One went into read only mode, and I could copy all the data to a replacement drive (I think - it's possible there's corruption in files I haven't noticed).
Recently one I bought in 2016 failed hard. Was working fine one day, next reboot the machine didn't even recognise it had the drive plugged in. (I also didn't lose any data, because I don't trust _any_ drives, and have multiple backups of it...)
The puts me at a 66% failure rate over ~5 years for SSDs in my personal machines. I don't recall a spinning rust drive failing in that time, and I have quite a lot more of those in service.
Although personally I’ve never had it happen, I work closely with our support team who handles these issues. It happened to around 5% of staff towards the 4th year of use and about 100% more on the 5th year. Majority shut off and never came back on. A small percentage were able to turn on and we could retrieve all the useful data after a clone. Thankfully we have a policy that critical things should never be keep local.
In a PC, no. In a server, yes... About half a dozen of them. Later determined they were part of a defective batch from Kingston. They just stopped working. No recovery of data was achieved.
Ouch i hope you had mixed drives in a raid, its things like this which are the reason i make a point of mixing disks up a little in an array, its never happened to me but ive heard enough stories in the past to be prepared
I've had in the neighborhood of a hundred fail in servers at work. With two models (different manufacturers) a large number of drives failed within the first weeks of use. The other model fails randomly in our systems with raid controllers and sas expanders, possibly connected to a firmware issue on the ssds. In any case, the failure mode is the same -- the disk disappears from the bus, never to be seen again. The failure rate is much less than when we ran similar number of conventional hard drives, but the failure mode is much worse. We don't have a particularly high write rate, so none of the failures were connected to media wearout (to my knowledge)
No actually save for a freind hamfistedly managing to snap the sata connector on an OCZ drive some years ago ive yet to actually have one fail and my oldest one in regular use is about 6 years old
6 comments
[ 2.9 ms ] story [ 23.0 ms ] threadOne went into read only mode, and I could copy all the data to a replacement drive (I think - it's possible there's corruption in files I haven't noticed).
Recently one I bought in 2016 failed hard. Was working fine one day, next reboot the machine didn't even recognise it had the drive plugged in. (I also didn't lose any data, because I don't trust _any_ drives, and have multiple backups of it...)
The puts me at a 66% failure rate over ~5 years for SSDs in my personal machines. I don't recall a spinning rust drive failing in that time, and I have quite a lot more of those in service.