[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: firefox-37, where to put

On Friday 03 April 2015 20:22:18 Chris Bannister wrote:
> On Fri, Apr 03, 2015 at 05:39:32PM -0400, Gene Heskett wrote:
> > Usually less than 4 megabytes of free space left at the end of the
> > disk with a used 4k per sector disk on the cable, but could be zero
> > for a 512 byte per sector used disk since there is not normally an
> > alignment problem with the old 512 byte per sector formatting.  I
> > have 1Tb disks that look alike at first glance.  The 512 byte per
> > sector pair is heavier and a wee bit thicker because it likely has
> > two platters in it, while the 4k version pair is a bit lighter and
> > thinner, I presume because there is only one platter in those two
> > disks.
> >
> > The commodity drives I have coming will be, at 2Tb, 4096 bytes per
> > sector, and linux must align its writes with a Read-Modify-Write
> > cycle updating the whole 4k just to change one byte if things don't
> > start on a sector boundary.  There's a pretty good speed penalty for
> > doing that. A disk that can write at 120+ megs a second when aligned
> > can be turned into a 20 megs a second slowpoke. if miss-aligned.
> Ahh!, weren't you mentioning performance problems?
> There was a recent discussion on the netbsd-users list about 4k disks.
> In particular this:
> http://mail-index.NetBSD.org/netbsd-users/2015/03/27/msg015986.html

Bootability doesn't seem to be a problem unless I try to boot a kernel so 
old it has no clue what to do with this disk: (from sudo 
smartctl -a /dev/sda)

>Sector Sizes:     512 bytes logical, 4096 bytes physical

No clue but following a suggestion, the first partition actually starts 4 
mebibytes into the disk, which seems to guarantee proper alignment with 
the physical sectors.

Cheers, Gene Heskett
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page <http://geneslinuxbox.net:6309/gene>

Reply to: