this controller (LSI SAS1068E) is a catastrophe for grub2 and linux in general.
YOU CAN NOT USE IT TO BOOT LINUX FROM IT.
YOU WILL HAVE TO BOOT LINUX / XEN FROM AN HARDDISK CONNECTED TO THE INTERNAL ICH10R Controller CONTROLER!
THEN YOU CAN SETUP/INSTALL/RUN (CITRIX) XEN BY BOOTING FROM THE WEB-GUI-CD-MOUNT INTERFACE.
HOWEVER THE SAS-Controler CAN ONLY BE USED FOR STORING DATA. (with jumper removed) NOT FOR BOOTING!
i have installed ubuntu server on a ssd that is connected to one of 6 INERNAL ON BOARD SATA-ICH10R-ports directly on the MOBO. (i bet grub won’t boot anything installed on the jumper-removed SAS-Controler that has become SATA then)
i was not able to get the f****** soft sas raid controler of the X8DT3-F to work under linux (xen Centos 5.X, debian/ubuntu) ……. so what you can do is disable software raid modus by removing this jumper and then it acts as “normal” sata controler that you can use linux software raid mdadm but you can not boot from it.
BUT I GUESS (NOT FOR SHURE) THAT LINUX WILL NOT BOOT FROM IT!
(SO ALL BOOTABLE STUFF NEEDS TO BE ON THE HARDDISKS CONNECTED DIRECTLY TO ONBOARD SATA CONTROLER)
it identifies under linux as: LSI MegaRAID 1068E 8-Port SAS Controller LSI Logic / Symbios Logic MegaRAID SAS 8208ELP (rev 08) LINUX Firmware 80-00163-01_RevD_SAS_Driver_UG-1
But you can also spend days and nights getting the driver to work under linux. Tell me if you were lucky.
Btw: after removing the jumper, xentos (lol xen hypervisor from citrix iso is based on centos… so i call it xentos 
) reports “task nash blocked for more….” and stops booting…… so……… i guess……… install xen on ubuntu server seems to be the best solution or try kvm. (i am still trying)
mdadm example:
46. lvm on softraid
list any existing soft raids.
mdadm -A -s
3. daten partition / festplatten zu einem raid1 zusammenfassen
/sbin/mdadm --create /dev/md0 --level 1 --raid-devices=2 /dev/hdc3 /dev/hdd3
4. create lvm on soft-raid device md0 (raid1 using hdc3 and hdd3):
one needs ceate a new partition on md0:
fdisk /dev/md0
n
p
1
[enter] accepting default values.
[enter] accepting default values.
w
one needs to tag the Partitions-ID with »8e« for LVM
fdisk /dev/md0
t
8e
w
pvcreate - initialize a disk or partition for use by LVM
pvcreate -ff /dev/md0
y
now create a new Volume Group using md0 naming it “lvmData1″:
vgcreate VG1 /dev/md0
check what volume groups are there:
[root@xen56 /]# pvs
PV VG Fmt Attr PSize PFree
/dev/hda3 VG_XenStorage-d56a511d-0144-fd13-0349-63976dbc8f46 lvm2 a- 23.98G 23.96G
/dev/md0 VG1 lvm2 a- 24.00G 24.00G
[root@xen56 /]# pvdisplay
--- Physical volume ---
PV Name /dev/hda3
VG Name VG_XenStorage-d56a511d-0144-fd13-0349-63976dbc8f46
PV Size 23.99 GB / not usable 6.77 MB
Allocatable yes
PE Size (KByte) 4096
Total PE 6140
Free PE 6135
Allocated PE 5
PV UUID mGNLUx-XYWy-ngCw-XE6I-8Ykw-BXlP-R5oV8N
--- Physical volume ---
PV Name /dev/md0
VG Name lvmData1
PV Size 24.00 GB / not usable 1.25 MB
Allocatable yes
PE Size (KByte) 4096
Total PE 6143
Free PE 6143
Allocated PE 0
PV UUID SUqneo-gk4u-RCf8-p3nW-swSc-NmKa-ZwT0wv
PE are like sectors/blocks on normal harddisks.
Now lets create a new usable LVM-partition called "lvmData1":
lvcreate – create a logical volume in an existing volume group
with the maximum available PEs:
lvcreate -n lvmData1 -l 6143 VG1
with the given size:
lvcreate -n lvmData1 -L 6.0GB VG1
now lets format the newly formatted lvm-parition with an filesystem:
mkfs.ext3 /dev/VG1/lvmData1
lets mount it:
mkdir -p /mnt/VG1/lvmData1
mount /dev/VG1/lvmData1 /mnt/VG1/lvmData1
and begin testing your lvm on Soft-RAID1 
like remove one harddisk… and see if all the data is still there…
ONE IS DONE 
LSI MegaRAID 1068E 8-Port SAS Controller LSI Logic / Symbios Logic MegaRAID SAS 8208ELP (rev 08)
Firmware 80-00163-01_RevD_SAS_Driver_UG-1
LINUX SETUP NOTEZ v1.0
PROBLEM:
On servers with on-board AHCI based SATA controllers, controllers can be setto pure HBA mode or RAID mode using system BIOS. System BIOS changes the PCIsub-system IDs based on this mode.
LSI’s software RAID driver (megasr) can claim RAID mode AHCI controllers andprovide RAID 0, 1, 10 & 5 functionality. The megasr driver considers all fourPCI IDs to decide if it supports a controller.
The native ahci driver Linux kernel looks at only the vendor & device IDs. Itignores the subsystem IDs. As a result it ignores the RAID mode. While thisbehavior always existed, it poses a problem only in Red Hat Enterprise Linux 5(RHEL5). All other Red Hat distributions (RHEL3, RHEL4) load the driver froma driver disk (DUD) if it exists before loading the native drivers from theinstallation media. To override native ahci and use megasr, users just had tous the megasr DUD. Once megasr is loaded, these older versions added megasrto the initial RAM disk image (initrd) and removed ahci from it.
RHEL5 installation loads native ahci driver even when megasr DUD is present.It also includes ahci in the initrd. This is an undocumented deviation fromthe normal behavior. Red Hat initially set to resolve this in the next RHEL5update (RHEL 5.1). But later it rejected the issue.
SOLUTION:
1. To prevent ahci from loading ahead of megasr, use “noprobe” option at the installation kernel boot prompt (in addition to “dd” option). This
lets you install RHEL5 on megasr virtual drive.
boot: linux dd noprobe=ata1 noprobe=ata2 noprobe=ata3 noprobe=ata4
2. At the last step of the installation (after all the packages are
installed) RHEL5 prompts you to reboot. At this point, the initrd is
built with ahci driver. Megasr must be added and ahci must be deleted
from initrd before you reboot.
3. Pressing Ctrl+Alt+F2 would take you to a text console prompt. Run
megasr_install script from the root (/) directory. This script will
modify /etc/modprobe.conf and /etc/modprobe.d/blacklist if required.
It will create a new initrd image.
4. The megasr_install script will be provided in RHEL5 DUD floppy image.
5. If the script is in floppy or an USB flash drive, you have to mount it
first. In the normal system, we mount using /dev/sd<x> device names
to mount a device. In the pre-installation environment, you may not
have these names created. You can find the major and minor numbers of
your device by reading /proc/partitions file. Using those numbers in
mknod, you can create your own device name.
03:00.0 SCSI storage controller: LSI Logic / Symbios Logic MegaRAID SAS 8208ELP/820
03:00.0 SCSI storage controller: LSI Logic / Symbios Logic MegaRAID SAS 8208ELP/8208ELP (rev 08)
Subsystem: Super Micro Computer Inc Unknown device 0001
Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B-
Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR-
Latency: 0, Cache Line Size: 256 bytes
Interrupt: pin A routed to IRQ 16
Region 0: I/O ports at d000 [size=256]
Region 1: Memory at fabec000 (64-bit, non-prefetchable) [size=16K]
Region 3: Memory at fabf0000 (64-bit, non-prefetchable) [size=64K]
Expansion ROM at fac00000 [disabled] [size=2M]
Capabilities: [50] Power Management version 2
Flags: PMEClk- DSI- D1+ D2+ AuxCurrent=0mA PME(D0-,D1-,D2-,D3hot-,D3cold-)
Status: D0 PME-Enable- DSel=0 DScale=0 PME-
Capabilities: [68] Express Endpoint IRQ 0
Device: Supported: MaxPayload 4096 bytes, PhantFunc 0, ExtTag+
Device: Latency L0s <64ns, L1 <1us
Device: AtnBtn- AtnInd- PwrInd-
Device: Errors: Correctable- Non-Fatal- Fatal- Unsupported-
Device: RlxdOrd+ ExtTag- PhantFunc- AuxPwr- NoSnoop+
Device: MaxPayload 128 bytes, MaxReadReq 512 bytes
Link: Supported Speed 2.5Gb/s, Width x8, ASPM L0s L1, Port 0
Link: Latency L0s <64ns, L1 <1us
Link: ASPM Disabled RCB 64 bytes CommClk- ExtSynch-
Link: Speed 2.5Gb/s, Width x4
Capabilities: [98] Message Signalled Interrupts: 64bit+ Queue=0/0 Enable-
Address: 0000000000000000 Data: 0000
Capabilities: [b0] MSI-X: Enable- Mask- TabSize=1
Vector table: BAR=1 offset=00002000
PBA: BAR=1 offset=00003000
Capabilities: [100] Advanced Error Reporting
It’s a shame……. SuperMicro boards are professional Server boards and cost professional money and have professional problems….. .
But sometimes……. like this time…… they like to build LSI onBoard SAS/SATA Controlers named
LSI MegaRAID 1068E 8-Port SAS Controller LSI Logic / Symbios Logic MegaRAID SAS 8208ELP (rev 08)
which is a Software RAID that needs drivers for the operation system to work properly make life harsh…….
It is almost useless for Debian / CentOS / Ubuntu users…… (which is not a minority)
so if you are planing on buying such a board……. remember: Windows, Red Hat und SUSE ONLY!
which means……… citrix xen is (as downloadable bootable iso from citrix) NOT SUPPORTED BY THIS BOARD……..
never the less…… nerds try to make it work…….
workaround1:
get a separate NAS drive station…… an leave the FUCKING 8x HARDDISK BAY OF YOUR SUPERMICRO SHIT SERVER UNUSED! (great job! shake hands supermicro & LSI….. you suck! can’t you just open sources of the controler’s driver to the community? is it so hard to include the GPL in your sources?)
nag the support guy:
MY MESSAGE TO LSI SUPPORT:
Hi,
i have problems getting your controler to work with xen (centos-release-5-4.el5.centos.1, basically the same as red hat RHEL5).
should the RHEL5 drivers not also work with CentOS?
i am quiet new to centos/redhat.
do you have any ideas/howtos how to get this to work?
would be a shame if xen was not compatible with this hardware.
best regards
LSI ANSWER:
XenServer is not supported. Please review the link I previously sent. All supported operating systems are listed.
Best Regards,
Michaela Koert
Technical Support Engineer
Global Support Services
LSI Corporation
4165 Shackleford Road
Norcross, GA 30093
Phone: 1-866-625-3993
Fax: 1-678-728-1214
Email: support@lsi.com
On 18.11.10 15:58, Support, Software wrote:
Service Request Number: 2-276487353
CQ ID:
Date Opened : 11/18/2010 08:47:27 AM
Abstract : Linux support inquiry
Customer/Reference Number (if applicable):
Serial Number (if applicable):
Supported operating systems for the 8208ELP are available in the link below. Please note that RHEL and SLES are the only supported flavors of Linux. Which specific RHEL/SLES updates are you have trouble with?
http://www.lsi.com/channel/products/raid_controllers/megaraid_8208elp/index.html#Driver
http://forums.citrix.com/message.jspa?messageID=1400884
Re: LSI MegaRAID SATA/SAS 8204ELP Ctrlr – No hard drives found during installation
Hi,
I checked with the website of LSI and found that the raid controller 8204ELP uses the “megasr” module for linux operating system. The controller chip integrated into the card is the same which is available with LSISAS1068E but the Fusion-MPT SAS card uses the module “mptsas”. There must be some difference in internal hardware configuration of these two cards. You could ask from LSI support for the surce code of the “megasr” which can be compiled with XenServerDDK to detect the disk drives.
Regards,
Vikas Varma
i tried …… 13.15_RedHat5.zip …….. let me know if anyone had success with this constellation.
workaround2:
there is a mysterious jumper http://ubuntuforums.org/showthread.php?t=1519373 enables “pass through” mode.
which disables the MEGA-CRAP RAID and then it acts as a “normal” SATA controler which allows linux software raid with mdam (available under xen and almost any linux).
but it will not be bootable unless you flash the firmware with some mysterious firmware…… (downloadable where?)
workaround3:
http://forums.gentoo.org/viewtopic-t-731366.html
meh, I resolved it. (for gentoo)
The controller is actually an Intel SASMF8I, but looks like and is also identified as LSI 8208ELP.
It’s using LSISAS1068(E) controller chip, which is actually supported by fusion MPT/SAS driver, it’s just the PCI id of this card is not recognized by the driver..
Thus, added the id (0×0059) to the mptsas_pci_table of mptsas.c and it works nicely.
here’s the (gentoo) patch http://bugs.gentoo.org/show_bug.cgi?id=325805
http://hwraid.le-vert.net/ticket/4
Hi folks,
I’ve recently had the (mis-)fortune to run into an LSI 8208ELP (PCI ID: 1000:0059, vs LSI 1068E 1000:0058) integrated into a Supermicro motherboard. Learned a bunch of crap about the card etc. Turns out it works with the mptsas driver if you do the following:
0. i assume you are root/super user.
1. check that mptas module is loaded
lsmod|grep mptsas
if not go:
modproble mptsas
2. now there is a file existing that is not even writable as root… make it writable…
chmod u+rw /sys/bus/pci/drivers/mptsas/new_id
3. insert some magic numbers…… i have no idea what this means…….
echo “1000 0059″ > /sys/bus/pci/drivers/mptsas/new_id
echo “0×1000 0×0059″ >/sys/bus/pci/drivers/mptsas/new_id
Makes the mptsas driver recognize the chip as supported by the driver. Works fine with that. The only trick is getting the script into the initrd so you can have it online before pivot_root; got a script to do that, which I’ll send to you if you’re interested.
If you -really- want their shitty raid5, you can always use the megasr driver; someone’s created a deb package for it, at mentors.debian.net /debian/pool/non-free/m/megasr/
Hope this helps / makes it into your wiki David Bronaugh
check
dmesg
ioc1: : Capabilities={Initiator}
mptsas 0000:03:00.0: setting latency timer to 64
mptbase: ioc1: ERROR – Unable to allocate Reply, Request, Chain Buffers!
mptbase: ioc1: ERROR – didn’t initialize properly! (-3)
mptsas: probe of 0000:03:00.0 failed with error -3
….. looks like it helped me not……
…… to be continued.
workaround4:
download intel drivers
http://downloadcenter.intel.com/Detail_Desc.aspx?agr=Y&DwnldID=18570
http://asteriskpbx.ru/wiki/intel-raid-sasmf8i-install-howto
I will not describe all actions taken, I will describe just the right steps for installing Linux Gentoo on this RAID controller. Instead of looking for the driver to withdraw lspci, pay attention to what is written on the box In my case it was Intel ® RAID Controller SASMF8I. This is the starting point.
- We go to the site Intel, and shakes latest version: Intel ® ESRT-2 SATA & SAS Software RAID Driver for Linux
- http://downloadmirror.intel.com/18570/eng/ESRT2_Linux_v.13.21.0614.2010_u1.zip
- In the resulting file with a name ESRT2_Linux_v.13.21.0614.2010_u1.zip are binary drivers for RedHat & CentOS. On to the greatest of happiness, there’s a build environment drivers for any Linux.
- Next steps cite the comments of the driver assembly.asteriskware tmp #
- asteriskware tmp #
- asteriskware tmp # ls
- ESRT2_Linux_v.13.10.0708.2009.zip
- asteriskware tmp # mkdir sas
- asteriskware tmp # mv ESRT2_Linux_v.13.10.0708.2009.zip sas
- asteriskware tmp # cd sas/
- asteriskware sas # unzip ESRT2_Linux_v.13.10.0708.2009.zip
- Then go to the folder with the source for RedHat / CentOs:asteriskware SAS # cd shim/library/rhel5-32 /
- asteriskware sas # cd shim/library/rhel5-32/
- asteriskware rhel5-32 #
- Execute the command
- asteriskware rhel5-32 # make -C /usr/src/linux M=$PWD
- make: Entering directory `/usr/src/linux-2.6.29-gentoo-r5′
- LD /tmp/sas/shim/library/rhel5-32/built-in.o
- CC [M] /tmp/sas/shim/library/rhel5-32/../../open_source/oss/lin/lin_oss.o
- CC [M] /tmp/sas/shim/library/rhel5-32/../../open_source/osl/lin/osl_char.o
- CC [M] /tmp/sas/shim/library/rhel5-32/../../open_source/osl/lin/linux_osl.o
- /tmp/sas/shim/library/rhel5-32/../../open_source/osl/lin/linux_osl.c: In function ‘setup_sglist_each_sg’:
- /tmp/sas/shim/library/rhel5-32/../../open_source/osl/lin/linux_osl.c:2031: warning: comparison is always false due to limited range of data type
- LD [M] /tmp/sas/shim/library/rhel5-32/megasr.o
- Building modules, stage 2.
- MODPOST 1 modules
- CC /tmp/sas/shim/library/rhel5-32/megasr.mod.o
- LD [M] /tmp/sas/shim/library/rhel5-32/megasr.ko
- make: Leaving directory `/usr/src/linux-2.6.29-gentoo-r5′
- asteriskware rhel5-32 #
- O wonder! The driver met (before that was downloaded a similar file from LSI, but only an attempt to build on Linux asteriskware 2.6.29-gentoo-r5 ended unsuccessfully).
- File megasr.ko – it’s just something that we do not have enough to gain a sense of inner peace.
- By the way, the assembly megasr.ko made in the already deployed AsteriskWare copy on another machine, as well as the architecture was identical, we simply copied the file into the installed system. To do this right during the installation process, it is necessary to unpack the kernel in RAM, so it will need a minimum of 4G. So if you putAsteriskWare the machine controller SASMF8I, do not be lazy, install and under VMWare or some other system, it is easier, because you can just copy the file.
- Once the target AsteriskWare appeared binary megasr.ko, it must register in the image modules, loadable up to boot. This is done using a simple trick:cd / tmp / mkdir myinitramfs & & cd myinitramfs CP / boot / initramfs *. MV initramfs * initramfs.gz initramfs.gz gunzip cpio-I <initramfs CP / root/library/rhel5-64/megasr.ko lib / modules / * / kernel / drivers / scsi vi etc / modules / scsi
- cd /tmp/
- mkdir myinitramfs && cd myinitramfs
- cp /boot/initramfs* .
- mv initramfs* initramfs.gz
- gunzip initramfs.gz
- cpio -i < initramfs
- cp /root/library/rhel5-64/megasr.ko lib/modules/*/kernel/drivers/scsi
- vi etc/modules/scsi
- Next you need to add the line above: megasr, well, pack it all back:
- find. | Cpio-H newc-o | gzip -9> / boot / initramfs *
That’s it. Reboot.
PS It seems easy, but the whole internet is full of conflicting information about how to do it, and all because lspci identifies the controller as LSI MegaRAID SAS 8208ELP
