• unpowered SSD's lose data

    From Felix@none@not.here to aus.computers,alt.os.linux.mint on Fri Nov 28 22:06:51 2025
    From Newsgroup: alt.os.linux.mint


    apparently they need to be powered up regularly

    https://share.google/pT5s7Z0hsWCDK0wFD
    --
    Linux Mint 22.2

    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Paul@nospam@needed.invalid to aus.computers,alt.os.linux.mint on Fri Nov 28 12:40:36 2025
    From Newsgroup: alt.os.linux.mint

    On Fri, 11/28/2025 6:06 AM, Felix wrote:

    apparently they need to be powered up regularly

    https://share.google/pT5s7Z0hsWCDK0wFD


    # Unwrapped link shortener...

    https://www.xda-developers.com/your-unpowered-ssd-is-slowly-losing-your-data/

    Yes, there is a lot to know about storage, and also
    about shortcomings in SSD firmware. Not all firmwares
    are created equal.

    NAND wears as you write it. 600 write cycles per cell is a typical number
    for TLC flash. Either you are using those 600, or if the drive has
    "freshen" code, it can use some of those cycles (as long as the drive
    is powered up occasionally versus leaving it on the shelf).

    I would think a TLC based SSD, left unpowered on the shelf for ten
    years, would be in very poor shape. Some brands would be "good but slow",
    other drives might well be undetectable.

    The message here, is you should be keeping backup copies. And if a
    drive has become unbearably slow, you can do a Macrium backup and
    a Macrium restore, to try and freshen it.

    Samsung had to issue a firmware update for one of their products,
    as it became evident it needed freshen code to make the drive
    "look like it was normal". Whereas my Lexar NS100 (my cheap "scratch drives"), the initial evidence is, powering them up does... nothing!!!
    Does not help to just power those ones. I would have to resort
    to backup and restore to freshen. (Macrium does a TRIM before it
    does the restore.) I can use Macrium, because it backs up both
    NTFS and EXT4, and all my Linux use EXT4 for that reason.

    Paul
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Lawrence =?iso-8859-13?q?D=FFOliveiro?=@ldo@nz.invalid to aus.computers,alt.os.linux.mint on Fri Nov 28 21:42:18 2025
    From Newsgroup: alt.os.linux.mint

    On Fri, 28 Nov 2025 12:40:36 -0500, Paul wrote:

    NAND wears as you write it.

    ItrCOs not even physical wear and tear. As I understand it, it has to do
    with accumulation of the wrong charges in the wrong places; if there
    was a way to reset those charges back to where they should go (with
    some additional circuitry in the right places, perhaps), the drive
    would be as good as new.
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Paul@nospam@needed.invalid to aus.computers,alt.os.linux.mint on Fri Nov 28 18:27:26 2025
    From Newsgroup: alt.os.linux.mint

    On Fri, 11/28/2025 4:42 PM, Lawrence DrCOOliveiro wrote:
    On Fri, 28 Nov 2025 12:40:36 -0500, Paul wrote:

    NAND wears as you write it.

    ItrCOs not even physical wear and tear. As I understand it, it has to do
    with accumulation of the wrong charges in the wrong places; if there
    was a way to reset those charges back to where they should go (with
    some additional circuitry in the right places, perhaps), the drive
    would be as good as new.


    It can have an infinite life, via thermal annealing to correct
    the defects. They've tested that in the lab, lift the silicon
    to a certain temperature, and the cells are writable again
    (for their 600 cycles).

    The problem with the lab discovery, is incorporating a heater
    to apply the annealing.

    But it would also not be "good for business", if NAND lasted
    forever, so you can see why they would be less than
    enthusiastic about the prospect.

    It would be easier to just go back to SLC :-) That would fix it.

    Paul
    --- Synchronet 3.21a-Linux NewsLink 1.2