Raid-Z1 or Raid-Z2

[Chaoticlusts]

Hey all

I'm right at the final point of rebuilding my NAS..have backed up all the data onto independent HD's and am about to reinstall FreeNAS and add the new drives...my question is do I go for Raid-Z1 or Z2

I had 4 1.5TB drives before so the answer was easy Z2 would be overkill and remove a ton of storage....now I'm looking at 8 drives(still 1.5's)...how risky would Z1 be? It's not an absolute tragedy if I lost everything (though it would be a complete pain in the arse nothing on there will be irreplaceable or at least anything that is would be stored on a separate location as well)

the trade-off is this is the most HD's I can put in this system so the next time I run out of space I'll need to move to larger discs (presumably 3TB's or larger by then) so I really want this to last me as long as possible (the 4.5 I have right now is full)...but similarly it's useless if the whole thing has a high chance of failing...given my experience with HD's I've *never* had 2 fail in a short period of time let alone simultaneously so Z1 seems enough but I'm not an IT expert simply a longterm high usage layman so there maybe scenario's I'm not taking into account that could kill a single parity system....I'm pretty confident 10.5TB (marketing) would last me long enough 9 I'm a little dubious about but would probably be ok basically I'd have more freedom not to worry about it with 10.5 so if Z2 is overkill I'd prefer to avoid it but if Z1 is being risky I'm willing to take the space hit if this makes sense

(I'm sorry if this has come up before but I can't get the search to allow me to search for 'raid-z2' just searches for raid >_< I really am trying to research this myself too...but people on here know a hell of a lot more than I do )

If anymore information is needed that I left out please let me know and I'll add it

 

[Chaoticlusts]

ok after a couple more hours of research it seems over 5 drives it's quite recommended you move to double parity (hell Sun recommends you move to triple at 8 drives but ah heck that ) so looks like I'm going that way

just thought I'd post that in here if anyone looks at this with the same question

not particularly thrilled about reducing my capacity another 1.5 but well...would be worse to lose the whole array

 

[Emulex]

seems like RAID-Z3 is the best according to article. but man in my life i've seen far too many dual disk failures that i'm a firm believer in raid-10 given drive prices. plus there is a level of performance that raid-z3 seems to not deliver on random writes

 

[sub.mesa]

For 4K sector drives you should adhere to these recommended disk sizes:

RAID-Z: 3 or 5 disks
RAID-Z2: 6 or 10 disks

Making groups of x disks means you have multiple vdevs, like:

vdev1: RAID-Z2 of 10 disks
vdev2: RAID-Z2 of 10 disks

A mirror is not that secure; it would not survive one failure + uBER; RAID-Z2 would.

Regarding random IOps; the more vdevs the higher your IOps performance. But generally, you can get alot more speed by opting for an SSD to function as SLOG and L2ARC. This will accelerate both random read and random write, and let's the HDD focus on sequential I/O so that means massive speedups.

The problem is: you need an SSD with supercapacitor to use it as SLOG disk safely. That means one of the following:
- Intel G3
- Marvell C400
- Sandforce SF2000 enterprise family (like Vertex 3 Pro)

Each of these should have a supercapacitor, making the SLOG feature relatively safe to use. Without supercapacitor, a corrupted SLOG could kill your entire pool. SSDs corrupt easily on unclean shutdowns; be careful with powering down. SSDs can die instantly if they are powered down on the wrong moment.

 

[nanaki333]

seems like RAID-Z3 is the best according to article. but man in my life i've seen far too many dual disk failures that i'm a firm believer in raid-10 given drive prices. plus there is a level of performance that raid-z3 seems to not deliver on random writes

what happens when both disks die on the same side in raid 10?

 

[Emulex]

same thing that happens when raid-z2 has three disk simultaneous failure.

 

[sub.mesa]

But you can see the difference: RAID-Z2 guarantees you two full failures; a mirror only guarantees one with the potential of more but that is not guaranteed. A mirror is much simpler to implement and avoids things like the raid5 write hole and other things relevant to parity RAID. However, RAID-Z has no write hole due to there being no read-modify-write cycles in RAID-Z; all writes are full stripe block writes.

 

[nanaki333]

same thing that happens when raid-z2 has three disk simultaneous failure.

i know. i was sorta joking. raid10's redundancy is nice if you have 1 failure on each side, while other raid setups like z2 can lose any of the 3. i'm not a huge fan of raid10. i personally find it more of a gamble, but that's just me.