If it didn't actually zero out the data then parity/data reconstruction for failed disks would not work - you'd end up with garbage since blocks that parity though were zero were actually not.
Simon.
-----Original Message----- From: owner-toasters@mathworks.com [mailto:owner-toasters@mathworks.com] On Behalf Of Blake Golliher Sent: Friday, February 17, 2006 10:15 PM To: Adrian Ulrich Cc: toasters@mathworks.com Subject: Re: wiping disks on an old F230
Or maybe, it just unallocated those blocks, and says "that data there? just ignore it". Which to me is more likely.
-Blake
On 2/17/06, Adrian Ulrich toaster@blinkenlights.ch wrote:
I don't belive spares_zero acutally overwrites data,
It does.
Proof: Adding a 'zeroed-out' disk to an aggr/trad_volume takes just a few seconds.
Why? -> Because the filer *knows* that the disk is filled with \0's so the parity-disk does not have to care about the new disk.