvious, you will end up with a corrupted filesystem.

There are work arounds for this, like ZFS which has filesystem level health checking, but this completely invalidates RAID anyway with RAIDZ. And even still, don't use RAIDZ. Use zfs + an offline backup.

Storage is cheaper than ever before and there is no excuse for not having an offline backup of your critical data. RAID is not that, and it never will be.

>the failure mode matters when you've lost all your data anyway

>you will end up with a corrupted filesystem.
yes. on that single disc. we are still talking about raid1 right?
should not affect the other discs and those should still have a correct filesystem, right?

k

Parity parity parity. Not a backup.

RAID is never a backup jesus.


Make real backups, only use RAID if you need redundancy.

Only if the error is immediately detected. Which RAID does nothing to support, which basically means you have a case where the drive completely fails to power on (in this case the power circuitry is probably fucked and the actual data on the disk is fine. either way dead disk, but no corruption), or you need some mechanism at the filesystem level to detect when something isn't healthy. Otherwise errors will just be propagated between the drives.

>mfw you lose all your data

>mfw i dont even have raid because yolo
>mfw jokes on you

>INTERNET EXPLORER

RAID is used to keep a system online in case of drive failure. You shouldn't rely on it to protect your data in case of file corruption caused by drive failure or software malicious or otherwise. Offline backups are basically the only thing that can really protect your data. If you can make at least two backups and store one or more of them offsite that's probably the best solution.

Allright, assuming that I know right and its not the hdd itself that decides which sector to write in, I can imagine that the corrupted drive flags a sector for write when in reality its already used on the correct disc thus corrupting it.
Still doubt it that this is anywhere common and to be calculated with, complete failure sure is more common thus raid 1 is good against it.

Servers and workstations that need more storage than ssd can deliver. If you're working with lots of video for example and need like 10TB+ kind of storage, you don't really have a choice.

I've got a pair of 2TB hdds in RAID 0 for my steam library because big ssds didn't exist back when I built this and my 512gb ssd fills up way too quick. For like $150 I get 3.6TB at about half the sequential speed of an sata ssd.

...

I think realistically the main concern is with software. Keep an offline backup of your stuff. You should really keep multiple offline backups but even just one will reduce the damage caused by software.

what kind of offline backup should I be using?
1:1 so the hdd is replaceable fast but if I get a ransomware and dont realise it soon enough the backup might copy all the crap too at night?
or some daily change only type in which case the backup needs much more space?

If you expect your data collection to continue growing 1:1 might not be enough but you can always buy another drive or drives. Make regular frequent backups but I'd also recommend storing very important (maybe personal?) information completely separately from the other backup. Stuff that you definitely cannot lose but wont be accessed frequently anyway.