Re: [zfs-discuss] Filesystem Benchmark

2007-11-14 Thread Peter Tribble
On 11/14/07, Gary Wright <[EMAIL PROTECTED]> wrote: > > Hope you don't mind me asking but we are planning to use a CX3-20 Dell/EMC > SAN connected to a T5220 server (Solaris 10). Can you tell me > if you were forced to use PowerPath or have you used MPXIO/Traffic Manager. > Did you use LPe11000-E

Re: [zfs-discuss] Filesystem Benchmark

2007-11-14 Thread Gary Wright
Hi Cesare, Hope you don't mind me asking but we are planning to use a CX3-20 Dell/EMC SAN connected to a T5220 server (Solaris 10). Can you tell me if you were forced to use PowerPath or have you used MPXIO/Traffic Manager. Did you use LPe11000-E (Single Channel) or LPe11002-E (dual channel) HBA

Re: [zfs-discuss] Filesystem Benchmark

2007-05-09 Thread Selim Daoud
go ahead with filebench and don't forget to set set zfs:zfs_nocacheflush=1 in /etc/system (if using nevada) s. On 5/9/07, cesare VoltZ <[EMAIL PROTECTED]> wrote: Hy, I'm planning to test on pre-production data center a ZFS solution for our application and I'm searching a good filesystem benchm

Re: [zfs-discuss] Filesystem Benchmark

2007-05-09 Thread Ming Zhang
On Wed, 2007-05-09 at 21:09 +0200, Louwtjie Burger wrote: > > > LUN are configured as RAID5 accross 15 disks. > > Won't such a large amount of spindles have a negative impact on > performance (in a single RAID-5 setup) ... single I/O from system > generates lots of backend I/O's ? yes, single io

Re: [zfs-discuss] Filesystem Benchmark

2007-05-09 Thread Louwtjie Burger
> LUN are configured as RAID5 accross 15 disks. Won't such a large amount of spindles have a negative impact on performance (in a single RAID-5 setup) ... single I/O from system generates lots of backend I/O's ? ___ zfs-discuss mailing list zfs-discuss

Re: [zfs-discuss] Filesystem Benchmark

2007-05-09 Thread Ming Zhang
On Wed, 2007-05-09 at 16:27 +0200, cesare VoltZ wrote: > Hy, > > I'm planning to test on pre-production data center a ZFS solution for > our application and I'm searching a good filesystem benchmark for see > which configuration is the best solution. > > Server are Solaris 10 connected to a EMC C

Re: [zfs-discuss] Filesystem Benchmark

2007-05-09 Thread Richard Elling
cesare VoltZ wrote: Hy, I'm planning to test on pre-production data center a ZFS solution for our application and I'm searching a good filesystem benchmark for see which configuration is the best solution. Pedantically, your application is always best. -- richard _

Re: [zfs-discuss] Filesystem Benchmark

2007-05-09 Thread Rayson Ho
Tried filebench before?? http://www.solarisinternals.com/wiki/index.php/FileBench Rayson On 5/9/07, cesare VoltZ <[EMAIL PROTECTED]> wrote: I used in the past iozone (http://www.iozone.org/) but I'm wondering if there are other tools. Thanks. Cesare _

[zfs-discuss] Filesystem Benchmark

2007-05-09 Thread cesare VoltZ
Hy, I'm planning to test on pre-production data center a ZFS solution for our application and I'm searching a good filesystem benchmark for see which configuration is the best solution. Server are Solaris 10 connected to a EMC Clariion CX3-20 with two FC cable in a total high-availability (two H