Re: Ever had an SSD fail?
Sure :-) We sell SSD's mostly as boot drives in "industrial process control" PC's. The endurance of an SSD depends greatly on how you configure your OS and apps. An SSD in read-only mode can last for ages. I have some firewalls booting Linux from read-only CF cards, running for almost a decade. Same thing for simple DOS-based systems that don't ever need to write to the drive (or scarcely). Same thing for Windows Embedded with EWF locked all the time.
But I also know cases where a SCADA app (configured to log data or keep a persistent image on disk) can thrash a decent 2.5" SSD in three months. Spinning rust still has its merits. Yes it can fail too - but it's not *prone* to fail in some deployments where SSD's *are* prone to fail pretty soon. And, in terms of spinning rust, you'd better shop for the *lowest* capacity currently available on the market = the simplest and proven construction, the lowest data density. The terabyte race is not a nice prospect in that context.
Ironically, most people still think that strictly nothing beats the endurance of an SSD in the role of a Windows boot drive... any SSD, in unmodified stock Windows, running Windows Update, an antivirus, a dozen self-updating apps etc.
You know - you install Windows on your shiny new expensive SSD, ohh the joy of how *fast* it is, then you go entertain yourself with something else... and a couple months down the road, when the SSD slows down noticeably, or fails outright, you tend to blame the piece of the SSD, or the early SSD model, or the brand... "Gosh, the SSD's were *crap* a year ago... must've been a bad batch or something... let me have a new one, that will surely last longer!" ;-)
It hardly comes across your mind that maybe the SSD thing is *principally* wrong for the position.