On 11/27/12 1:52 AM, "Grégory Giannoni" <s...@narguile.org> wrote:

>Le 27 nov. 2012 à 01:17, Erik Trimble a écrit :
>> On 11/26/2012 12:54 PM, Grégory Giannoni wrote:
>>> [snip]
>>> I switched few month ago from Sun X45x0 to HP things : My fast NAS are
>>>now DL 180 G6. I got better perfs using LSI 9240-8I rather than HP
>>>SmartArray (tried P410 & P812). I'm using only 600Gb SSD drives.
>> That LSI controllers supports SATA III, or 6Gbps SATA.   The Px1x
>>controllers do 6GB SAS, but only 3GB SATA, so that's your likely perf
>>difference.  The SmartArray Px2x series should do both SATA and SAS at
>The SSD drives I'm using (Intel 320 600GB) are limited to 270MB/sec ; So
>I don't think that SATA II is limiting.
>> That said, I do think you're right that the LSI controller is probably
>>a better fit for connections requiring a SATA SSD.  The only exception
>>is having to give up the 1GB of NVRAM on the HP controller. :-(
>I don't think that this is a real issue when using a bunch of SSDs. I
>even wonder if the NVRAM is not slowing down writings. My tests were done
>with ZIL enabled, so a power loss shouldn't damage the datas.

HP recommends to disable the write accelerator on SSD-only volumes.

>>> [...]
>> Is the bottleneck the LSI controller, or the SAS/SATA bus, or the PCI-E
>>bus itself?  That is, have you tested with LSI 9240-4i  (one per 8-drive
>>cage, which I *believe* can use the HP multi-lane cable), and with a LSI
>>9260-16i or LSI 9280-24i?   My instinct would be to say it's the PCI-E
>>bus, and you could probably get away with the 4-channel cards.  i.e.
>>4-channels @ 6Gbit/s = 3 GBytes/s > 4x PCI-E 2.0 at 2GB/s
>The first bottleneck we reached (DL 180 / standard 25 drives bay) was the
>HP controller (both P410 AND P812 reached the same perfs : 800MB/sec
>writing, 1.3GB/sec reading).
>With LSI 9240-8I, we reached 1.2GB/s writing, 1.3Gb/s reading.
>The LSI 9240-4I was not able to connect to the 25-drives bay ; Not tested
> LSI 9260-16I or LSI 9280-24i.
>The results were the same with 10 or 25 drives, so I suspected either the
>PCI bus, either the expander in the 25-drives bay (HP 530946-001).
>Plugging the disks directly to the LSI card allowed to gain few MB/s :
>the expander was limiting a bit, but moreover, it disallowed to use more
>than 1 disk controller !
>By replacing the 25-drives bay by three 8-drives bays (507803-B21), the
>system was able to use 3 LSI 9240-8I, with this 4.4GB/sec reading rate.

That's correct that you've run into the limitation of the expander on the
25-disk drive backplane. However, I'm curious about the 8-drive cage you
mention. I use that cage in the ML/DL370 G6 servers. I didn't think it
would fit into a DL180 G6. How is this arranged in your unit? What does
the resulting setup look like? Sine the DL180 drive cages are part of the
bezel, do you just have three loose cages connected to the controllers?

Also, with three controllers, didn't you max the number of available PCIe

Anyway, the new HP SL4540 server is the next product worth testing in this
realmŠ 60 x LFF disks.

Edmund White

zfs-discuss mailing list

Reply via email to