Notes in-line Regards Jonathan Lewis http://www.jlcomp.demon.co.uk/faq/ind_faq.html The Co-operative Oracle Users' FAQ http://www.jlcomp.demon.co.uk/seminar.html Optimising Oracle Seminar - schedule updated Sept 19th ----- Original Message ----- From: "Matthew Zito" <mzito@xxxxxxxxxxx> To: <jonathan@xxxxxxxxxxxxxxxxxx> Cc: "Oracle-L" <oracle-l@xxxxxxxxxxxxx> Sent: Friday, December 17, 2004 2:04 PM Subject: Re: Storage array advice anyone? : : When I'm talking about virtualization, I am talking about using large : numbers of spindles to satisfy I/O, and in better examples, software on : the array to rebalance I/O. In that case, your argument should FIRST be: "You need a lot of spindles per application." The fact that you might have some cutesy bit of software that makes it easier to handle a lot of spindles is a minor detail. : : We can look at the other case - I could have a bunch of RAID-10 volumes : and any given read could span two drives. When I then have 64 I/Os : that are pending against that lone RAID-10 volume, wouldn't I be better : off having those I/Os against 50 disks instead of 14? : You might be ... but another way of looking at it is that if you're going to do a bad database design and write appalling code (that's a rhetorical "you", not a personal "you") then I'd rather you wrecked the I/O performance on your set of 14 discs than wreck my performance by hammering my set of 14 disks as well. : And as far as sharing disks, even EMC's RAID-1 implementation splits a : disk into a set of volumes and mirrors them to another drive. There's : always the possibility of spindle contention with any RAID group, : including a RAID-10 volume. Which is just one example of why virtualization can be seen as is a con-trick. The vendor "virtualizes" a small number of spindles into a large number of hypervolumes, then tells you that your LUN is striped across lots of hypervolumes - but forgets to mention that your LUN is striped, checkered, zig-zagged and self-contending on a small set of spindles. Then when you ask for the management tools to investigate your performance problem, they either don't exist, or cost a small fortune - "but you don't need to know because it's virtualized and striped and there can't be a problem." : How well an array copes with that is a : factor of workload, cache, and the elegance/functionality of the array : OS. Like everything else in the world, quality makes a difference. : That's why I suggested that you vet your vendors heavily. I thought that was a practice that died out in the late 18th century : : Thanks, : Matt : : -- : Matthew Zito : GridApp Systems : Email: mzito@xxxxxxxxxxx : Cell: 646-220-3551 : Phone: 212-358-8211 x 359 : http://www.gridapp.com : : : -- //www.freelists.org/webpage/oracle-l