ML110 G5 running W2K with intermittent BSOD

[Robert Zels]

Hello,

i have a strange problem with a new ML110 G5 server (E2160 CPU, 512MB RAM, 2x 160GB SATA, GB NIC, DVD).

An older HP E800 server was replaced by this new server as the mainboard shows up some errors on POST. So i inserted the NETRAID-1M controller from the E800 into the ML110G5 to avoid installing the whole Windows 2000 Server from scratch. When systems comes up it asked for some new hardware found. Ok, i installed the network drivers from Broadcom Version 10.62b as the drivers from HP Version 10.78 did not recognize the right hardware. The driver aarahci.sys was installed by it´s INF file and works also fine. Chipset drivers for the Intel chipset have also been installed successfully. Same with the Matrox video driver. The system works fine and stable in this combination. I know that HP does not support W2K on this machine, but it runs in the way expected.

Because the SCSI discs are of the same age than the old server i wanted to copy the old partition over to the new Embedded HOSTRAID Controller in a second step. The attached
two 160G SATA drives have been configured as MIRROR. The MIRROR was set up correctly and was in condition OPTIMAL. I used Acronis to copy the whole partition to the new location and removed the old drives and the old NETRAID1M controller and booted up the new machine.

The machine booted up and i have cleaned some old drivers and special monitoring tools for the old RAID-controller. I rebooted the machine a few times and tested it very well.
Next day i checked the event log for errors and did not find anything. So machine can be used in production environment.

A day later i got a call that the new machine stays in a BSOD with an STOP error INACCESSIBLE_BOOT_DEVICE. We tryed to start the system in secure mode but no way to get it up. So i copied the partition again from the old drives to the new ones and the system was running again. I booted the cloned system also with a bootable CD to check if there where file system errors or media defects on the new discs with no result. So we connected the machine with the production network again and it works for round about a day. Next morning same BSOD. Hacks, viruses etc are not the reason because no client was running in the same network during the night and the server was the only activ machine in the network. The machines (client and server) are all not
connected to the internet in any way.

Now i removed both 160GB SATA discs and inserted the old NETRAID1M controller together with the three old SCSI disks and booted the system from this volume. Everything works
now for more than five days, so it seems to be a problem with the onboard HOSTRAID controller/driver. I think that the driver writes garbage to the discs when running under
Windows 2000 Server which will result in a damaged boot partition or environment.

Did someone else has a running W2K server in combination with an embedded HOSTRAID controller.

Was this driver realy tested with a W2K server. The INF file shows up with detailed installer information for this operating system, but on the HP driver site only W2K3 und up will be supported. Using latest from HP site.

All other hardware (Memory, SATA drives) has been tested in other systems with no result for any defect.

Maybe there is a registry value to set in case of timing problems under W2K or i have to use an older driver?

Any hint is welcome. This took me a lot of time but it is very difficult to replace a system running stable during the last 7 years with a new hardware and a complete new OS (W2K3) with all the same settings and getting all services (4D Server) up and running without
updating every software package.

Regards

Robert

[Gary Antonio Benavides]

Did you installed all the drivers from the HP ProLiant ML110 G5 Easy Set-up CD? did you update all the firmware in this new server? Is there any unknown device on device manager? What is the blue screen error code?

HP ProLiant ML110 G5 Easy Set-up CD
http://h20000.www2.hp.com/bizsupport/TechSupport/SoftwareDescription.jsp?lang=en&cc=us&prodTypeId=15351&prodSeriesId=3577708&prodNameId=3577715&swEnvOID=1005&swLang=8&mode=2&taskId=135&swItem=MTX-43b9aecd93834f978134bb10a9

[Robert Zels]

Yes, i have applied all drivers and BIOS (system 2008.06.24 / embedded controller 5.0.0.2???) with the latest release date expect the one for the Broadcom NIC (10.62b not 10.78). Device manager in Windows shows no unknown device. I get also no warning or error during boot. Also there is no critical entry in the eventlog.

After a while/day we will get the BSOD:
STOP 0x0000007B (0x81FD7BB0, 0xC0000102, 0x00000000, 0x00000000)
INACCESSIBLE_BOOT_DEVICE

It is not possible to boot the system anymore as it looks like some files of the OS are corrupt.

[Robert Zels]

All with the same or nearly the same problem should use a real HARDWARE RAID controller instead of this Adaptec pseudo Raid controller.
I have placed an ARECA ARC-1200 in one of the PCIe slots and the system works together with the older Windows 2000. No problem anymore, better performance and a true RAID.
This is an additional 100 EUR for the controller but saves many hours for trouble shoting.

Regards

Robert