Do you recommend Supermicro AOC-SAT2-MV8 as PCI-X SATA controller?

[Stephane LAPIE]

Hello Moritz,

On 04/05/2012 02:36 AM, Moritz Wilhelmy wrote:
> Hello Stephane,
>
> Thanks for the informative reply!
>
> On Thu, Apr 05, 2012 at 01:18:11 +0900, Stephane LAPIE wrote:
>> Right now I'm using this controller :
>>
>> mpt0 at pci0:6:1:0:    class=0x010000 card=0x10aa1734 chip=0x00541000
>> rev=0x01 hdr=0x00
>>     vendor     = 'LSI Logic / Symbios Logic'
>>     device     = 'SAS1068 PCI-X Fusion-MPT SAS'
>>     class      = mass storage
>>     subclass   = SCSI
>> mpt1 at pci0:6:7:0:    class=0x010000 card=0x10aa1734 chip=0x00541000
>> rev=0x01 hdr=0x00
>>     vendor     = 'LSI Logic / Symbios Logic'
>>     device     = 'SAS1068 PCI-X Fusion-MPT SAS'
>>     class      = mass storage
>>     subclass   = SCSI
>>
>> http://www.lsi.com/products/storagecomponents/Pages/LSISAS1068.aspx
>>
>> The only problem is, that card is an OEM version that was a bit tricky
>> to find.
>> I got it off an auction site in Japan, but I guess it covers pretty much
>> what you want,
>> in that it will advertise disks as physical volumes by default.
>
> I see. I'll try to get my hands on one, but anything that's hard to buy
> is also not an option (as always with universities, there is already
> enough bureaucracy involved). Unfortunately, this limits the choice of
> hardware :-(
I see :(
>> My only complaint on FreeBSD 9.X so far is that,
>>  it has trouble doing SMART probes on the disks using passthrough :)
>
> Uh, that's bad with that many disks... How do you work around it, if at
> all?
I didn't manage yet :( I'm pretty much stuck...
>
>> Performance is top notch, even though I had to tweak my /boot/loader.conf
>> to have a long SCSI timeout for disk detection (as you can imagine,
>> 15 disks will take a LONG time to detect.) :
>> kern.cam.scsi_delay=15000
>
> How's the performance of the AOC-SAT2-MV8 compared to this one?
I don't remember the exact figures. But given that the mvs(4) sometimes
hanged up (most of the time silently, really bad times would have drives
flap in and out), I say that mpt(4) comes out on top easy. Other than
that... My testing was done using all of my Gbit interfaces and trying
to max them all at once, and I could bust out something like 300-400MB/s
in pure transfer speeds using HTTP.

I'm not sure if the zpool iostat are to be trusted, but I reckon mvs(4)
would go up to 400MB/s while mpt(4) did go around 500-600MB/s... (take
that with a healthy grain of salt)

Cheers

-- 
Stephane LAPIE, EPITA SRS, Promo 2005
"Even when they have digital readouts, I can't understand them."
--MegaTokyo


-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 262 bytes
Desc: OpenPGP digital signature
Url : http://lists.freebsd.org/pipermail/freebsd-hardware/attachments/20120405/57038153/signature.pgp