Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Non-ECC memory corrupted my hard drive image [video]

youtube.com

Post reply on HN
Print view View original

81–90 of 161 posts

Re: Non-ECC memory corrupted my hard drive image [video]

#81

Post by pflanze »

tasubotadas wrote: post #7
TacticalCoder wrote: post #2

> Why ECC Memory Is So Important Except that it's not for many use cases. It's great for servers but for people on their personal and/or work computer, it's simply not that useful. Seriously: which percentage of developers have ECC on their development machine(s)? As developers we live in a world of SSH, cryptographic hashes, checksums everywhere, Git repositories (that is a big one), Merkle trees, digital signatures…

When people complain about these random OS crashes and freezes it's usually RAM corruption at fault. Per 1gb of RAM, you can expect to see 266 bit errors per month[1-2] if you are using your PC 16h per day. Multiply that by 64GB or 128GB of RAM and it's crazy to think that you won't run into any of the stability issues. [1-2] https://static.googleusercontent.com/media/research.google.c... [1-2] https://en.wikipedia.o…

> Per 1gb of RAM

The Google study you linked says "25,000 to 70,000 errors per billion device hours per Mbit".

Assuming bits, {25000 to 70000}/1e9 * 3016 1000 = 12 to 33.6 bit errors per month. Assuming bytes, it's 96 to 268 bit errors per month.

Apparently you've meant bytes, not bits. (I'm a pedant, but was also just unsure and interested in the numbers.)

FWIW, a comment I ran across that confirms toast0's view[1]: "It's a bimodal distribution - you either have many errors (due to a defect somewhere) or basically zero. If you're on the good side of the distribution, with only extremely rare errors, then you probably don't need ECC. But without ECC, you don't know whether you need ECC!"

[1] https://www.realworldtech.com/forum/?threadid=198497&curpost...

Re: Non-ECC memory corrupted my hard drive image [video]

#82

Post by vladvasiliu »

ilyt wrote: post #46
vladvasiliu wrote: post #34

Earlier quoted context omitted.

> Since then I'm very paranoid, always run memtest on any new RAM I buy overnight, and have ECC where it's possible to have it. Yeah, I do the same, but I've learned that you have to do it regularly. In one of my desktop machines, the RAM ran fine for like two years. Then, all of sudden, random Firefox segfaults, etc. Whipped up a memtest ISO, and sure enough, one of the sticks was bad.

Weirdly enough I had same case but memory turned out to be fine, replacing power supply fixed the issue. I ran test, saw memory is bad, replaced sticks, same problem, put the sticks back in and decided to just run it (it was gaming PC). Few months later powersupply outright died (had ~8years at that point), replaced it with good one, no memory errors.

In my case it was clearly a bad RAM stick. Took it out, OK. Switched them around: errors. Replaced it with a new one, back to OK.

In this particular case, a bad PSU would be the end of the PC. It's an HP dekstop mini. Basically a laptop without a screen, powered by and external adaptor that puts out a single 12V line. All further conversions are done on the motherboard somehow.

Re: Non-ECC memory corrupted my hard drive image [video]

#83

Post by vladvasiliu »

my123 wrote: post #56
OJFord wrote: post #36

Earlier quoted context omitted.

I realise that's blurrier when it comes to laptops, but AIUI it's more a case of whether the motherboard than supports it than about the AMD chip. i.e. given a desktop CPU, as far as I know you can put it in a motherboard that either does or does not support ECC RAM.

Ryzen APUs, which include almost all AMD laptops, actually have ECC fused off in silicon unless you buy the "Pro" variant.

Huh. I didn't know that.

My particular laptop does have a "pro" CPU. However, I would be surprised to no end to learn that it supports ECC. This particular model sports an MBP-level price tag [0], but is absurdly cheaply built. Even for "customer facing components", that are easy to compare, such as the screen (terrible colors) and case (creaks if you look at it wrong). HP doesn't offer ECC RAM, not even as an upgrade, so I really don't think the additional lines are physically present.

---

[0] I don't remember the specific number, but it was within 100 € of a 14" M1 MBP with 32 GB RAM and 512 GB SSD. That's counting a RAM (8 -> 32) and SSD (256 -> 512) upgrade which were made with components bought separately (though they were rather high-end).

Re: Non-ECC memory corrupted my hard drive image [video]

#86

Post by layer8 »

gruez wrote: post #55
Brian_K_White wrote: post #45

Earlier quoted context omitted.

The problem with untrusworthy memory (or any other component) is not that your system crashes, it's that it doesn't .

I don't doubt that non-ECC hardware experiences some non-zero number of bitflips per year. I'm just doubting the parent commenter's claim that non-ECC ram is causing computers to crash "frequently".

And the parent is pointing out that not crashing on bit flips is exactly the problem.

Re: Non-ECC memory corrupted my hard drive image [video]

#87

Post by IYasha »

metadat wrote: post #77
IYasha wrote: post #73

Earlier quoted context omitted.

Unfortunately, background scrubbing significantly increases power consumption and impacts performance as well.

Do you want (1) a higher rate of stable and correct computations to be performed at a slightly higher energy cost, or (2) a demonstrably less reliable device at a slightly lower energy efficiency? I'll go for #1 in most cases, as long as the system is to be relied upon for anything deemed important.

Me too, of course. I'm just highlighting downsides so people know what to expect.

Re: Non-ECC memory corrupted my hard drive image [video]

#88

Post by dale_glass »

everybodyknows wrote: post #80
dale_glass wrote: post #12

Earlier quoted context omitted.

> But really most of all I live in a world of Git repositories. My entire Emacs config is versioned under Git (I know YMMV but I like it that way). Some people version under Git their entire user dir. > Tell me how my lack of ECC is going to really make life miserable here? Git will break if RAM is bad just like anything else. That it checksums everything won't save you from checking in corrupt data, the filesystem i…

How trustworthy is git-fsck for detecting random-bit corruption?

Depends on how you look at it.

For actual verification, unless there's a bug my understanding is that very trustworthy. But by that point it's already too late. Okay, you know something is broken, but that won't give your good data back.

But that only tells you that Git data is intact and that all the hashes match. If git got a corrupted file to start with, then correctly hashed it, everything will verify 100% and still be broken.

Re: Non-ECC memory corrupted my hard drive image [video]

#89

Post by layer8 »

metadat wrote: post #78
aborsy wrote: post #66

Earlier quoted context omitted.

Synology offers ECC in 2023 consumer modes such as 923. Still the experience that synology btrfs provides is nowhere as good as ZFS (due to a lot of limitations).

I bought a 6 disk Synology a few months ago and it came with ECC by default. I did a cursory web search about this just now and ECC support appears to be the norm for 22 (as in the year 2022) model revisions and newer (thankfully!).

It’s because they use AMD CPUs now (already in the 21 models). The trade-off is that those CPUs have worse hardware codec support than the Intel ones they previously used, if you want to do video transcoding.
Post reply on HN