Marvel AOC-SASLP-MV8 - 8 Sata PCI-E 4x non-raid controller card

Now for my own question.

What is INT13?

I remember people talking about this earlier in the thread, but do I need to change something on here to be able to boot from the controller for WHS?
 
Now for my own question.

What is INT13?

I remember people talking about this earlier in the thread, but do I need to change something on here to be able to boot from the controller for WHS?

If you want to boot from a drive attached to the SASLP card, you need to leave INT13 enabled. If you want to boot from a drive attached to a mobo sata port, you need to disable INT13.
 
Ok cool.

Yea I figured i just try it.
I have 2008 booting off the controller now, and I am in the process of installing my Virtual WHS.

One concern I did have though is earlier I pulled a drive and then added a new and the OS froze and I had to reboot.

Anyone experienced anything like that? HotSwap capability is pretty important to me. I dont have the onboard SATA controller enabled but I dont think setting AHCI or IDE in the mobo's BIOS would even matter.

It didnt for my SAT2-MV8 at least.
 
Hi guys!

Have done a clean install of Ubuntu Kermic 2.6.31-kernel and my RAID5 array is working... at 25MB/s read with hdparm.

I've just installed the svn driver from http://sourceforge.net/projects/scst/ , so I suppose the cause is somewhere else than in the drivers. Any suggestions? Could it be some settings with the array for instance? I've set it up with mdadm (sw Ubuntu raid)

I've got:

2,0GHz Opteron s939
DFI LANPARTY SLI-DR
2GB RAM

Here's my lsmod output:
libsas 56456 1 mvsas
scsi_transport_sas 35488 2 mvsas,libsas

I've got the Marvell Technology Group Ltd. MV64460/64461/64462 System Controller, Revision B (rev 01).

Though, I didn't get to compile the driver properly, got some Warning errors with stops:

make -C /lib/modules/2.6.31-14-server/build SUBDIRS=/root/sas/scst/trunk/mvsas_tgt BUILD_INI=m
make[1]: Entering directory `/usr/src/linux-headers-2.6.31-14-server'
Building modules, stage 2.
MODPOST 1 modules
WARNING: "scst_rx_data" [/root/sas/scst/trunk/mvsas_tgt/mvsas.ko] undefined!
WARNING: "scst_unregister_session" [/root/sas/scst/trunk/mvsas_tgt/mvsas.ko] undefined!
WARNING: "scst_rx_mgmt_fn" [/root/sas/scst/trunk/mvsas_tgt/mvsas.ko] undefined!
WARNING: "scst_register_session" [/root/sas/scst/trunk/mvsas_tgt/mvsas.ko] undefined!
WARNING: "scst_aen_done" [/root/sas/scst/trunk/mvsas_tgt/mvsas.ko] undefined!
WARNING: "scst_tgt_cmd_done" [/root/sas/scst/trunk/mvsas_tgt/mvsas.ko] undefined!
WARNING: "scst_cmd_init_done" [/root/sas/scst/trunk/mvsas_tgt/mvsas.ko] undefined!
WARNING: "scst_rx_cmd" [/root/sas/scst/trunk/mvsas_tgt/mvsas.ko] undefined!
WARNING: "__scst_register_target_template" [/root/sas/scst/trunk/mvsas_tgt/mvsas.ko] undefined!
WARNING: "scst_unregister_target_template" [/root/sas/scst/trunk/mvsas_tgt/mvsas.ko] undefined!
WARNING: "scst_to_tgt_dma_dir" [/root/sas/scst/trunk/mvsas_tgt/mvsas.ko] undefined!
WARNING: "scst_register" [/root/sas/scst/trunk/mvsas_tgt/mvsas.ko] undefined!
WARNING: "scst_unregister" [/root/sas/scst/trunk/mvsas_tgt/mvsas.ko] undefined!
make[1]: Leaving directory `/usr/src/linux-headers-2.6.31-14-server'



Any suggestions? I didnt get som much out of the README file in this dir.
 
Try the dd test:

Code:
dd if=/dev/zero of=test.file bs=1M count=1K; sync; date

hdparm doesn't really tell much.

Those warning are ok, don't worry about them.
 
Thanks for the reply! :)

The output seems much better, though I experience only 25MB/s read-speeds with Samba file share. I though it may be caused by the disks since hdparm shows the same read results.

I previously set up the array with a Silicon 3114 controller, then I experienced up to 37MB/s read with hdparm and Samba. It's now worse on the new controller, in which I have hard to believe it should be.

The dd test seemed to give much better output, though I was hoping for more..?

1024+0 records in
1024+0 records out
1073741824 bytes (1,1 GB) copied, 7,7276 s, 139 MB/s
on. 11. nov. 13:26:27 +0100 2009



Could it be some other settings in the system causing this? The RAID array settings maybe?
 
whats the output from /sys/block/mdX/md/stripe_cache_size and /sys/block/mdX/bdi/read_ahead_kb ? Also whats your chunk size? (/proc/mdstat will tell you that)
 
stripe_cache_size :256
read_ahead_kb: 128

md1 : active raid5 sdc1[0] sde1[1] sdd1[2]
1953519872 blocks level 5, 4k chunk, algorithm 2 [3/3] [UUU]

Some bad values? :)
 
4K chunk on what size drives? Sounds WAY too low.

do this and retest using bonnie++:

echo "4096" > /sys/block/mdX/md/stripe_cache_size
echo "8192" > /sys/block/mdX/bdi/read_ahead_kb

bonnie++ command:

sync; bonnie++ -n 0 -f -b
 
Ran "sync; bonnie++ -n 0 -f -b"

You must use the "-u" switch when running as root.
usage: bonnie++ [-d scratch-dir] [-s size(Mb)[:chunk-size(b)]]
[-n number-to-stat[:max-size[:min-size][:num-directories]]]
[-m machine-name]
[-r ram-size-in-Mb]
[-x number-of-tests] [-u uid-to-use:gid-to-use] [-g gid-to-use]
[-q] [-f] [-b] [-p processes | -y]

Version: 1.03c


U got some switch wrong there..

I'm not taking any chances myself...
 
-u <username with access to the array>

what i run on my box: sync; bonnie++ -u wibla -n 0 -f -b
 
Here's some output :) Thanks for the help so far!

------Sequential Output------ --Sequential Input- --Random-
-Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP
myhost 4G 54637 10 23462 4 64761 4 116.9 0
Myhosy.no ,4G,,,54637,10,23462,4,,,64761,4,116.9,0,,,,,,,,,,,,,
 
I follow the thread quite a while, and since I plan to upgrade my controllers (throw them away, 3 of them) and replace them with the supermicro aoc-saslp-mv8, I was just wondering what kind of SFF 8087 breakout cables should I buy.
There are two versions available, straight ones and reverse. What's the difference.
My server case is a Lian li PC-A77.
 
anybody out there?
Anyway, queries continue.
My server board is an Asus P5Q. It features 1 PCI-E 2.0 16x (graphics card), and 2 PCI-E 1x.
The aforementioned controller is a PCI-E x4 if I am not mistaken.
Would that be a problem, in terms of compatibility and speed issues? Does the add-on card fit in a pci-e x1 slot?
Your help is greatly appreciated.


Andreas
 
anybody out there?
Anyway, queries continue.
My server board is an Asus P5Q. It features 1 PCI-E 2.0 16x (graphics card), and 2 PCI-E 1x.
The aforementioned controller is a PCI-E x4 if I am not mistaken.
Would that be a problem, in terms of compatibility and speed issues? Does the add-on card fit in a pci-e x1 slot?
Your help is greatly appreciated.

In terms of compatiblity, no issue whatsoever as long as the PCI-E x1 slot has an open front.. All PCI-E cards are designed to work in slower slots due to it being specified in the PCI-E design specs. However, the PCI-E x1 slots on the Asus P5Q, like most other motherboards, have a closed front.

So outside of modding or cutting open that PCi-E x1 slot, no that SATA controller will not fit into a PCI-E x1 slot. People do cut open the PCI-E x1 slot BTW:
http://www.overclock.net/overclock-net-folding-home-team/439190-great-news-pci-e-x1-folding.html
 
Can I fit the controller into the Pci-e 16x slot (graphics card slot) since no graphics card is needed for Whs.
Would I, in that case, get full potential (speed) of the controller?
 
Can I fit the controller into the Pci-e 16x slot (graphics card slot) since no graphics card is needed for Whs.
Would I, in that case, get full potential (speed) of the controller?

I think it depends on the mobo. I have two of the SASLP cards working fine in the x16/x8 slots in a Gigabyte MA790GP-UD4H mobo.
 
Can I fit the controller into the Pci-e 16x slot (graphics card slot) since no graphics card is needed for Whs.
Would I, in that case, get full potential (speed) of the controller?

Yes it will work.

You can run the card in any PCIe slot, as they are backwords and forwards compatible by design.
 
So i just read 342 posts.

You guys talk to much lol

But my question was answered on page 17 (serves me right for reading the whole frigin thread!!)

Will work with windows 7 ultimate :D. I am in the process of deciding to use whs, orrrr windows 7. still a toss up. but it works for both so I think I can just decide as I go once I have all my gear :)

to bad a fair priced 16 port never came out. 419 for 16 ports or under a 100 for 8 ports.. cheaper to build a whole new pc for a second card rather than buy the 16port. lol (or I think i will buy a 3 pci-e board)

Has anyone found a reliable board offering 4 pci-e with onboard hdmi? I may not need the hdmi but would like the availibility if needed. I did see the 3 slot board someone psoted earlier. and may go with that for my choice.

Thanks for all the info guys


Shawn
 
So I have attached the controller to my Debian sid system, build a linux software RAID-1 array and am running dd to see whether I get any errors for a while...

Try the dd test:

Code:
dd if=/dev/zero of=test.file bs=1M count=1K; sync; date

hdparm doesn't really tell much.

Those warning are ok, don't worry about them.

Test on my IBM 7k1000.Bs
Code:
dd if=/dev/zero of=/dev/md5 bs=1M count=1K; sync; date
1024+0 records in
1024+0 records out
1073741824 bytes (1.1 GB) copied, 11.7771 s, 91.2 MB/s
Tue Nov 17 02:58:19 CET 2009
BackServ:/home/gwk# dd if=/dev/zero of=/dev/md5 bs=1M count=1K; sync; date
1024+0 records in
1024+0 records out
1073741824 bytes (1.1 GB) copied, 11.5374 s, 93.1 MB/s
Tue Nov 17 02:59:20 CET 2009
 
so there is a new, interesting development on this issue: my card does not work either. I get a ton of mvs_abort_task errors and apparently, I am not alone

Code:
[  972.820024] /build/buildd/linux-2.6-2.6.31/debian/build/source_amd64_none/drivers/scsi/mvsas/mv_sas.c 1669:mvs_abort_task:rc= 5

If anyone has a solution on how to use this card in Linux for a raid array, let em know :)

Hi guys!

Have done a clean install of Ubuntu Kermic 2.6.31-kernel and my RAID5 array is working... at 25MB/s read with hdparm.

I've just installed the svn driver from http://sourceforge.net/projects/scst/ , so I suppose the cause is somewhere else than in the drivers. Any suggestions? Could it be some settings with the array for instance? I've set it up with mdadm (sw Ubuntu raid)

I've got:

2,0GHz Opteron s939
DFI LANPARTY SLI-DR
2GB RAM

Here's my lsmod output:
libsas 56456 1 mvsas
scsi_transport_sas 35488 2 mvsas,libsas

I've got the Marvell Technology Group Ltd. MV64460/64461/64462 System Controller, Revision B (rev 01).

Apparently my 2.6.31-1-amd64 Kernel has a different module?
Code:
HamCore:/home/gwk# uname -r
2.6.31-1-amd64

HamCore:/home/gwk# lsmod|grep sas
mvsas                  48224  0
libsas                 50536  1 mvsas
scsi_transport_sas     33936  2 mvsas,libsas
libata                190940  4 ata_generic,libsas,pata_jmicron,ahci
scsi_mod              172992  5 sd_mod,mvsas,libsas,scsi_transport_sas,libata

HamCore:/home/gwk# lspci| grep -i marvell
02:00.0 SCSI storage controller: Marvell Technology Group Ltd. MV64460/64461/64462 System Controller, Revision B (rev 01)

Looks like there is light at the end of the tunnel in the next kernel:
http://bugzilla.kernel.org/show_bug.cgi?id=14534
 
Last edited:
so there is a new, interesting development on this issue: my card does not work either. I get a ton of mvs_abort_task errors and apparently, I am not alone

Code:
[  972.820024] /build/buildd/linux-2.6-2.6.31/debian/build/source_amd64_none/drivers/scsi/mvsas/mv_sas.c 1669:mvs_abort_task:rc= 5

If anyone has a solution on how to use this card in Linux for a raid array, let em know :)



Apparently my 2.6.31-1-amd64 Kernel has a different module?
Code:
HamCore:/home/gwk# uname -r
2.6.31-1-amd64

HamCore:/home/gwk# lsmod|grep sas
mvsas                  48224  0
libsas                 50536  1 mvsas
scsi_transport_sas     33936  2 mvsas,libsas
libata                190940  4 ata_generic,libsas,pata_jmicron,ahci
scsi_mod              172992  5 sd_mod,mvsas,libsas,scsi_transport_sas,libata

HamCore:/home/gwk# lspci| grep -i marvell
02:00.0 SCSI storage controller: Marvell Technology Group Ltd. MV64460/64461/64462 System Controller, Revision B (rev 01)

Looks like there is light at the end of the tunnel in the next kernel:
http://bugzilla.kernel.org/show_bug.cgi?id=14534

I have the same exact problem. And you are right, we are not alone. I just bought 4 new 1.5tb Samsungs and this card to run in RAID5 or RAID6 and now realized everyone is having these problems. This is very frustrating.
 
I've configured this card in a Gigabyte GA-EP45-DS5.

Despite what Gigabyte say, if you have a PCI-e video card in the first PCI-e x16 slot then the second PCI-e x8 slot isn't reserved just for Crossfire switching but can support a RAID card such as the AOC-SASLP-MV8. But I have had to retain BIOS F7 inorder to eanble the GSATA secondary satat 2 ports as valid disk extensions (not just RADI1 backup). I can get 8 disks off the motherboard.

I require this card to support 8 x SATA 1.5TB disks in JBOD.

My problem is when I boot the computer the card BIOS is processed, I can change the card settings, but I don't see any physical disks.

When WHS starts the card shows 8 amber lights but WHS still does not see any of the disks which is not unexpected.

I am using 2 SFF-8087 'straight' cables with 2 x 4 SATA II connectors. This is my second set of cables.

Two questions:

1. do I need special Supermicro IPASS cables (CBL-SFF8087OCF-05M) or will any SFF-8087 breakout cable suffice.

2. I could alternatively purchase the older generation Promise SuperTrak EX8350 PCI-E 8 SATA RAID Controller but I beleive there will be a marked drop in performance.
 
I've configured this card in a Gigabyte GA-EP45-DS5.

Despite what Gigabyte say, if you have a PCI-e video card in the first PCI-e x16 slot then the second PCI-e x8 slot isn't reserved just for Crossfire switching but can support a RAID card such as the AOC-SASLP-MV8. But I have had to retain BIOS F7 inorder to eanble the GSATA secondary satat 2 ports as valid disk extensions (not just RADI1 backup). I can get 8 disks off the motherboard.

I require this card to support 8 x SATA 1.5TB disks in JBOD.

My problem is when I boot the computer the card BIOS is processed, I can change the card settings, but I don't see any physical disks.

When WHS starts the card shows 8 amber lights but WHS still does not see any of the disks which is not unexpected.

I am using 2 SFF-8087 'straight' cables with 2 x 4 SATA II connectors. This is my second set of cables.

Two questions:

1. do I need special Supermicro IPASS cables (CBL-SFF8087OCF-05M) or will any SFF-8087 breakout cable suffice.

2. I could alternatively purchase the older generation Promise SuperTrak EX8350 PCI-E 8 SATA RAID Controller but I beleive there will be a marked drop in performance.

Are you connecting straight to the disk or are you using some sort of backplane? If youre using a backplane then take a disk out and connect it directly.

Are you sure the card is working in that slot?
Try a different one.
 
Swap the GPU and SASLP

Really its going to be the only way to find out if you have a bad card.
 
I've configured this card in a Gigabyte GA-EP45-DS5.

Despite what Gigabyte say, if you have a PCI-e video card in the first PCI-e x16 slot then the second PCI-e x8 slot isn't reserved just for Crossfire switching but can support a RAID card such as the AOC-SASLP-MV8. But I have had to retain BIOS F7 inorder to eanble the GSATA secondary satat 2 ports as valid disk extensions (not just RADI1 backup). I can get 8 disks off the motherboard.

I require this card to support 8 x SATA 1.5TB disks in JBOD.

My problem is when I boot the computer the card BIOS is processed, I can change the card settings, but I don't see any physical disks.

When WHS starts the card shows 8 amber lights but WHS still does not see any of the disks which is not unexpected.

I am using 2 SFF-8087 'straight' cables with 2 x 4 SATA II connectors. This is my second set of cables.

Make sure the SFF-8087 to 4-SATA cables you're using are the "forward breakout" cables. If you by chance got ahold of "reverse breakout" cables then your disks won't show up.
 
I've tried the 2.6.32 kernel and it doesn't help. Apparently the updated drivers for mvsas haven't made it to the mainline yet or even the 2.6.32-git releases. Sucks. :(
 
Damn windows 7 for not having software raid 5!!!! :(

I may need to switch to a server 2008 system :(

but since this is my primary system I find it a bit hard to use server, since I than get a lot of program errors with rare programs I use :(

Shawn
 
IS it for certain this card will not work with Windows XP? Not listed on the spec page but not sure if this is due to the old age of Windows XP.
 
Are there any confirmed 775 core-2 motherboards that will support two of these cards in PCIe slots? I see it posted two cards are working in a Gigabyte MA790GP-UD4H MB but I have a Core 2 CPU sitting around to use.
 
Has anyone tried OpenSuSE/Fedora/CentOS with these yet?

I've read somewhere someone was having problems with the areca 16 port which uses the same marvell chipset, and they were using the latest OpenSuSE :( Is linux support broken in *all* the distros?
 
As an eBay Associate, HardForum may earn from qualifying purchases.
okay, thanks. I do not damn Fresno,CA and no Fry's but I will be in San Jose for Christmas so maybe i will wait til then

thanks
 
Back
Top