On 2011-06-14 at 15:47 -0400, Adam Levin wrote: > One of the most interesting aspects of putting together an implementation > of ZFS, to me, is figuring out what kind of infrastructure to build it on. > We're all used to "enterprise" class storage that has a lot of built-in > protection against hardware failures. ZFS mitigates a lot of that and > allows the use of much less expensive infrastructure, which I find > fascinating.
The latest ;login: magazine from USENIX had a good and thought-provoking article, "System Impacts of Storage Trends: Hard Errors and Testability" by Steven R. Hetzler; index & article: http://www.usenix.org/publications/login/2011-06/index.html http://www.usenix.org/publications/login/2011-06/pdfs/Hetzler.pdf The hard figures put on failure rates, combined with current hard-drive capacities and the impact on RAID made me realise that my mental model had become rather dated. The article itself is not about ZFS, but hard disk failure rates and SSD failure rates. My takeaway was that I should be looking at ZFS sooner rather than later and relying upon enterprise class disks + RAID without checksums in the FS layer is becoming increasingly foolhardy. -Phil _______________________________________________ Tech mailing list [email protected] https://lists.lopsa.org/cgi-bin/mailman/listinfo/tech This list provided by the League of Professional System Administrators http://lopsa.org/
