Most of what you are referring to are software bugs. You know perfectly well what I’m talking about as it relates to Guild Wars 2.
No, I’d missed that part, sorry.
A power failure is very unlikely. Also, if the power fails during an install Windows uses a journaling file system, which means that the incomplete write would be marked as a failed write.
Nothing in your post relates to actual data “corruption” – unless there is a hardware issue such as bad RAM.
[Replied to this part at the top]
[…]
Incorrect data written to a file due to software bugs is not actual data corruption – it’s just incorrect data. Hardware problems can cause actual data corruption when the data would otherwise have been written correctly. THAT is what I was talking about and there’s a big difference between the two.
Hopefully that clears things up a bit.
A bit, yeah. My disagreement with you is mostly* in your claim that bad hardware is the only cause of data corruption.
Although rare (as you correctly pointed out), power failures/fluctuations can cause some SSDs (and probably some HDDs) to write bad data. Since, in this scenario, the write completes, a journaling filesystem won’t help. If the power supplied to the drive was momentarily out of spec, so you can’t really claim it’s “bad” hardware because of that one instance. Moving up the chain, the power supply must be bad then, right? Well, yeah, the computer’s power supply did momentarily put out “bad” power, but only because somebody crashed into a utility pole 5 blocks away, breaking one of the power lines, which caused surging and fluctuations in the remaining lines while things settled back down. This caused the power coming into your house to momentarily go out of spec. So the power supply isn’t “bad” per se, it just wasn’t designed to handle power that dirty. You could put a surge suppressor or UPS between your computer and the wall, but then I’ll just bring up lightning strikes and EMPs. Now, I’m not suggesting we stop using surge suppressors or UPSs. I’m just pointing out that everything is designed to work within a certain set of constraints, and that it’s pretty much always possible for the universe to put you outside of them. Once you’re in such a situation, the device isn’t “bad” simply because it did the wrong thing.
ECC memory solves a real problem (more so, as we start putting more RAM in our machines), but that doesn’t mean a flipped bit in non-ECC memory automatically indicates “bad” RAM, it’s just that the people who design consumer RAM are willing to tolerate an error rate that is, IMHO, too high.
Same thing for some causes of data rot… A few years ago, I had a drive spontaneously lose part of its filesystem data (hooray for backups!). Was the drive “bad”? Depends on how much you believe in the law of averages. The MTBF (Mean Time Between Failures) is exactly that: the mean, or average, not the minimum, time between instances of the drive writing, say, 0xFF, and then reading that same byte back as something else. Now, it’s been years since that drive tried to lose my data, and it’s worked fine ever since then. AFAICT, despite its early attempt at losing of some of my data, the drive hasn’t exceeded its MTBF spec. So, the drive is probably fine. I was just unlucky, both in the timing and in which particular piece of data it got wrong.
None of these issues are caused by bugs or bad hardware. With the possible except of dirty power (depending on where you live), they’re pretty rare, but they do happen.
*As to the role that bugs play, I believe you and I define “data corruption” differently. Shall we agree to disagree?
This isn’t fun!