New array stats

We've just installed the 2 new Clariions. We are in initial setup / burn phase,
which is a nice way of saying it only works on my Windows boxes so far, and
the EMC tech is busy/beavering, working on a Solaris box (based on the lack
of idle time on his terminal session) and the fact the box just rebooted.

Oh, and 1 Linux box. A Dual Xeon, not the Opterons like I wanted. Oh well.

But, on the other hand, this thing is FAST!

Here are the Bonnie++ tests to far:

Local mirrored disk to get a feel for baseline local disk:

\nVersion  1.03       ------Sequential Output------ --Sequential Input- --Random-\n                    -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--\nMachine        Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP  /sec %CP\nflat             2G  3621  96 26504  15 12258   4 13283  65 30537   5 454.2   0\n\n\nEMC Array:\nExt3:\n\nVersion  1.03       ------Sequential Output------ --Sequential Input- --Random-\n                    -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--\nMachine        Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP  /sec %CP\nflat             2G 34080  98 143791  89 36309  15 31671  71 105523  25  1522   3\n\n\nExt2:\n\nVersion  1.03       ------Sequential Output------ --Sequential Input- --Random-\n                    -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--\nMachine        Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP  /sec %CP\nflat             2G 40038  97 277351  61 36711  11 29625  70 99576  20  3090   6\n\n

When I go to XFS on Suse, I expect to get about 20% better than Ext2
performance.

277MB per second write, which is obviously based on cache acks, but that
is fine since I'll never stream that much. It also means that single files
writes are being load balanced across the dual port 2GB fibre channel card.

99MB per second read is very good, expecially when a typical database will be
reading multiple luns (and yes, my tests showed they scaled), and then finally,
the 3090 seeks per second.

These numbers are between 3 and 20 times better than my current system.

I think I'll keep it.

Post #178,564 by broomberg 10/8/04 11:46:14 PM Reply	New array stats We've just installed the 2 new Clariions. We are in initial setup / burn phase, which is a nice way of saying it only works on my Windows boxes so far, and the EMC tech is busy/beavering, working on a Solaris box (based on the lack of idle time on his terminal session) and the fact the box just rebooted. Oh, and 1 Linux box. A Dual Xeon, not the Opterons like I wanted. Oh well. But, on the other hand, this thing is FAST! Here are the Bonnie++ tests to far: Local mirrored disk to get a feel for baseline local disk: \nVersion 1.03 ------Sequential Output------ --Sequential Input- --Random-\n -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--\nMachine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP\nflat 2G 3621 96 26504 15 12258 4 13283 65 30537 5 454.2 0\n\n\nEMC Array:\nExt3:\n\nVersion 1.03 ------Sequential Output------ --Sequential Input- --Random-\n -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--\nMachine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP\nflat 2G 34080 98 143791 89 36309 15 31671 71 105523 25 1522 3\n\n\nExt2:\n\nVersion 1.03 ------Sequential Output------ --Sequential Input- --Random-\n -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--\nMachine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP\nflat 2G 40038 97 277351 61 36711 11 29625 70 99576 20 3090 6\n\n When I go to XFS on Suse, I expect to get about 20% better than Ext2 performance. 277MB per second write, which is obviously based on cache acks, but that is fine since I'll never stream that much. It also means that single files writes are being load balanced across the dual port 2GB fibre channel card. 99MB per second read is very good, expecially when a typical database will be reading multiple luns (and yes, my tests showed they scaled), and then finally, the 3090 seeks per second. These numbers are between 3 and 20 times better than my current system. I think I'll keep it.
Post #178,579 by folkert 10/9/04 12:15:09 AM Reply	And it is cheaper too ?! -- [link\|mailto:greg@gregfolkert.net\|greg], [link\|http://www.iwethey.org/ed_curry\|REMEMBER ED CURRY!] @ *i_wethey* No matter how much Microsoft supporters whine about how Linux and other operating systems have just as many bugs as their operating systems do, the bottom line is that the serious, gut-wrenching problems happen on Windows, not on Linux, not on Mac OS. -- [link\|http://www.eweek.com/article2/0,1759,1622086,00.asp\|source] Here is an example: [link\|http://www.greymagic.com/security/advisories/gm001-ie/\|Executing arbitrary commands without Active Scripting or ActiveX when using Windows]
Post #178,584 by broomberg 10/9/04 12:32:42 AM Reply	Oh yeah Last array was bargain basement noname about $50K per TB. This one is brand name about $12K per TB.
Post #178,596 by deSitter 10/9/04 9:24:37 AM Reply	You are the performance master (bow) -drl

Welcome to IWETHEY!