Old work laptop - worth keeping or worth risking its return?

[KlaxonCow]

[edit] @K_C: I think once I saw some documentary about certain Google servers they really have "a metal shredder" the reduce HDs that need to be destoryed into pulp.

Well, yeah. But that's Google.

One, they can certainly afford it. To actually commission a dedicated device, just for their hard drive smashing needs.

Two, they're sitting on top of one of the biggest - if not THE biggest - "data silos" in existence. So they've got to take this sort of thing seriously, and basically be "the gold standard" for data privacy policies.

Three, I don't doubt their "metal shredder" really, really shreds it so small that it's essentially a powder. Yes, that's more than good enough to be sure that it's impossible to recover from.

I just mentioned fire, as I'm inspired by experience.

I bought a hard drive ages back - it used a molex power connector, so we're talking pre-SATA here (and I remember that it was 250GB capacity because this was gargantuan storage for the time) - and then I got a system fan for the case. But the fan didn't come with a fan header, it actually plugged into a molex for power. The way this dodgy little fan worked was that you plugged it into a molex power connector and then it had another molex power connector coming from it. Like, a "passthrough" kind of thing.

To be honest, in hindsight, it was a very dodgy set-up and I shouldn't have trusted it. But I plugged the fan into a molex from the power supply and then plugged the hard drive into the molex "extender" thing on the fan.

Well, their molex extender thing was dodgy as hell. There was a short circuit. The hard drive got so hot that the plastic melted and the motherboard was visibly fucked.

I remember it well, as I'd just paid good money for that hard drive and I'd just copied over photos of my brother's wedding on it. But, no, I had not got around to backing those up, as I'd only just got them off the camera after coming back from the wedding. So, yeah, I lost irreplaceable family photos to that shitty fan setup. I mean, all was not lost, as plenty of other folks took photos at the wedding and I went around gathering those up to replace the ones I lost - and probably now have more photos of that wedding than anyone - but my own photos were gone. Sent to digital oblivion.

So, yeah, my other advice is that if you see any dodgy power thing like that - where you don't plug it in directly, but have to "daisy chain" power to devices - then don't do it. It's not worth it (I mean, it's lucky that it happened while I was around, to smell burning, and then turn things off - if I'd not been in the room, then it could have easily started a house fire).

But I was inspired to also suggest fire, after seeing the state of the bloody hard drive after what was, after all, only a short circuit - for a relatively short time, as I smelt burning and then disconnected everything. It was physically mangled. Actually melted. You ain't coming back from that. And so I imagine that if you made a little "garden bonfire" and really burnt a hard drive, that would definitely physically mangle it to a level that absolutely no-one could ever recover it from. Particularly if, after it cooled down, you took a hammer to it as well.

To be absolutely sure, kill it with fire. Then bludgeon it repeatedly with a sledgehammer. That'll surely do it.



(Did you know that this music video was one of the first projects that Aardman Animations - of Wallace and Gromit fame - ever did? This is the full video that includes the credits at the end.)

[zip accident]

There are probably three main things to consider, is the battery replacable, do you want to play games and if so is the graphics card up to playing games (some on-CPU graphics are quite good, but a lot aren't), and how old is the machine? Not being able to replace the battery would be bad, a five year old machine would probably be borderline worthless, and graphics standards change.

[Thoughtcat]

Thanks for all the advice, guys!

[noodles]

You'll probably have to replace the SSD anyway due to wear, along with the battery, and it will need more RAM to really remain usable.

BTW, what is the lifespan of your average SSD? I could never find any real data on that in the early days for obvious reasons, but now that some years have gone by, I wonder if anyone has some good stats on this and by brand if possible ?

[Dio]

The stress tests tend to show that if they last the first year or two they usually go well past their stated write endurance (by, like 10x).

[Little stoned kiwi]

My take is self-biased experience and anocdotal instances, and think "it depends". I work on the front line supporting corporate PCs that the users torture by having the filesystem 97% full and half the company's SharePoint site syncing through OneDrive. They fail. Probably around the 4 year mark. If you aren't watching the SMART alerts they can catch you off guard by just not working one day.
Another factor is what's under the hood of the drive.

Of course it's possible 95% of systems will run for 10 more years without problem, I only look at the bad ones. However I don't trust them to be any more reliable than mechanical drives and consider them to fail more dramatically.

[KlaxonCow]

Being solid state - hence no moving parts - then you'd reckon that they ought to be vastly more reliable.

But this is offset by the fact that the manufacturing yield on flash memory is actually a bit shit.

So what Dio's saying makes abundant sense. That if you've got a good batch - the manufacturing yield was good - then it'll actually last a very long time (x10, as Dio suggests) because, you know, solid state and no moving parts, so less to go wrong and only thermal "wear and tear" to worry about.

But if you got a shit batch - poor yield - then it'll fail pretty quickly, which offsets that and means, on average, it's about the same as a mechanical drive. Even though, by rights, as solid state, it really ought to be vastly better.

Apparently, when you buy, say, a 32GB USB stick, it's quite possible that it's actually a 64GB stick but when they test the yield on them, half of it's no good, so they program the internal controller to dodge the bad bits and only use the good bits, and report itself as being a 32GB stick to the OS.

Yes, they're actually manufacturing, say, a bunch of 128GB sticks at the factory. But because the yield on flash is very hit-and-miss, a bunch of them will end up being reprogrammed to report as 64GB sticks, another really bad bunch will report as being 32GB sticks and so forth. Then they label them up accordingly and sell them.

So, actually, the manufacturing for this range of USB sticks is identical, but depending on how good the yield is on them, they'll relabel and sell them as lesser sticks for a cheaper price. Which is a good way to still make money, even off the bad yield. "Ah, though it wasn't a good yield, this one still has 16GB that's good, so let's just program the controller to report 16GB, only use the good bits of the flash and sell it as a 16GB stick".

Thus, it very much is a "it depends" answer. If it was a very good batch, then that flash will last way longer than expected - no moving parts, less to go wrong - but flash memory is a bit dodgy to manufacture, so if it's a bad batch then it'll fail miserably in short order. And the balance of this is that, generally, on average, it's probably no better than a mechanical drive. Though, by rights, if it was possible to always guarantee a good yield every single time, it ought to be so much better for being solid state.

The manufacturing can be so dodgy on flash that, as I say, they actually build that into their overall strategy. We'll make 128GB sticks and the ones that come out poorly can be relabelled as 64GB or 32GB - disregarding the half / three quarters of the flash that's not come out good.

Inside a USB stick is a microcontroller, so it can be reprogrammed with a "bad sectors" table that tells it not to use certain bad bits of the flash and only utilise the bits of it that passed the tests (and the microcontroller then presents it to the OS as if it's one contiguous bit of memory, but it's very possible it absolutely isn't).

The logic of this, though, is that you should probably buy the largest available capacity. Because that means it was a good yield for all of it to have passed the tests. Buying less capacity costs less, but it was supposed to be a higher capacity, but had a poor yield on it. So it's been reprogrammed and relabelled as a lesser capacity - but this also implies that, if a big chunk of it came out bad, then even the bits that did pass the tests as "good" was from a generally poor yield, so - by association - you might suspect that it's not got as much durability to it. So, if you can afford it, get the highest capacity in a range. Bet on the better initial yield as probably being indicative of it being more reliable in the long term.

[Little stoned kiwi]

That they have disabled the faulty cells doesn't mean the remaining cells are any less reliable. It happens in everything silicon.
I suspect buying a larger drive (larger than you need) allows the wear to be spread over a larger surface, giving it a better service life. That's speculation though.

[Orava]

A Windows 8 something laptop that overheats randomly? Worth keeping as an emergency contact being evicted/homeless?

[Orava]

Ok I am on the hunt for a new laptop any good ones?

[venusian]

I recently bought myself a Lenovo Yoga Slim 7 Pro X 14. Ryzen 7 6800HS (so goes fast and comes with the most recent integrated GPU) and a 3K 120Hz screen. For me it's the optimal combination of performance and slimness and lightness. I really like it.

[Foozwah]

Or, you know, to be really sure, melt the bugger first and then smash it with a hammer.

And then melt down the hammer. And smash that with another ha... Oh hang on, I see a flaw in my attempted "next-levelling".