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

I tested four NVMe SSDs from four vendors – half lose FLUSH'd data on power loss (2022)

twitter.com

Post reply on HN
Print view View original

161–170 of 261 posts

Re: I tested four NVMe SSDs from four vendors – half lose FLUSH'd data on power loss (2022)

#161

Post by Sakos »

wannacboatmovie wrote: post #67
bugfix wrote: post #63

Earlier quoted context omitted.

I had this exact experience with my workstation SSD (NTFS) after a short power loss while NPM was running. After I turned the computer back on, several files (package.json, package-lock.json and many others inside node_modules) had the correct size on disk but were filled with zeros. I think the last time I had corrupted files after a power loss was in a FAT32 disk on Win98, but you'd usually get garbage data, not al…

They may be pointing to unallocated space which on a SSD running TRIM would return all zeros. NTFS is an extremely resilient yet boring filesystem, I cannot remember the last time I had to run chkdsk even after an improper shutdown.

As somebody who worked as a PC technician for a while until very recently, I've run chkdsk and had to repair errors on NTFS filesystems very, very, very often. It's almost an everyday thing. Anecdotal evidence is less than useful here.

Re: I tested four NVMe SSDs from four vendors – half lose FLUSH'd data on power loss (2022)

#162

Post by kristopolous »

Sakos wrote: post #160
kristopolous wrote: post #154

Under long term heavy duty, I've routinely seen cheap modern platter outperform cheap brand name NVME. There's some cost cutting somewhere. The NVMEs can't seem to sustain throughput. It's been pretty disappointing to move I/O bound workloads over and not see notable improvements. The magnitude of data I'm talking about is 500-~3000GB I've only got two NVME machines for what I'm doing so I'll gladly accept that it's…

1) Nobody says otherwise about cheap anything NVMe. They're pretty terrible once they've exhausted the write cache. This is well-known and addressed in every decent review by reputable sites. 2) Sustaining throughput seems the least of our problems when some unknown number of NVMe SSDs might be literally losing flushed data.

Is this expected with say Samsung evo 9X0 pro? Or is there another tier above consumer level gear? Is there something I should go with?

Re: I tested four NVMe SSDs from four vendors – half lose FLUSH'd data on power loss (2022)

#163

Post by rft »

automatic6131 wrote: post #130
lapsis_beeftech wrote: post #119

Earlier quoted context omitted.

Is nitter.net still functional for you? The service was accessible earlier this year from my location but now it is all 403 Forbidden.

nitter.net is, the mirrors are not. iirc, twitter turned off all anonymous access, unless you come from a search engine, then you get a limit number of requests. So zedeus came up with an idea to make a massive pool of search engine'd api tokens and use those to keep nitter up. The mirrors would have to copy that idea, and few (0?) have atm.

The current state is to use guest accounts [1]. If you use that branch instead of master and get a collection of accounts, you can run your own instance. For my personal use instance this is working fine with the initial set of accounts I put in. No idea how long this will work and/or when it is going to get merged, it is a moving target.

I think the token workaround you mentioned is the old way that no longer works, but I am not sure.

[1] https://github.com/zedeus/nitter/pull/985

Re: I tested four NVMe SSDs from four vendors – half lose FLUSH'd data on power loss (2022)

#164

Post by kristopolous »

efxhoy wrote: post #159
kristopolous wrote: post #154

Under long term heavy duty, I've routinely seen cheap modern platter outperform cheap brand name NVME. There's some cost cutting somewhere. The NVMEs can't seem to sustain throughput. It's been pretty disappointing to move I/O bound workloads over and not see notable improvements. The magnitude of data I'm talking about is 500-~3000GB I've only got two NVME machines for what I'm doing so I'll gladly accept that it's…

I think cheaper QLC chips use a part of their storage space as SLC, which is fast to write. But once you’ve written the fast part that fits in the SLC cache write throughput quickly tanks as it has to push the data further in to the slower QLC parts.

Yeah I guess it works well for how most people use computers which is not actually for computation...

Modern platter is actually pretty decent and cheap. It's probably still the way to go for large loads unless you have a grove of money trees

Re: I tested four NVMe SSDs from four vendors – half lose FLUSH'd data on power loss (2022)

#165

Post by matja »

r1ch wrote: post #54

We shipped a shader cache in the latest release of OBS and quickly had reports come in that the cached data was invalid. After investigating, the cache files were the correct size on disk but the contents were all zero. On a journaled file system this seems like it should be impossible, so the current guess is that some users have SSDs that are ignoring flushes and experience data corruption on crash / power loss.

Journaling filesystems (including NTFS, and ext3/ext4 using default mount options) typically only track file structure metadata in the journal, so that is WAI - the filesystem structure was not corrupted, but all bets are off when it comes to the contents of the files.

Re: I tested four NVMe SSDs from four vendors – half lose FLUSH'd data on power loss (2022)

#166

Post by andybak »

WendyTheWillow wrote: post #158
andybak wrote: post #155

Earlier quoted context omitted.

I can't take this seriously. Sorry.

Yet again monetary selfishness wins over principle. Maybe you should stay on Twitter. Edit: Based on your website, who exactly are you trying to reach? Prospective clients? On Twitter? You say "users of your app", but it seems like that isn't a huge volume of people, and were you to move to Bluesky, they would almost certainly follow... Unless it's not worth it for them?

You're looking at my personal website. And I'm really not engaging any more.

Re: I tested four NVMe SSDs from four vendors – half lose FLUSH'd data on power loss (2022)

#167

Post by dspillett »

bugfix wrote: post #63
r1ch wrote: post #54

We shipped a shader cache in the latest release of OBS and quickly had reports come in that the cached data was invalid. After investigating, the cache files were the correct size on disk but the contents were all zero. On a journaled file system this seems like it should be impossible, so the current guess is that some users have SSDs that are ignoring flushes and experience data corruption on crash / power loss.

I had this exact experience with my workstation SSD (NTFS) after a short power loss while NPM was running. After I turned the computer back on, several files (package.json, package-lock.json and many others inside node_modules) had the correct size on disk but were filled with zeros. I think the last time I had corrupted files after a power loss was in a FAT32 disk on Win98, but you'd usually get garbage data, not al…

> but you'd usually get garbage data, not all zeros.

You are less likely to get garbage with an SSD in combination with a modern filesystem because of TRIM. Even if the SSD has not (yet) wiped the data, it knows that a block that is marked as unused can be retuned as a block of 0s without needing to check what is currently stored for that block.

Traditional drives had no such facility to have blocks marked as unused from their PoV, so they always performed the read and returned what they found which was most likely junk (old data from deleted files that would make sense in another context) though could also be a block of zeros (because that block hadn't been used since the drive had a full format or someone zeroed free-space).

Re: I tested four NVMe SSDs from four vendors – half lose FLUSH'd data on power loss (2022)

#168

Post by vladvasiliu »

kristopolous wrote: post #162
Sakos wrote: post #160

Earlier quoted context omitted.

1) Nobody says otherwise about cheap anything NVMe. They're pretty terrible once they've exhausted the write cache. This is well-known and addressed in every decent review by reputable sites. 2) Sustaining throughput seems the least of our problems when some unknown number of NVMe SSDs might be literally losing flushed data.

Is this expected with say Samsung evo 9X0 pro? Or is there another tier above consumer level gear? Is there something I should go with?

I don't know about the 950 pro specifically, but when I bought my 980 pro I looked into this, and it seemed that this drive does have a drop in write speed after a while (can't remember how long) but "low speed" wasn't that low. Again, don't remember specifics, but it was above 1 GB/s. Other drives fared much worse, with a drop coming in sooner and going lower.

Depending on your needs this can be an issue.

For me, using this drive for random "office" work, I figured I'd never feel it in practice. This drive is supposed to support PCIE 4, but my laptop only does 3. This also "helps", since it won't fill whatever cache it uses as quickly. In practice, it was able to write 100 GB at the top speed. Didn't bother to test more. The only time I've ever written that many data at a time was restoring a backup when I bought the drive. Since my backup was on 2.5" spinning rust, it wasn't an issue.

Re: I tested four NVMe SSDs from four vendors – half lose FLUSH'd data on power loss (2022)

#169

Post by WendyTheWillow »

andybak wrote: post #166
WendyTheWillow wrote: post #158

Earlier quoted context omitted.

Yet again monetary selfishness wins over principle. Maybe you should stay on Twitter. Edit: Based on your website, who exactly are you trying to reach? Prospective clients? On Twitter? You say "users of your app", but it seems like that isn't a huge volume of people, and were you to move to Bluesky, they would almost certainly follow... Unless it's not worth it for them?

You're looking at my personal website. And I'm really not engaging any more.

Serious question; what responsibility do you have as a consumer of a product, if any, beyond yourself?

Re: I tested four NVMe SSDs from four vendors – half lose FLUSH'd data on power loss (2022)

#170

Post by kristopolous »

vladvasiliu wrote: post #168
kristopolous wrote: post #162

Earlier quoted context omitted.

Is this expected with say Samsung evo 9X0 pro? Or is there another tier above consumer level gear? Is there something I should go with?

I don't know about the 950 pro specifically, but when I bought my 980 pro I looked into this, and it seemed that this drive does have a drop in write speed after a while (can't remember how long) but "low speed" wasn't that low. Again, don't remember specifics, but it was above 1 GB/s. Other drives fared much worse, with a drop coming in sooner and going lower. Depending on your needs this can be an issue. For me, us…

Thanks. I guess I can't really blame them for cutting corners on a use case that 99.9% of their customers never see.

I'd probably advocate the same.

Post reply on HN