←back to thread

203 points binwiederhier | 10 comments | | HN request time: 0.647s | source | bottom
1. RedShift1 ◴[] No.45050411[source]
Is it actually killing the SSD (SSD can no longer be used) or just corrupting the data on the SSD? It's hard to make out from all the comments and news articles.
replies(3): >>45050605 #>>45051152 #>>45053977 #
2. crest ◴[] No.45050605[source]
"Just" corrupting your filesystem...
replies(1): >>45051375 #
3. asimovfan ◴[] No.45051152[source]
some data might be worth way more than any SSD.
replies(1): >>45053258 #
4. dspillett ◴[] No.45051375[source]
Relative seriousness, both drive damage and filesystem damage are both bad but by slightly different degrees.

There is more chance of being able to fix data corruption, than being able to fix a bricked drive or one with unbearable blocks.

replies(1): >>45053272 #
5. X0Refraction ◴[] No.45053258[source]
If it is then storing it without backups sounds like a bad idea
6. dspillett ◴[] No.45053272{3}[source]
Self rely as I'm too late to edit out a slide-keyboard error: unbearable -> unreadable
7. toast0 ◴[] No.45053977[source]
I've seen lots of SSDs die suddenly (no longer visible on the bus), so I would assume that is what is happening based on the words people are using. I've yet to see an SSD fail to read only mode like they're supposed to... and there's rarely any warning, just working or dead (although I did have a couple that went from working to terribly slow while doing a large reallocation, and we replaced those rather than find out what would happen over a longer term)

That said, people use words with a different meaning all the time, and data corruption could fit as a failure.

replies(1): >>45054226 #
8. RedShift1 ◴[] No.45054226[source]
Failing to read-only is only an Intel thing, I've not seen any other SSD do that...
replies(2): >>45055048 #>>45060304 #
9. toast0 ◴[] No.45055048{3}[source]
I've not seen an Intel SSD do it either, although I've seen many of them escape their earthly existence :P

There was a firmware bug, but updating the firmware was inconvenient, and the specific interaction that caused the failure wasn't stated, so I couldn't avoid whatever it was; seemed connected to being pretty idle... we had a second data center as an untested "warm" failover target, and disks would tend to die over there where nothing significant was happening.

10. kimixa ◴[] No.45060304{3}[source]
I had a Crucial drive fail to read-only.

I got the data off, but most of the data wasn't really that important so there might have been dead regions.

I feel that many consumers won't really know if it's still readable, I'd suggest that 90% of people just have a single drive, and windows doesn't cope with a non-writable root drive particularly well.