On Tue, 13 Jun 2017, John D. Baker wrote: > On Tue, 13 Jun 2017, Patrick Welche wrote: > > > > Doesn't RAIDframe assume 512-byte sectors? So, 64k stripe size would > > > be SecPerSU=128, yes? > > > > I think it's SecPerSU=64 blocks = 32k. Two data disks in one write, so > > 2 * 32k = 64k = MAXPHYS written in one write. > > > > (Please correct me if I'm wrong: I just make a 3 * 2T raid 5 this > > morning ;-) )
Clarification from mlelstv, prompting me to re-read "raidctl(8)" more closely shows where I got confused: StripeUnit (SU) == data component (disk) This is what I'd misunderstood the whole time. It turns out the 2TB disks for the planned RAID-5 also have 512-byte native sectors, not 4K sectors. So no alignment constraints after all. I have a 3-disk RAID-5 across substantially smaller disks (with native 512-byte sectors) whose configuration I shall have to revisit. Fortunately, its contents are easily expendable for testing. -- |/"\ John D. Baker, KN5UKS NetBSD Darwin/MacOS X |\ / jdbaker[snail]mylinuxisp[flyspeck]com OpenBSD FreeBSD | X No HTML/proprietary data in email. BSD just sits there and works! |/ \ GPGkeyID: D703 4A7E 479F 63F8 D3F4 BD99 9572 8F23 E4AD 1645
