RE: [OT] 8-port AHCI SATA Controller?
- From: "Roger Heflin" <rheflin@xxxxxxxxx>
- Date: Tue, 31 Jan 2006 14:48:41 -0600
> -----Original Message-----
> From: linux-kernel-owner@xxxxxxxxxxxxxxx
> [mailto:linux-kernel-owner@xxxxxxxxxxxxxxx] On Behalf Of
> Lennart Sorensen
> Sent: Tuesday, January 31, 2006 12:58 PM
> To: Sander
> Cc: linux-kernel@xxxxxxxxxxxxxxx; jgarzik@xxxxxxxxx
> Subject: Re: [OT] 8-port AHCI SATA Controller?
>
> On Tue, Jan 31, 2006 at 07:50:07PM +0100, Sander wrote:
> > Actually, I need 24 ports :-) But 3x SX8 sets me back 540 dollars
> > according to pricewatch, which is less than half.
>
> I know with older promise controllers, it wasn't possible to
> run more than 2 in one system as far as I remember due to
> some dma issues. Not sure if that applies to the SX8.
>
> If it turns out the SX8 has issues (like the one pointed out
> earlier about number of commands to the card at once) or that
> it can't have 3 cards in one system at once, then what? Are
> you then out $540 + the cost of a better controller?
> Certainly worth finding out before spending the money.
>
> > Fakeraid controllers are less expensive, and would do too of course
> > :-)
>
> Of course those aren't hardware, and are only meant for small
> toy raids for windows users. The rest of use treat them as
> ide/sata controllers only. I haven't seen one of those with
> more than 4 ports either. If the SX8 is one, then I must
> admit I haven't looked at it before. I try to avoid hardware
> from promise whenever possible.
>
Highpoint has some that I believe are software raidish.
They do have on-board parity generators that are used when
you use there binary only modules.
I have heard that they will work with later kernels (2.6.15+)
since the highpoint are a standard Marvell chipset, and they
seem to be fairly price competitive with JBOD raid controllers,
and have some controllers that have more than 8 ports, the price
per port may be better on the larger controllers.
Using 3 disk software stripe (linux) or 3 disks software stripe
(binary modules) I have got IO rates of 135MB/second sustained
over 90-100 GB read/write tests.
Roger
-
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/
- References:
- Re: [OT] 8-port AHCI SATA Controller?
- From: Lennart Sorensen
- Re: [OT] 8-port AHCI SATA Controller?
- Prev by Date: Re: vgacon scrolling problem [Was: Re: 2.6.16-rc1-mm4]
- Next by Date: Re: GPL V3 and Linux - Dead Copyright Holders
- Previous by thread: Re: [OT] 8-port AHCI SATA Controller?
- Next by thread: Re: [OT] 8-port AHCI SATA Controller?
- Index(es):
Relevant Pages
- Re: [OT] 8-port AHCI SATA Controller?
... I know with older promise controllers, it wasn't possible to run more ... If
it turns out the SX8 has issues (like the one pointed out earlier ... to avoid hardware
from promise whenever possible. ... (Linux-Kernel) - Re: [OT] 8-port AHCI SATA Controller?
... > I know with older promise controllers, it wasn't possible to run more ...
> sure if that applies to the SX8. ... >> Fakeraid controllers are less
expensive, and would do too of course :-) ... send the line "unsubscribe linux-kernel"
in ... (Linux-Kernel) - 5.1 -> 5.2 upgrade, Promise FastTrak device /dev/ar0 missing?
... The /dev/ar0 device shows up on the 5.1 kernel. ... Pentium Pro MTRR support
enabled ... 2 ports with 2 removable, ... # RAID controllers interfaced
to the SCSI subsystem ... (freebsd-hackers) - 5.1 -> 5.2 upgrade, Promise FastTrak device /dev/ar0 missing?
... The /dev/ar0 device shows up on the 5.1 kernel. ... Pentium Pro MTRR support
enabled ... 2 ports with 2 removable, ... # RAID controllers interfaced
to the SCSI subsystem ... (freebsd-hackers) - Re: USB host controllers
... > they'll appear on their respective Host Controllers Advanced ... > My
Intel 865 motherboard has 8 Ports and in Device Manager ... of my USB devices across
ports. ... USB2 TV tuner, whose power requirements exceed the single port "limit" ...
(microsoft.public.windowsxp.general)