WAFL and RAID overhead is not 5% - it's 10% off of the top (and it's specifically WAFL overhead - the RAID overhead comes in the parity disks).

The 'extra space 20%' for snapshots with LUNs will vary based on the application and the snapshot retention. It may be 0% in some cases, or as much as 100% in others (this is the 'it depends' part).

That is not bad if you consider that every disk subsystem has RAID overhead (NetApp is no more than Hitachi or anyone else) and that snapshots are backups - if you didn't need snapshots, there would be only the 10% overhead.

You can put SATA in place (the 3000 series supports it natively, even on boot volumes), but there are serious performance risks with doing so: SATA can do about 50% of the IOPS of an FCAL drive for the same amount of latency. What this means is, if you need to get 150 IOPS per disk, SATA will have double the latency of FCAL. That may not impact your application, or it may - this is where the analysis needs to focus (bad for Exchange or any other OLTP database, but probably just fine for data warehousing).

SATA does have higher failure rates than FCAL - this will probably change over time, but expect SAS to be the 'new hotness' - FCAL goodness, with SATA ease - and 2.5" drives are going to be the new thing as well.. time will tell if NetApp will go this direction or not.

Hope this all helps,

Glenn

-----Original Message----- From: owner-toasters@mathworks.com [mailto:owner-toasters@mathworks.com] On Behalf Of margesimpson@hushmail.com Sent: Thursday, July 27, 2006 10:00 PM To: toasters@mathworks.com Subject: RE: Storage space overheads!

Hi Glenn D, Glenn W, Andrew, Michael, Holland and Holland, Thank you all for your feeback.

To summarize, the following is true (i.e., about storage sacrifice): - Disk are right sized (realigning size in blocks from different vendors...eg., for 72GB we get 68GB) before using by NetApp. - dual parity (2 disks on RG of 16 => 14D+2P) - spare disk/s - WAFL & RAID overhead (~ 5%). - aggr reserve 5% default - vol snapshot reserve 20% default - for LUNs: 50% for base snapshot + extra space 20% of remaining

data space - user data should always be below 90% to avoid performance bottleneck and defrag issues.

That's sounds bad to me, but if you all feel to be true then its must be a solution.

How about putting SATA disks on the primary filers to save costs? Whats are the problems in replacing FC disks with SATA? Whats record about SATA disks, I heard they are terrible? (Failure rates is abnormal)

Thanks again for feedback. Marge.

Concerned about your privacy? Instantly send FREE secure email, no account required http://www.hushmail.com/send?l=480

Get the best prices on SSL certificates from Hushmail https://www.hushssl.com?l=485