Sector counts for replacing ZFS disk with different model

Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.

danwood82

Member
Feb 23, 2013
66
0
6
My OmniOS/napp-it server has dropped a drive (and looks like a second is on the way out, reallocated sector count appears to be rising), and like an idiot, I never ordered up any spares when I set the thing up - they're currently all Seagate Constellation CS 3TB disks, but that disk is EOL, and I can only get hold of it for >twice what it's worth, and on backorder.

I figured I'd swap it out for an HGST Deskstar NAS 3TB, but I'm not sure if that'll work. I figure it'll need to be either exactly the same sector count or higher... but how can I tell? Nobody seems to publish sector count data these days. Are all modern type drives of the same stated capacity interchangeable, or is this still a potential minefield?
 

Scott Laird

Active Member
Aug 30, 2014
316
147
43
I believe sector count for N TB hard drives is standardized now, but I can't find a reference to this anywhere.

I'm fairly confident that it's true for 3 TB drives, at least. I have a ton of different brands and models of 3T drives in my ZFS server. Some Seagate, some WD, some HGST, of a bunch of different generations. Some of them were 520-byte Netapp drives that were reformatted to 512-byte sectors. Every single one of the drives is 5860533168 sectors according to blockdev --getsz in Linux.
 
  • Like
Reactions: danwood82

danwood82

Member
Feb 23, 2013
66
0
6
Fantastic, many thanks for the detailed response!
Turns out, "5860533168" was the magic Google search term I needed... turned up a review of the HGST that quotes the exact same number, and reference to the Seagate stating it too.

Very useful info to know!
 

danwood82

Member
Feb 23, 2013
66
0
6
Just wanted to say thanks again. I got the replacement Deskstar NAS drives through - they were indeed byte-exact matches for the Seagate drives, and the swap and resilver went through without a hitch. Back up and running!
 

adgenet

Member
Apr 12, 2016
40
17
8
I thought when letting ZFS manage the partitioning, as in adding whole disks to a vdev as opposed to a pre-configured partition, it adds an 8MB partition to the end of the disk to account for potential differing sector counts between models of disks of the same overall size?
I was under the impression that this feature lets you replace, say a 1TB disk with another 1TB disk without worrying about the exact number of sectors.
My pool has disks that look like this in fdisk:
Code:
Device                                                       Start         End     Sectors  Size Type
/dev/sda-part1        2048 11721027583 11721025536  5.5T Solaris /usr & Apple ZFS
/dev/sda-part9 11721027584 11721043967       16384    8M Solaris reserved 1
Never had to deal with this directly but I thought that's how it worked