I flashed my Supermicro X9SAE motherboard to v2.0a (form v1.0), now my system hangs displaying "SunOS Release 5.11 Version 11.1 64-bit Screen. Copyright,,,,,"
The grub menu appears ok before this, from which I have tried playing around with kmds debug modes and kernel debug (which hangs)(https://blogs.oracle.com/dmick/entry/diagnosing_kernel_hangs_panics_with).
I have attempted booting form the live media, but this hangs in the same place. Supermicro support sent me firmware v2.00 (build date 07/31/2012) earlier today, I re-flashed but same problem. I am hoping they can provide me with v1.0. Could this be the problem - incompatible bios? Most appreciative of any assistance that can be provided.
My system specs;
Supermicro X9SAE Motherboard
Intel Xeon E3-1275v2 Processor
4 x 8GB DDR3-1333 ECC x8 SERVER SDRAM Module (HYNIX ORIGINAL)
1 x IBM ServeRAID M1015 SAS RAID Controller
8 x SATA Hitachi drives 7k4000 3TB
850W Corsair unit
PCI Slot 10
PCI Address 01:00
MPT Firmware Revision 14.00.00.00-IT
SAS Address 500605B0:04EF1210
NVDATA Version 0E.03.00.04
Hi there, I have similar socket 1155 supermicro boards and ran into the same issue. I managed to get the system to boot by disabling Intel-AES in the BIOS. If that doesn't work, try disabling VT-D and IOMMU. Hopefully you aren't using this as a Virtualbox host because the VT-D and IOMMU really help to accelerate that.
Tom, thanks for your quick response. I will check this tonight. Also Supermicro have just sent me the original firmware ( x9sae_t4182.zip), so I will attempt to reflash if the BIOS settings fail to boot the system.
Thanks Tom, your suggestion worked! I also re-flashed the motherboard with the latest 2.0a firmware. Intel-AES was disabled by default, but I disabled Intel VT-D from the following menu and the system now boots fine.
I am unsure if this is a supermicro bios issue or on Oracle's side.. I'll probably end up opening a ticket in both locations. I love the super-micro boards so and their barebone systems are used for most of my system builds. However I did not notice this behavior with the AMD boards I used in VM Hosts even though they are also supermicro. So maybe its a Solaris VT-D implementation issue.
Are you seeing a message from Solaris at boot that says "Bios miss-represents the presence of IOMMU"? I get that on some of the systems that exhibit this issue, yet there there is no doubt that they DO have IOMMU capabilities.
Interestingly I have found a similar issue on two additional systems. In both cases I had to disable Intel Trusted Execution Technology (TXT) in the BIOS. The Solaris 11 LiveCD boots fine with this enabled, but the installed OS will only boot if TXT is disabled.
Both are using an Ivy Bridge CPU, but with two different chipsets (both Intel) on motherboards from different manufacturers (one genuine intel).