5 ms·

Since a 7200 rpm disk can roughly transfer ~350 MB/s, it may take around to 21.6 hours to fill the 26 TB disk, and around 24 hours to fill a 30 TB disk. A 10,00

Since a 7200 rpm disk can roughly transfer ~350 MB/s, it may take around to 21.6 hours to fill the 26 TB disk, and around 24 hours to fill a 30 TB disk. A 10,000 rpm disk can transfer at around 600 MB/s, so it'll take roughly more than half then time of a 7200 rpm HDD.

KptMarchewa 5y ago
That's actually 2-3x speed I remember from where I was still using HDDs personally... They've actually got better.
hulitu 5y ago
Ok. So then i should not check for bad blocks when formatting a new partition ?
- thequux 5y ago
  There hasn't been any benefit to doing so for at least two decades. Most[1] disks since then transparently remap bad blocks, so you'll never encounter a read error from a just-written block, nor will you ever encounter a write error unless the disk is out of spare blocks. If the disk is out of spare blocks, the bad block check won't save you; the disk simply shouldn't be used. [1] I don't know of any counterexamples, but neither do I know that there are none.
  - rbanffy 5y ago
    > If the disk is out of spare blocks, the bad block check won't save you; the disk simply shouldn't be used. While I completely agree that data is more valuable than the spinning metal it lives on, if the bad block consumption rate is constant, there is little reason to assume the drive will die. You can probably use it as one of a dozen drives in a RAID-[0,1] read cache until it decides to self-retire.
xxpor 5y ago
Assuming a fixed number of platters, shouldn't (physically) sequential read and write speeds go up as capacities go up? The physical size is fixed, as is the rotation speed, so given the density increase, more data should be moving past the head for a given amount of time, right? Assuming also you don't hit the 6 gbit/sec limit of SATA III of course.

5 ms·

Since a 7200 rpm disk can roughly transfer ~350 MB/s, it may take around to 21.6 hours to fill the 26 TB disk, and around 24 hours to fill a 30 TB disk. A 10,00

by webmobdev 5y ago

KptMarchewa 5y ago
That's actually 2-3x speed I remember from where I was still using HDDs personally... They've actually got better.
hulitu 5y ago
Ok. So then i should not check for bad blocks when formatting a new partition ?
- thequux 5y ago
  There hasn't been any benefit to doing so for at least two decades. Most[1] disks since then transparently remap bad blocks, so you'll never encounter a read error from a just-written block, nor will you ever encounter a write error unless the disk is out of spare blocks. If the disk is out of spare blocks, the bad block check won't save you; the disk simply shouldn't be used. [1] I don't know of any counterexamples, but neither do I know that there are none.
  - rbanffy 5y ago
    > If the disk is out of spare blocks, the bad block check won't save you; the disk simply shouldn't be used. While I completely agree that data is more valuable than the spinning metal it lives on, if the bad block consumption rate is constant, there is little reason to assume the drive will die. You can probably use it as one of a dozen drives in a RAID-[0,1] read cache until it decides to self-retire.
xxpor 5y ago
Assuming a fixed number of platters, shouldn't (physically) sequential read and write speeds go up as capacities go up? The physical size is fixed, as is the rotation speed, so given the density increase, more data should be moving past the head for a given amount of time, right? Assuming also you don't hit the 6 gbit/sec limit of SATA III of course.