Update on Samsung SSD Reliability

Post by YoumuChan » Sat, Feb 04, 2023, 3:28 AM UTC

Aardwolf wrote: post #55

pentamassiv wrote: post #25

Earlier quoted context omitted.

A few months ago I already updated my Samsung SSD by following this procedure: https://askubuntu.com/a/1386451 . Theoretically they provide an image to boot from to do the update, but the image seems very outdated and did not recognize my keyboard so it was unusable.

I now found and followed this: https://blog.quindorian.org/2021/05/firmware-update-samsung-... And it seems to have worked. After extracting this updater tool and running it, smartctl kept showing the old firmware version (3B2QGXA7), but after reboot it now shows the new version (5B2QGXA7). I took the risk of running this while the OS (Archlinux) was running with the disk mounted (this is the OS install disk), and at…

  isoinfo -R -i xxx.iso -x /initrd | gzip -dc | cpio -idv --no-absolute-filenames "root/fumagician*"

Post by donmcronald » Sat, Feb 04, 2023, 3:35 AM UTC

vlovich123 wrote: post #132

rogers18445 wrote: post #129

Earlier quoted context omitted.

Those writes would have to be significantly smaller than the SSD's page (sector) size which is 512 bytes or 4 KiB. And would have to be written to different pages in rapid succession (to be flushed apart) - a standard serial write wouldn't trigger this even if it's 1 byte at a time, the OS FS cache would buffer it. It would have to be very misbehaving software or deliberate sabotage.

I’m pretty sure SSDs can only do 4kib aligned writes regardless of the FS sector size (under the hood it’s a write amplification unless the OS or controller manage to coalesce them. But yea, it depends on how things are getting flushed, but generally I wouldn’t expect too much magic unless you get lucky. It sounds like a small bug in the OS (ie these kinds of wires should be matched in memory in the application).

Post by HyperSane » Sat, Feb 04, 2023, 3:41 AM UTC

smiley1437 wrote: post #20

I've been trying to find a decent endurance NVME in the m.2 form factor for write-heavy applications and it appears that true 2-bit MLC has all but disappeared, replaced by 3-bit TLC and higher (with commensurate loss of endurance) The high endurance SSDs appear to be only available in u.2\u.3\hhhl and god-help-me EDSFF form factors Any suggestions? Micron's 7450 isn't readily available

Post by donmcronald » Sat, Feb 04, 2023, 3:43 AM UTC

Havoc wrote: post #109

walrus01 wrote: post #46

I wish more independent review organizations would conduct destructive "write lifespan until ultimate failure" real world tests on SSDs. With a mixture of real world large contiguous files and small random writes. Real ultimate write lifespan on 3-level-cell and QLC consumer grade SSDs varies wildly for things of the same capacity and similar price. Such as this series of tests from 7 years ago: https://techreport.co…

>destructive "write lifespan until ultimate failure" real world tests on SSDs >from 7 years ago It's from 7 years back for good reason. They stopped doing those tests when it became impractical as endurance increased. The drives are now good enough that you can't wear them out fast enough to make sense in a review setting ...unless fundamentally broken like these

Post by userbinator » Sat, Feb 04, 2023, 3:46 AM UTC

wahern wrote: post #138

userbinator wrote: post #111

For the same price, you can get twice the space for 1/4 the endurance, thrice the space for 1/8th the endurance, and now four times the space for 1/16th the endurance. Most people don't realise that is a horrible tradeoff, because NAND flash marketing and terminology like "TLC" or "QLC" is intentionally deceptive and manufacturers have been very secretive about the true endurance specifications, as well as trying to…

> as well as trying to overprice SLC out of production. Is it even possible to buy SLC drives any more? For the past 5+ years the only outlet I've been able to find that even advertise SLC is https://www.delkin.com/ , and you need to speak to sales to even get a price. I just assumed they and any other similar suppliers bought giant lots of chips at the tail end of SLC production and jack up the price on every new or…