On Sat, 5 Apr 2003 05:18, Emile van Bergen wrote: > > It should be easy enough to implement with LVM or EVMS. Why not try it > > out and see what happens? > > I might do just that. If you'll help me devise some nice bonnie++ tests > for the benchmark :)
I suggest that you first do some tests with two contiguous partitions, and then create two interleaved partitions and do tests on one of them. If my guesses are correct then a simple Bonnie++ test with default options will provide results to convince you that such interleaving won't work. I expect that one of the interleaved partitions will show considerably less performance than the contiguous partition on the slowest part of the disk did. Also do tests with the -y option and two copies of bonnie++ running on different partitions at the same time. If I am incorrect then let me know and I'll help you devise other tests to determine the relative merits of interlaved vs contiguous partitions. One of the reasons I'm not particularly excited about this idea is that my impression that LVM and EVMS are not yet ready for serious wide-spread use. Even if LVM or EVMS in such a use does provide significantly better (say 50%) performance I would not be inclined to use it for that reason (I'd rather just pay more money for faster disks). -- http://www.coker.com.au/selinux/ My NSA Security Enhanced Linux packages http://www.coker.com.au/bonnie++/ Bonnie++ hard drive benchmark http://www.coker.com.au/postal/ Postal SMTP/POP benchmark http://www.coker.com.au/~russell/ My home page
