100TB
That's a lot of smut.
Seagate has set a course to deliver a 48TB disk drive in 2023 using its HAMR (heat-assisted magnetic recording) technology, doubling areal density every 30 months, meaning 100TB could be possible by 2025/26. HAMR overcomes the tendency of smaller magnetised areas in current PMR (perpendicular magnetic recording) technology to …
I'm looking right now to upgrade my current 1TB disk, which suddenly has panic attacks of crushing inadequacy after it saw this article. The reason is, of course, in no way shape or form related the above-mentioned materials. Nu-uh. No Sir. No way, no how. Nope. Nothing to see here, move along...
Was waiting for some joker to complain about rebuild time, which is inevitable in every article about "bigger hard drives coming". If you aren't already using forms of RAID that distribute parity instead of putting it on a single drive, and support the failure of more than one drive in a group, then you deserve a week long rebuild.
No one will force you to buy 100TB drives, you can keep using 500GB drives and get great rebuild times if sticking with antiquated RAID5 is that important to you.
Well - for a start - RAID5 *does* distribute parity across drives, but the main benefit to doing so is random write (and a lesser extent read); rebuilds require reading the whole rest of the stripe anyway, so only gain from the improvements to concurrent access.
Multiple parity (substantially) improves your odds of a successful rebuild - but it certainly doesn't make it faster - it can only be slower than an equivalent single-parity configuration as there is more computational overhead.
Ultimately, any RAID rebuild requires reading data equivalent to the capacity of the entire (sub)unit being rebuilt, and writing out the entire rebuilt drive - possibly whilst also satisfying routine IOs.
Depends on how they increase the storage size whether by increasing the track density or the DPI and if say you assume the drive will be 4 times bigger by doubling both then it will still only take twice as long to access all of it as you will get double the data in each rotation of the platter so ZFS Raid-Z3 will probably suffice and it doesn't even matter if they are shingled because even Solaris introduced sequential re-silver I think in version 11.2 in 2014. If you don't use the raid array whilst its re-silvering you're hardly working the actuator heads as they will be gently clicking over from one track to the next and the fluid bearings won't wear out and also being enterprise drives they will be adequately cooled so no problems there either. I've sequentially re-silvered an idle ten drive Raid-Z2 array of 3TB Toshiba DT01ACA300's in 7 hours with an I7-4820K CPU processing 1 GB's and my other array in another PC with an E5-2670 v1 CPU does 1.3 GB's and in enterprise systems with more grunt they should go even faster, so yes it could take days but so what? with ZFS the procedure should eventually complete successfully no matter how long it actually takes.
I believe the idea is that the write head isn't powerfully enough to actually write on it own. The big issue has been making the write head smaller, but HAMR bypasses that. The laser is smaller then the write heads writing area, but as the write head can't write without the laser it can only write to the area that the laser has heated.