On 2014-07-07 12:32 (GMT+0200) jdd composed:
Hello reading the other thread about advantage of 4096 sector size disk, I noticed one of my (spare) 1Tb backup disk is given by hdparm to have 4096 sector size, and I wanted to use it.
however I'm puzzled.
what ever I do, hdparm always says logical 512 / physical 4096, fdisk and YaST 512/512
first tryed with yast, modifying the filesystem option to 4096, then tried manually with:
fdisk -b 4096 /dev/sdp
remove partitions, create new as default -hole disk), write
disque /dev/sdp : 8001.6 Go, 8001639088128 octets, 1953525168 secteurs Unités = secteur de 1 × 4096 = 4096 octets Taille de secteur (logique / physique) : 4096 octets / 4096 octets
but hdparm says:
Logical Sector size: 512 bytes Physical Sector size: 4096 bytes
and fdisk -l
Disque /dev/sdp : 1000.2 Go, 1000204886016 octets, 1953525168 secteurs Unités = secteur de 1 × 512 = 512 octets Taille de secteur (logique / physique) : 512 octets / 512 octets
so what?
So read the whole other thread again, paying closer attention to: http://www.seagate.com/tech-insights/advanced-format-4k-sector-hard-drives-m... I don't know that operating systems will be overhauling themselves any time soon, if ever, to make a minimum filesystem allocation size 4096 bytes instead of 512 bytes. It's much more wasteful. Also, 512 is intertwined with too many kernel, support system and application components. IOW, this 512/4096 thing is really about I/O efficiency. Just because disk makers have agreed they are committed to 4096 internally does not mean the rest of the universe must make the same atomic adjustment. So for the foreseeable future and beyond, the only things you should be concerned with are: 1-what is reported for optimal I/O size (minor concern) 2-I/O alignment to physical disk sectors (primary concern) The filesystems you use are likely already using a blocksize that matches the "advanced format" physical sector size, leaving the only concern #2, ensuring logical disk divisions are aligned to 4k multiples. You don't want to see tools mentioning an obsolete CHS of ?/255/63 or ?/240/63 as have been typical since a couple of decades or so ago. So, watch what happens when you create /dev/sdX1. Except on a disk with genuine 512 byte physical sectors, you do not want to see: Device Boot Start End Blocks Id System /dev/sda1 63 80324 40131 83 Linux Start = 63 == bad. Block count not evenly divisible by 8 == bad. On disks with 4096 byte physical sectors, you do want to see e.g.: Device Boot Start End Blocks Id System /dev/sda1 2048 821247 409600 83 Linux Start = 2048 == good. Block count evenly divisible by 8 == good. If using GPT, the desired start sector 2048 offset will happen automatically. Don't pay much attention to the logical/physical lies often reported about disks more than two years old. If in doubt about the physical sector size and it bothers you, consult the manufacturer specifications for your disk. Also for most uses, it hurts nothing to assume 4096 byte sectors in partitioning disks that have 512. The only situations where a 4096 blocksize assumption can hurt is for filesystems purposely configured with small blocksize because they play host to a huge number of small files. WRT $SUBJECT, you want a filesystem blocksize of 4096, not a sector size of 4096. Leave the OS, tools and applications to continue using the 512 byte standard sector size as always since the birth of the IBM PC. -- "The wise are known for their understanding, and pleasant words are persuasive." Proverbs 16:21 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org