The question is about understanding a failure or fault of an SSD drive that is not total, and not predicted or visible in SMART.

The FAQ for Server Fault says: "Server ... hardware, software ... Operations, maintenance, and monitoring". I believe the question fits these categories.

I'm not asking so much about a specific device or model, but whether the problem is repairable (using knowledge of SSDs outside my understanding) or predictable (using knowledge of SMART outside my understanding).

Please can you help me rephrase my question so that it's not considered off-topic?

  • 5
    I can't help thinking that the question should have been closed as 'not constructive' rather than off topic, but on the whole I don't see it as a good question. The answer to any hard drive failure question is "Diagnose it, if it's bad throw it away and get another one, if the data on it was important then you'll have a backup, so restore the backup". There's not much else that can be said. This isn't a SSD issue (normal drives can fail after 3 months) or a fault with SMART (it tries to predict when a failure might happen, absence of a prediction isn't a predition that it won't fail anyway) – Rob Moir Mar 6 '13 at 17:24
  • Sorry, then the aim of my question was not clear. I didn't necessarily want to diagnose what's wrong with this particular drive. I wanted to know: why is auto reallocate failing (am I doing it wrong?) and is there anything I missed in the SMART data (am I reading it wrong?). I've edited the question to make it clearer why I posted it here and what I'm looking for. Thanks for the advice. – qris Mar 6 '13 at 17:45

I don't see any reason it should be closed as off-topic. It may be Too Localized (another close reason), and you might never get a good answer.

Note: Kingston drives are ultra-cheap crap that I would recommend against using in any professional environment. It's incredibly unlikely that the problem is repairable. Grab what data you can, RMA the drive, and use backups to restore whatever you couldn't get off the drive.

  • 2
    Agreed - About the only reason I can think of that this got closed as off-topic is the possibly-incorrect assumption that nobody in their right mind would put a Kingston SSD in a server destined for use in a professional environment (i.e. people incorrectly assumed it was a "home use" question). That said, RobM's comment pretty much sums up how we treat hard drive failure questions. We should probably find a canonical questionfor them. – voretaq7 Mar 6 '13 at 17:29
  • This is an offline content cache server for schools in rural zambia. All the content is replaceable. It has to be cheap because schools are budget-limited and there will be a lot of them. But it also has to be reliable because it might be 8 hours each way on bad roads to replace. I'm not allowed to ask here what drives are not "ultra-cheap crap". I thought maybe a SECURITY ERASE might fix the bad sectors, but wanted opinions of others, or on what I might have missed in the SMART data. – qris Mar 6 '13 at 17:33
  • 3
    Cheap has different meanings. Most people think of cheap as low CapEx, which Kingston certainly is. But in "professional" situations TCO should be the measure of cheap. The Kingston drives are going to be somewhat more likely to fail that many other brands, and the drive/downtime OpEx to replace is certainly going to eat up any savings in CapEx. There are reviews around the net for reliable SSDs (I personally use SandForce 2000 or newer based products). – Chris S Mar 6 '13 at 17:56
  • Thanks Chris. I agree with your assessment of the opex of using cheap drives, but I'm having difficulty persuading the project owner. I did try to find those reliability surveys without much luck (no hard data yet). It seems from my research that this particular drive is an SSDNow V+200 which is a SF-2281 controller? So this drive would meet your criteria? – qris Mar 6 '13 at 20:07

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .