Formating a new hard drive

Regular format overrides all data on the partition. Quick doesn't so your old data will still be there and will be overridden only as the new partition fills up.

You should usually use quick format.


LE: OK, regular format also performs scans as Davros pointed out.
 
Doing a full format will allow the drive to find and re-map any bad sectors that may exist on the drive. It's just a one-off thing anyway generally speaking, so it shouldn't be that much of a bother. :)
 
Doing a full format will allow the drive to find and re-map any bad sectors that may exist on the drive. It's just a one-off thing anyway generally speaking, so it shouldn't be that much of a bother. :)
Considering all HDDs are formated in the factory to remap those bad sectors as needed, what good does it do to redo it at home? It's not like 20 years ago when low-level formatting actually made some sense :)
 
Considering all HDDs are formated in the factory to remap those bad sectors as needed, what good does it do to redo it at home?
Theoretically, new bad sectors could have appeared during transit from said factory. In any case it doesn't exactly hurt to do it.

How long would it take to full format a 4TB drive?
Less than an afternoon, but not by much I suspect... Formatting my 2TB drive when I bought it a couple years ago took quite a while!

HDDs curiously behave a bit like tape drives these days, with the ever-decreasing data transfer rate/capacity ratio they're displaying as storage goes ever up and up and up.
 
Grall said:
Theoretically, new bad sectors could have appeared during transit from said factory. In any case it doesn't exactly hurt to do it.
Theoretically anything can happen, including the drive exploding during the initial format :)
Grall said:
Less than an afternoon, but not by much I suspect.
I want your afternoons :(
Assuming average-ish 100MB/s speed for the entire HDD (less on the inside, more on the outside edge) it would take around 11 hours to perform a full-format on 4TB drive. If you want to also actually do a proper check instead of just overwriting it with zeros as regular format does multiply that time a few times.
Grall said:
HDDs curiously behave a bit like tape drives these days, with the ever-decreasing data transfer rate/capacity ratio they're displaying as storage goes ever up and up and up.
Actually HDD speeds have scaled linearly with platter density.
 
Its a one shot deal think of it like running a burn in test when overclocking, yes you can get away with not doing one but its better to be safe.
ps: it would be much better to split that disk up into multiple partitions.
 
Actually HDD speeds have scaled linearly with platter density.
Obviously it hasn't, since you say it takes half a day to format a 4TB drive.

I can assure you it didn't take anywhere near that amount for any of the drives I've previously owned.
 
took 5 hours to full format new 2TB!:oops:
Was it some 5.xk RPM or 7.2k RPM drive? I took that 100MB/s from my 2TB 5200rpm drives. 7.2k ones are pulling off around 130-ish at similar platter densities. Even better would be if you'd say how many platters did the drive have.
 
Obviously it hasn't, since you say it takes half a day to format a 4TB drive.

I can assure you it didn't take anywhere near that amount for any of the drives I've previously owned.

Well, to be perfectly fair, we also have to consider that higher capacity drives have been adding more platters to the mix. Therefore, a linear relationship between platter density and sustained transfer speed could be entirely feasible, however the total time necessary to wipe the drive extended simply because of the additional platters. It looks like most of the 4TB units are using a five platter system, which is pretty crazy...

I think, as a general rule, hoho is probably very close to correct in that increases in sequential read/write speeds should be reflected in platter density increases. It just "makes sense" that, given the same linear speed under the RW heads, density is the only way for speed to change. More bits + same speed = more bitrate.

Edit: Anecdotally, it took almost 5 hours to format my first terabyte drive, it was one of the WD EARS drives when they were brand new and everyone was all scared of the 4K sectors. I too am a believer in a full format against a brand new disk.
 
Well, to be perfectly fair, we also have to consider that higher capacity drives have been adding more platters to the mix. Therefore, a linear relationship between platter density and sustained transfer speed could be entirely feasible, however the total time necessary to wipe the drive extended simply because of the additional platters. It looks like most of the 4TB units are using a five platter system, which is pretty crazy...

I think, as a general rule, hoho is probably very close to correct in that increases in sequential read/write speeds should be reflected in platter density increases. It just "makes sense" that, given the same linear speed under the RW heads, density is the only way for speed to change. More bits + same speed = more bitrate.

Edit: Anecdotally, it took almost 5 hours to format my first terabyte drive, it was one of the WD EARS drives when they were brand new and everyone was all scared of the 4K sectors. I too am a believer in a full format against a brand new disk.

Even with the same number of platters it depends on how the density was increased. 2x bits per track increase = 2x bitrate increase, 2x track density = no bitrate increase, 2x bits per track increase + 2x track density = 2x bitrate increase (2x longer to format)
 
Even with the same number of platters it depends on how the density was increased. 2x bits per track increase = 2x bitrate increase, 2x track density = no bitrate increase, 2x bits per track increase + 2x track density = 2x bitrate increase (2x longer to format)

Ah, yes I completely overlooked the TPI versus pure "BPI" aspect. Conceivably, both would likely increase as you drove the bits 'smaller', so you could theoretically double total the platter density, but only increase linear density by 1.4x (and then increase TPI by 1.4x) to give you the net doubling result. Your example of 2x TPI and 2x "BPI" would have resulted in a 4x increase in capacity, just FYI ;)

So, yeah, no linear speed increase is guaranteed. Good catch...
 
Your example of 2x TPI and 2x "BPI" would have resulted in a 4x increase in capacity, just FYI ;)

Edit: he, I first thought you were implying that I didn't know that the 3 examples did not have the same increase, then I thought you were right and I had made a typo, but then I noticed I hadn't made a typo so now I'm not sure any longer what you meant to to point out...
 
Last edited by a moderator:
That amount of time to do a brief test of the drive by doing a full format is nothing.

Before I add new drives into my unRAID array, I do 3 rounds of disk preclear. Each round reads the entire drive, writes full zeros to the drive, then performs a full read-back to validate that zeros were written. It also compares the SMART data at the beginning to the end and reports any possible issues or items to keep an eye on. The last time I precleared a 2TB WD Green drive it took 30 hours for 1 pass. The 3TB drives are up to 37 hours. I've had drives pass 2 full cycles but fail on the 3rd cycle. This is why regardless of the time it takes, I'll always run through 3 full cycles. (Preclear description)

It's always better to exercise your storage before you trust it with your data, especially with drives having a fatality bath-tub curve. Statistically speaking, the majority of drives die very early on or very late.
 
Was it some 5.xk RPM or 7.2k RPM drive? I took that 100MB/s from my 2TB 5200rpm drives. 7.2k ones are pulling off around 130-ish at similar platter densities. Even better would be if you'd say how many platters did the drive have.

It's a WD Caviar Green EARX 7200 rpm. I don't know how many platters and WD's website doesn't seem to mention it.
 
Last edited by a moderator:
Back
Top