Overstepping the aggregat limit

[igor]

Hiya!

Just a short question regarding RAID group aggregation. I've a 2020A running on DOT 7.3.4 and I'd like to attach two 14x1TB shelves (11 data, 2 parity, 1 spare) to one of it's controllers... As I remember there's a 16TB aggregate limit here, and each shelf has 8591973352 kbytes net capacity (just over 8TB) making both shelves have more just a drop over the limit.

If I go ahead with this, will the system allow it? If it does, can I run into some problems later on perhaps?

[aborzenkov]

You can create aggregate with 19 1TB data disks which means 23 disks in total (2 RAID_DP RGs). I would not try to oversubscribe it, really … but it is interesting to know what happens ☺

[Darkstar]

Simple: you can't overstep it. It's a hard limit that has to do with 32 bit block access. 0xffffffff * 4096 bytes/block = 16TB.

If you try to add more disks, the system will give you an error message.

In your case I would create 2 aggregates with ~8.5tb each, however, I would stick with one spare disk. You don't really need one spare per shelf

-Michael

[igor]

Thanks for your input Michael,

I was hoping to avoid creating 2 aggregates but it can't be done. 🙂

As for that 1 spare, I dunno... Regardless of NetApp's optimistic predictions, I've learned that it does take a long to time to rebuild a fully loaded 1 TB SATA disk. And since double failure is not a myth, and usually has to resolved while handling normal daily workload, I personally wouldn't feel comfortable waiting for ?? hours for it to finish - hoping that something else might not go wrong during that time with the same RAID group. That's a LOT of data to be lost.

Pozdrav,

Igor