Re: 2.6.38.4: xfs speed problem?

[Dave Chinner]

On Sat, May 07, 2011 at 12:09:46PM -0400, Justin Piszcz wrote:
> Hello,
> 
> Using 2.6.38.4 on two hosts:
> 
> Host 1:
> $ /usr/bin/time find geocities.data 1> /dev/null
> 80.92user 417.93system 2:19:07elapsed 5%CPU (0avgtext+0avgdata 105520maxresident)k
> 0inputs+0outputs (0major+73373minor)pagefaults 0swaps
> 
> # xfs_db -c frag -f /dev/sda1
> actual 40203982, ideal 40088075, fragmentation factor 0.29%
> 
> meta-data=/dev/sda1              isize=256    agcount=44, agsize=268435455 blks
>          =                       sectsz=512   attr=2
> data     =                       bsize=4096   blocks=11718704640, imaxpct=5
>          =                       sunit=0      swidth=0 blks
> naming   =version 2              bsize=4096   ascii-ci=0
> log      =internal               bsize=4096   blocks=521728, version=2
>          =                       sectsz=512   sunit=0 blks, lazy-count=1
> realtime =none                   extsz=4096   blocks=0, rtextents=0
> 
> --
> 
> Host 2:
> $ /usr/bin/time find geocities.data 1>/dev/null
> 54.60user 337.20system 48:42.71elapsed 13%CPU (0avgtext+0avgdata 105632maxresident)k
> 0inputs+0outputs (1major+72981minor)pagefaults 0swaps
> 
> # xfs_db -c frag -f /dev/sdb1
> actual 37998306, ideal 37939331, fragmentation factor 0.16%
> 
> meta-data=/dev/sdb1              isize=256    agcount=10, agsize=268435455 blks
>          =                       sectsz=512   attr=2
> data     =                       bsize=4096   blocks=2441379328, imaxpct=5
>          =                       sunit=0      swidth=0 blks
> naming   =version 2              bsize=4096   ascii-ci=0
> log      =internal               bsize=4096   blocks=521728, version=2
>          =                       sectsz=512   sunit=0 blks, lazy-count=1
> realtime =none                   extsz=4096   blocks=0, rtextents=0
> 
> 
> --
> 
> Host 1: RAID-6 (7200 RPM Drives, 18+1 hot spare)

Those will be 3TB drives

> Host 2: RAID-6 (7200 RPM Drives, 12)

and those are 1TB drives.

Different hardware is guaranteed to give you different performance,
especially from a seek capability perspective.

> Each system uses a 3ware 9750-24i4e controller, same settings.
> 
> Any thoughts why one is > 2x faster than the other?

Different filesystem sizes mean different directory, inode and data
layouts, especially if you are using inode64.

Cheers,

Dave.
-- 
Dave Chinner
david@xxxxxxxxxxxxx
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/