Updating SSD firmware on SAMSUNG drives...

[SiliconWizard]

Just asking for opinions here (although mine at the moment just seems to be: "if it ain't broken, don't fix it".)

In one workstation, I have two SAMSUNG SSDs, one 960 PRO NVMe (512GB), the other 860 EVO SATA (1TB).
The SAMSUNG Magician software shows an update available for both firmwares.

Is there any objective risk of updating? Any risk of data loss? Anyone has any experience doing that?

[ejeffrey]

Unless they have a changelog saying what is fixed and those apply to you I wouldn't update.  I always consider it a risk, most devices these days have fallback firmware and/or mechanisms to prevent partial or corrupted firmware updates, but I agree if it isn't broke don't fix it.  As far as I remember I have never updated firmware on any storage device.

[SiliconWizard]

My thought as well. The only time I updated a HDD firmware was a pretty long time ago on a HDD, and it was meant to fix a bug that could potentially corrupt data. But other than that, nope.
I haven't found a proper changelog for those updates... I think they are supposed to improve performance, from some unofficial info. But yeah, the drives just work fine, so I will most likely not bother.

I was just curious to know about people who would have done it on similar SSDs. Just to get some feedback.

[RoGeorge]

Just asking for opinions here (although mine at the moment just seems to be: "if it ain't broken, don't fix it".)

In one workstation, I have two SAMSUNG SSDs, one 960 PRO NVMe (512GB), the other 860 EVO SATA (1TB).
The SAMSUNG Magician software shows an update available for both firmwares.

Is there any objective risk of updating? Any risk of data loss? Anyone has any experience doing that?

There's always a risk, it depends.  I have a Samsung SSD 850 PRO 512GB, also NVMe, and updated it a couple of years ago, only weeks after I bought it.  I was using Samsung Magician and a Win10 machine back then, and it all went fine.  For me the update went OK, the old FW was working fine before the update, and apparently still works OK after the update.  That firmware was dealing with some speed loss when the disk was mostly full, but nothing noticeable without a benchmark.

What is the current FW update trying to fix or improve?

[SiliconWizard]

What is the current FW update trying to fix or improve?

Problem is, I don't know for sure. If anyone knows where the changelogs can be found...

[RoGeorge]

I couldn't find a list of fixes either, asking you in the hope Samsung Magician (which I don't have on Linux) can tell the changes.   

Their download page doesn't list the fixes, couldn't find any list in the .iso disk for FW upgrade either:
https://www.samsung.com/semiconductor/minisite/ssd/download/tools/#download_tab_0103_anchorpar3-st_semi_down_list_ex

[wraper]

In the past there were SSD which died or lost data due to bugs in older firmware. Like Crucial M4 which worked erratically after 5000 hours of operation. Samsung in particular had security problems in SED drives, so they were as secure as non-encrypted. https://www.syxsense.com/samsung-ssd-vulnerability/

[mrflibble]

If it's working fine as is, no need to upgrade firmware. I'd only consider flashing new firmware for a) significant performance increase under actual workload, or b) some vulnerability I actually care about for hardware in question.

If anyone knows where the changelogs can be found...

If anyone knew indeed. Firmware changelogs containing actual useful information are rare. And changelog entries like "have increased the number" or even "encryption make more mauve" do NOT count as useful, despite what some companies seem to believe...

[Wuerstchenhund]

Just asking for opinions here (although mine at the moment just seems to be: "if it ain't broken, don't fix it".)

In one workstation, I have two SAMSUNG SSDs, one 960 PRO NVMe (512GB), the other 860 EVO SATA (1TB).
The SAMSUNG Magician software shows an update available for both firmwares.

Is there any objective risk of updating? Any risk of data loss? Anyone has any experience doing that?

Do the update. The old "if it ain't broken, don't fix it" sayin' comes from a different time when many things were a lot simpler.

Developing and QA'ing firmware updates is expensive so manufacturers don't do it unless they have to. For SSDs, that usually means one of the following:

• To accomodate a new hardware revision
• To fix a bug
• To increase performance or durability

The updates for your drives are unlikely to exist solely for the first reason (new hardware support) as updates that only cover new hardware are not usually rolled out to existing drives unless there are also other things to fix (hw support only updates are normally released as intermediary updates that only get flashed into new production hardware).

So it's very likely either to fix some bugs or improve performance/durability, maybe even both. And in any case you want these updates as bug fixes often address issues with data corruption, security or other major issues which often aren't known in the public domain so the manufacturer wants to fix them before they become an issue (the absence of public release notes is often a good sign that this is the case).

SSD (and BIOS updates) should be treated like OS updates, i.e. you make sure you're on the latest version to avoid  problems later on. They are usually non-destructive (the time when a fw update wiped a SSD has long been gone) and are generally very reliable. You certainly don't want to risk your data encountering a problem that has already been fixed.

[Monkeh]

SSD (and BIOS updates) should be treated like OS updates, i.e. you make sure you're on the latest version to avoid  problems later on. They are usually non-destructive (the time when a fw update wiped a SSD has long been gone) and are generally very reliable. You certainly don't want to risk your data encountering a problem that has already been fixed.

Ahh.. I had to replace a motherboard earlier in the year because a BIOS update broke VT-d catastrophically and there's no way to roll back and no hope of getting it fixed.

[wraper]

I had to replace a motherboard earlier in the year because a BIOS update broke VT-d catastrophically and there's no way to roll back and no hope of getting it fixed.

Desolder Flash IC and program with $2.50 programmer

[Monkeh]

I had to replace a motherboard earlier in the year because a BIOS update broke VT-d catastrophically and there's no way to roll back and no hope of getting it fixed.

Desolder Flash IC and program with $2.50 programmer

Yes, and it's only a seven and a half hour flight away. Getting someone remotely to replace the board was a lot faster and less problematic than trying that..

[Wuerstchenhund]

SSD (and BIOS updates) should be treated like OS updates, i.e. you make sure you're on the latest version to avoid  problems later on. They are usually non-destructive (the time when a fw update wiped a SSD has long been gone) and are generally very reliable. You certainly don't want to risk your data encountering a problem that has already been fixed.

Ahh.. I had to replace a motherboard earlier in the year because a BIOS update broke VT-d catastrophically and there's no way to roll back and no hope of getting it fixed.

Most mainboards made in the last 10 years or so have a recovery function where you can flash a new BIOS in an unrecoverable board without any version checks, which should allow you to flash an older version.

Although, really, if a vendor releases a BIOS update which messes up something so basic as VT-d support then that says a lot about the (lack of) proper QA and I would seriously question if it's sensible to rely on that particular vendor's products.

[Monkeh]

SSD (and BIOS updates) should be treated like OS updates, i.e. you make sure you're on the latest version to avoid  problems later on. They are usually non-destructive (the time when a fw update wiped a SSD has long been gone) and are generally very reliable. You certainly don't want to risk your data encountering a problem that has already been fixed.

Ahh.. I had to replace a motherboard earlier in the year because a BIOS update broke VT-d catastrophically and there's no way to roll back and no hope of getting it fixed.

Most mainboards made in the last 10 years or so have a recovery function where you can flash a new BIOS in an unrecoverable board without any version checks, which should allow you to flash an older version.

Although, really, if a vendor releases a BIOS update which messes up something so basic as VT-d support then that says a lot about the (lack of) proper QA and I would seriously question if it's sensible to rely on that particular vendor's products.

No attempt at a recovery flash resulted in a downgrade. As for the vendor, that'd be Asus. One of their higher end boards, too. It's all a sea of garbage and assuming any of these vendors won't break things will bite.

Samsung are, perhaps, slightly more trustworthy - but it's worth noting the reason I took the opportunity to upgrade the BIOS on that board is it was out of service due to a desperately early failure of an extremely lightly used Samsung SSD..

[Wuerstchenhund]

No attempt at a recovery flash resulted in a downgrade. As for the vendor, that'd be Asus. One of their higher end boards, too. It's all a sea of garbage and assuming any of these vendors won't break things will bite.

Probably true, although I can't say as I stopped building PCs myself a very long time ago.

Samsung are, perhaps, slightly more trustworthy - but it's worth noting the reason I took the opportunity to upgrade the BIOS on that board is it was out of service due to a desperately early failure of an extremely lightly used Samsung SSD..

I'm not surprised, we had quite a few premature failures of Samsung SSDs so we stopped using them.

I'm not sure why they are so highly rated, they might be better than many other consumer-grade SSDs but even their enterprise SSDs left me underwhelmed.

[Monkeh]

I'm not sure why they are so highly rated, they might be better than many other consumer-grade SSDs but even their enterprise SSDs left me underwhelmed.

Because their early offerings were very fast and quite decent. Unfortunately, especially with the 'Evo' series in particular, they've gone down the QLC rabbit hole way, way too fast.

[Wuerstchenhund]

I'm not sure why they are so highly rated, they might be better than many other consumer-grade SSDs but even their enterprise SSDs left me underwhelmed.

Because their early offerings were very fast and quite decent.

True, but that was mostly because many other consumer SSDs were often worse

In the enterprise space Samsung SSDs have always been mostly one thing: cheap. They never stood out in either performance, reliability or endurance.

[Ed.Kloonk]

Who remembers when this happened in 2014..

https://www.extremetech.com/computing/190746-samsung-840-evo-has-serious-performance-bug-fix-on-the-way