Take the 2-minute tour ×
Server Fault is a question and answer site for professional system and network administrators. It's 100% free, no registration required.

It is well documented on Wikipedia that a head crash in a hard disk drive (HDD) may result in some data loss. However, there is not much being written about the different modes of a solid state drive (SSD) failure, and whether such failures would lead to significant data loss. Articles that I found on the web mentioned mostly about the complexity of recovering data from an SSD, but nothing is mentioned on the comparative likelihood of an irrecoverable data loss. Has anyone done such a comparison?

share|improve this question
    
As a single SSD or as a RAID member? –  ewwhite Jan 21 '13 at 8:48
    
@ewwhite, a single SSD. –  Question Overflow Jan 21 '13 at 9:04
    
The only single SSD failures I've had were due to write exhaustion or units that were DOA. –  ewwhite Jan 21 '13 at 9:09

2 Answers 2

In my experience when an SSD dies it's gone and you're reaching for your backups. You may get some data recovered by a specialist data recovery service but it's very unlikely you will be able to recover any data yourself.

share|improve this answer
    
I understand that backups are important, but data after the last backup will be loss if that is the case. –  Question Overflow Jan 21 '13 at 9:18
    
@QuestionOverflow Profesional backup may be faster - i do database server backups all 5 minutes. And even without backup some technology supports replicating changes out in real time. Also, the SSD may just be a CACHE, and not hold the "original" data. Some raid controllers can do that automatically. Also check adaptec hybrid raid (raid 10, mixing ssd and normal disc). –  TomTom Jan 21 '13 at 9:26
    
@QuestionOverflow: Yes but there are ways to mitigate against this for example using RAID for availability and backups for insurance. –  Iain Jan 21 '13 at 9:33
    
@TomTom and lain, thanks for both of your inputs. I was thinking about either some kind of continuous data backup plan or RAID 1 in addition to a daily backup. Just not sure which is better. –  Question Overflow Jan 21 '13 at 9:35

Technically I was told a lot of SSD failures are write failures - i.e. the data is there and can be read, just not written. Also "wear and tear" is destroying bits, not platters, so a Problem may be localized (data can be read EXCEPT a sector or so, while on a disc the head may be destroyed).

Not sure how much of that was "real", but that was info I was told when asking.

My own experience is more on the "ok, that damn Thing does not even Show up on the port" side, which indicates a total data loss.

share|improve this answer
    
That's my experience too - when an SSD dies the system can't see it any more. –  Iain Jan 21 '13 at 8:59
    
I think the problem is that there are 2 deaths here. One is "technical failure" (what we had), the other is "failure due to use" (max writes per sector). All my failures were fast (within a week) and basically faulty drives. I had not a single one fail after a LONG time of use. –  TomTom Jan 21 '13 at 9:03
    
Mine have been on older drives >1 year and relatively light usage. One was in my home desktop, it installed some updates and I rebooted - never came back. –  Iain Jan 21 '13 at 9:10
    
This seems to mean that sudden failure is more irrecoverable than failure due to wear and tear. –  Question Overflow Jan 21 '13 at 9:21
    
Yes and no. At the end I had hard disc brutally fail, too. And my sample size makes any statistical inclination totally idiotic (I.e. it has zero meaning). SO, the "how often DO they fail" part is impossible to answer. For example I lost 4 SSD within a week, but all were from the SAME order, SAME shipment - could be a production or even transport error of some sort. –  TomTom Jan 21 '13 at 9:24

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.