On 04/08/2011 08:50 PM, Levi Pearson wrote: > On Fri, Apr 8, 2011 at 6:33 PM, Nicholas Leippe<n...@leippe.com> wrote: > >> If you're worried about longevity/write wearing, don't. If you need >> convincing why it's not an issue, see this article: >> >> http://www.storagesearch.com/ssdmyths-endurance.html > It may be unrelated to write wearing, but I just had an Mtron SSD that > was a few years old go south on me. It's not completely dead, but I > was using it as my boot drive and performance had become terrible and > eventually got to the point where it would just hang after a short > while. It is an old one, and I'm not exactly sure how old because I > got it secondhand, but I was disappointed that it didn't last very > long for me. > > I know SSDs have some pretty fancy firmware in them to manage the > actual flash storage, so it's possible that the firmware had just got > the storage into an untenable state that could be recovered through > some sort of reformat, but it is an old part that didn't particularly > impress me with its performance when I first installed it, so I didn't > consider it worth the effort of trying to save. > > Anyway, my point is that whether it be by write wearing or something > else, SSDs can kick the bucket unexpectedly too, and they don't appear > to my admittedly non-representative experience to be exceptionally > reliable. > > --Levi > > /* > PLUG: http://plug.org, #utah on irc.freenode.net > Unsubscribe: http://plug.org/mailman/options/plug > Don't fear the penguin. > */
In our experience in a 24x7 environment running RAID we are seeing SSD failures at 1 - 1 1/2 year on average. --Henry /* PLUG: http://plug.org, #utah on irc.freenode.net Unsubscribe: http://plug.org/mailman/options/plug Don't fear the penguin. */