Slow Write Performance on 3ware controller

Nikolas Britton nikolas.britton at gmail.com
Sun Jul 23 16:28:04 UTC 2006


On 7/23/06, Jonathan Horne <jhorne at dfwlp.org> wrote:
> On Sunday 23 July 2006 06:04, Richard Collyer wrote:
> > Nikolas Britton wrote:
> > >> Would moving the top card to the bottom slot and putting the 9500S-12 in
> > >> the top slot screw with FreeBSD in anyway or will it just accept the
> > >> changes and keep on going?
> > >
> > > I don't see any problem doing that... FreeBSD would only care if it
> > > was the boot drive AND if it also changed the device name in /dev. So
> > > yes... try it... if it changes the /dev name just edit /etc/fstab with
> > > the new settings.
> >
> > Oh yes, had fun with fstab in single user mode yesterday ... I'd put a
> > spelling mistake in /etc/fstab that was fun as it was my first venture
> > into single user mode. Took me 20 mins to realise that only / was
> > mounted and that /use wasn't hence no editors or shells.
> >
> > As this is a production server (its only my home file server) and the
> > card was an ebay special at sub $200 I'm happy with the performance.
> >
> > I may play around with moving them when some more routine maintenance
> > comes up but as I am only writing to it over the 100Mbit network
> > 50MB/sec is more than enough for what I am looking at.
> >
> > Many thanks for the help.
> >
> > Cheers
> > Richard
> >
>
> all this 3ware discussion inspired me to check out my 6.1-releng server and
> see how its 3ware card stacks up to the previously posted scores.
>
> [root at zeus ~]# diskinfo -t /dev/twed2
> /dev/twed2
>         512             # sectorsize
>         360099151872    # mediasize in bytes (335G)
>         703318656       # mediasize in sectors
>         43779           # Cylinders according to firmware.
>         255             # Heads according to firmware.
>         63              # Sectors according to firmware.
>
> Seek times:
>         Full stroke:      250 iter in   5.702448 sec =   22.810 msec
>         Half stroke:      250 iter in   3.484361 sec =   13.937 msec
>         Quarter stroke:   500 iter in   5.728894 sec =   11.458 msec
>         Short forward:    400 iter in   2.178793 sec =    5.447 msec
>         Short backward:   400 iter in   3.040917 sec =    7.602 msec
>         Seq outer:       2048 iter in   0.502034 sec =    0.245 msec
>         Seq inner:       2048 iter in   0.521798 sec =    0.255 msec
> Transfer rates:
>         outside:       102400 kbytes in   2.872447 sec =    35649 kbytes/sec
>         middle:        102400 kbytes in   2.996709 sec =    34171 kbytes/sec
>         inside:        102400 kbytes in   2.341439 sec =    43734 kbytes/sec
>
> ive been a 3ware fan for many years now.  this pic was taken in 2001...
> http://www.dfwlp.org/~jhorne/pics/computerroom/raid01.jpg
> heh, you want to talk about pci placement based on fit of ide cables... lol
> try mine.  in the end, after 5 years, im impressed mine is still chugging
> away, at albeit the performance numbers of yesteryear.  oh, incase anyone
> would ask, my /dev/twed2 is a 4x120GB RRAID5 with seagate's (ST3120026A), in
> a dual 1ghz thats almost as old as the card.

IBM Deathstars, not so fond memories of those drives. Here's another
diskinfo from an 8 disk Maxtor 7L250S0 array connected to a HighPoint
2220, It's days away from being decommissioned and then rebuilt into a
backup array for the new one taking it's place (Areca ARC-1220 + 8
Maxtor 7V300F0s):

> diskinfo -t /dev/da0
/dev/da0
        512             # sectorsize
        1756440297472   # mediasize in bytes (1.6T)
        3430547456      # mediasize in sectors
        213541          # Cylinders according to firmware.
        255             # Heads according to firmware.
        63              # Sectors according to firmware.

Seek times:
        Full stroke:      250 iter in   4.068691 sec =   16.275 msec
        Half stroke:      250 iter in   3.614864 sec =   14.459 msec
        Quarter stroke:   500 iter in   6.100577 sec =   12.201 msec
        Short forward:    400 iter in   2.314725 sec =    5.787 msec
        Short backward:   400 iter in   2.492332 sec =    6.231 msec
        Seq outer:       2048 iter in   0.251374 sec =    0.123 msec
        Seq inner:       2048 iter in   0.299685 sec =    0.146 msec
Transfer rates:
        outside:       102400 kbytes in   0.712265 sec =   143767 kbytes/sec
        middle:        102400 kbytes in   0.698637 sec =   146571 kbytes/sec
        inside:        102400 kbytes in   0.690232 sec =   148356 kbytes/sec


-- 
BSD Podcasts @:
http://bsdtalk.blogspot.com/
http://freebsdforall.blogspot.com/


More information about the freebsd-questions mailing list