On Wed, Jan 20, 2010 at 10:25 AM, seth <ninja.c...@gmail.com> wrote: > Hel, > > hopefully a simple question. Our current network is NFS, with 8 frontends > that are read only to an NFS share. There is a single internal CMS server > that has write access to this share. This share is getting over worked so I > am moving it to a new backend. The original plan was just to replicate what > we have on faster disks, however reading I see that for the many small reads > we do (web frontends) on the really messed up file layout we have (some > directories have over 30k files in them) that iscsi could give us a > significant performance increase. > > (http://technomagesinc.com/papers/ip_paper.html)
This paper appears really old. Have you instrumented your own workload to see where the bottleneck actually lies? Disabling atime and adding a high-speed log device? You may find a load of things easier to do than to rebuild your environment using shared block devices...
_______________________________________________ storage-discuss mailing list storage-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/storage-discuss
