Re: [PERFORM] RAID stripe size question

[Alex Turner]

This is a great testament to the fact that very often software RAID will seriously outperform hardware RAID because the OS guys who implemented it took the time to do it right, as compared with some controller manufacturers who seem to think it's okay to provided sub-standard performance.

Based on the bonnie++ numbers comming back from your array, I would also encourage you to evaluate software RAID, as you might see significantly better performance as a result.  RAID 10 is also a good candidate as it's not so heavy on the cache and CPU as RAID 5.

Alex.

On 7/18/06, Luke Lonergan <[EMAIL PROTECTED]> wrote:

Mikael,

On 7/18/06 6:34 AM, "Mikael Carneholm" < [EMAIL PROTECTED]> wrote:

> However, what's more important is the seeks/s - ~530/s on a 28 disk
> array is quite lousy compared to the 1400/s on a 12 x 15Kdisk array

I'm getting 2500 seeks/second on a 36 disk SATA software RAID (ZFS, Solaris 10) on a Sun X4500:

=========== Single Stream ============

With a very recent update to the zfs module that improves I/O scheduling and prefetching, I get the following bonnie++ 1.03a results with a 36 drive RAID10, Solaris 10 U2 on an X4500 with 500GB Hitachi drives (zfs checksumming is off):

Version  1.03       ------Sequential Output------    --Sequential Input-   --Random-
                    -Per Chr-  --Block--  -Rewrite-  -Per Chr-  --Block--  --Seeks--
Machine        Size K/sec  %CP K/sec  %CP K/sec  %CP K/sec  %CP K/sec  %CP /sec %CP

thumperdw-i-1   32G 120453  99 467814  98 290391  58 109371  99 993344  94 1801   4

                    ------Sequential Create------ --------Random Create--------
                    -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
              files  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP

                 16 +++++ +++ +++++ +++ +++++ +++ 30850  99 +++++ +++ +++++ +++

=========== Two Streams ============

Bumping up the number of concurrent processes to 2, we get about 1.5x speed reads of RAID10 with a concurrent workload (you have to add the rates together):

Version  1.03       ------Sequential Output------   --Sequential Input-     --Random-
                    -Per Chr- --Block--  -Rewrite-  -Per Chr-  --Block--    --Seeks--
Machine        Size K/sec  %CP K/sec  %CP K/sec  %CP K/sec  %CP K/sec  %CP  /sec %CP

thumperdw-i-1   32G 111441  95 212536  54 171798  51 106184  98 719472  88  1233   2

                    ------Sequential Create------ --------Random Create--------
                    -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
              files  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP

                 16 26085  90 +++++ +++  5700  98 21448  97 +++++ +++  4381  97

Version  1.03       ------Sequential Output------   --Sequential Input-     --Random-
                    -Per Chr-  --Block--  -Rewrite-  -Per Chr-  --Block--   --Seeks--
Machine        Size K/sec  %CP K/sec  %CP K/sec  %CP K/sec  %CP K/sec  %CP  /sec %CP

thumperdw-i-1   32G 116355  99 212509  54 171647  50 106112  98 715030  87  1274   3

                    ------Sequential Create------ --------Random Create--------
                    -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
              files  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP

                 16 26082  99 +++++ +++  5588  98 21399  88 +++++ +++  4272  97

So that's 2500 seeks per second, 1440MB/s sequential block read, 212MB/s per character sequential read.
=======================

- Luke