The whole "SSD fails after X number of writes" thing is pretty much a myth now that most drives implement pretty aggressive write leveling. These modern drives introduce a mapping layer between the physical disk locations and the location written to by the OS. This means that writing to "KB 4242" on the disk actually is written to a different part of the disk each time it is written.
Some recent tests report about 1PB of writes before failure ( http://techreport.com/review/27436/the-ssd-endurance-experiment-two-freaking-petabytes). Let's say your caching system generated and trashed 20GB of data a day, that would leave you with about 130 years before the drive failed. So yes, it's limited but not very limited. On Fri, Mar 6, 2015 at 3:53 PM, Luc Prefontaine <lprefonta...@softaddicts.ca > wrote: > We have been running builds on the same SSDs, doing intensive logging, ... > for three years now. > > None deteriorated. > > Builds are mainly scrap & write thousands of small files plus a few big > ones (the targets). > > Write speed makes a huge difference for this kind of task. > > Aws allows to get VMs with SSDs of decent sizes and not to store only the > OS. > > I think they would restrict usage if > rewrites were a huge concern... > > Luc P. > > > On Friday, March 6, 2015 at 3:16:09 PM UTC-5, Michael Blume wrote: > > > > > > Possibly stupid question: can you just pretend you have more memory > than > > > you do and let the operating system do the heavy lifting? > > > > > As in, put the swap partition on the SSD and jack up the virtual memory > in > > the OS config? > > > > Isn't it a bad idea to put swap on an SSD, because of the limited number > of > > times an SSD byte can be rewritten before it sticks permanently? Thus > > making SSD more suited to storing stuff where speed counts, but which > > doesn't change very much, like the operating system kernel and modules > and > > core libraries, plus your most often used applications? Then you get > faster > > boot and app-startup times without constant writes to the SSD (just when > > something is upgraded). > > > > Of course, SSD being less well suited to frequently-written data would > also > > militate against using it for a cache managed by the application, rather > > than by the OS ... > > > > > > -- > > You received this message because you are subscribed to the Google > > Groups "Clojure" group. > > To post to this group, send email to clojure@googlegroups.com > > Note that posts from new members are moderated - please be patient with > your first post. > > To unsubscribe from this group, send email to > > clojure+unsubscr...@googlegroups.com > > For more options, visit this group at > > http://groups.google.com/group/clojure?hl=en > > --- > > You received this message because you are subscribed to the Google > Groups "Clojure" group. > > To unsubscribe from this group and stop receiving emails from it, send > an email to clojure+unsubscr...@googlegroups.com. > > For more options, visit https://groups.google.com/d/optout. > > > -- > Luc Prefontaine<lprefonta...@softaddicts.ca> sent by ibisMail! > > -- > You received this message because you are subscribed to the Google > Groups "Clojure" group. > To post to this group, send email to clojure@googlegroups.com > Note that posts from new members are moderated - please be patient with > your first post. > To unsubscribe from this group, send email to > clojure+unsubscr...@googlegroups.com > For more options, visit this group at > http://groups.google.com/group/clojure?hl=en > --- > You received this message because you are subscribed to the Google Groups > "Clojure" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to clojure+unsubscr...@googlegroups.com. > For more options, visit https://groups.google.com/d/optout. > -- “One of the main causes of the fall of the Roman Empire was that–lacking zero–they had no way to indicate successful termination of their C programs.” (Robert Firth) -- You received this message because you are subscribed to the Google Groups "Clojure" group. To post to this group, send email to clojure@googlegroups.com Note that posts from new members are moderated - please be patient with your first post. To unsubscribe from this group, send email to clojure+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/clojure?hl=en --- You received this message because you are subscribed to the Google Groups "Clojure" group. To unsubscribe from this group and stop receiving emails from it, send an email to clojure+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
