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.

I've had a SSD for a little over a year on a sever. And once in a while, the server will bluescreen/be completely unresponsive. I've found out that it's the SSD going offline while the system is running because after rebooting (soft reset), the BIOS won't detect the SSD--unless I completely power off the system and turn it back on. Then the SSD is detected. I've swapped cables etc. What's the cause for this? Can it be a bad SSD? (Doesn't make sense that it would just "go offline"). Running Windows Server 2008...the logs don't tell me anything either.

This is a OCZ oynx, firmware is the latest. My HDTune results show it has a lot of bad sectors but I'm not sure if I trust the results.

Edit

HD Sentinel surface test shows about 19 bad sectors. Once I get my backup ssd (corsair), I'll reformat/reinitialize it to see if that fixes the issue.

share|improve this question
    
Useless anecdotal data: I have the exact same thing with an OCZ Agility 3 on my desktop. –  Kenny Rasschaert Feb 17 '12 at 18:50
    
How many bad sectors? HDTune won't lie, it'll be retrieving the relocated sector count from the SMART table on the drive. –  Daniel Lawson Feb 21 '12 at 19:33
add comment

2 Answers

Most likely causes:

  1. Buggy firmware. SSDs in particular are very "new" in relative terms, and the firmware doesn't have the decades of testing and evolution that the firmware in spinning rust drives has had. This is why so many people still recommend Intel SSDs versus other brands with better price/capacity or price/performance: the Intel SSDs have proven to have very reliable firmware, while most other brands have had disastrous (data loss) bugs. If you're running the latest release firmware, all you can do is open a support case with OCZ and provide them whatever diagnostic data they ask for.
  2. Buggy SATA/SAS driver: check for updates (you don't mention an OS, but I assume windows). Support for features like TRIM was quite spotty for a while.
  3. Buggy BIOS/raid firmware: make sure these are at the latest available from your manufacturer\
  4. Actual hardware problem with the drive, either in the controller, flash chips, or even on the PCB or connectors. All you can do is RMA the drive.
share|improve this answer
add comment

Sectors on an SSD aren't mapped to real sectors. When you write to a "sector", the firmware for the SSD is actually writing that data to a yet-unused part of the underlying SSD. It will always choose the part that's least been written to, to accomplish "wear levelling". That said, no disk utility should see bad sectors unless something's gone wrong. I'd recommend replacing the drive and seeing if that helps.

By the way: each SSD ships with some percentage more space than it actually claims to have when it's polled by the OS. This extra buffer space is used when real sectors start dying due to too many writes. This combined with the write levelling is why SSD manufacturers are claiming that their devices have the same if not a longer mean time between failure as a physical drive. If you have an unusually high write load, though, this may not be true.

share|improve this answer
add comment

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.