RAM goes "missing" with IBM M1115 installed

[Markess]

This is a "new" Unraid build that includes an IBM ServeRAID M1115 in IT Mode (9211 8i P20 firmware). With the HBA installed, RAM in DIMM Slot A of the ASUS P9D-I motherboard goes unrecognized at the BIOS/UEFI level (board has two RAM slots). I'm wondering if this is some sort of little known behavior of the motherboard or HBA, or if I have a bad component(s), or made some other bone head error?

I've tried swapping RAM between the slots, as well as different brands/capacities. In each case, RAM in Slot A isn't recognized. The card seems to function fine. Unraid got the array set up OK (I didn't discover the issue till Unraid was up and running). The "missing" RAM behavior is the same with the drives disconnected.

When I remove the card all the RAM appears again. I tried some other cards, including a different SAS 2008 based card with the same P20 firmware (SM AOC-USAS2L8i) and all the RAM always appeared in BIOS.

This leads me to think its the IBM M1115, rather than the RAM or motherboard, but maybe I'm missing something?

If it wasn't a small form factor build I'd just keep the SM card in it, but its too tall. Before I buy another low profile card though, I thought I'd see if anyone had any ideas and/or recommendations. Build details below.

Thanks.

ASUS P9D-I Motherboard (Newest BIOS)
Intel i5-4570TE
2 x Micron 4GB PC3L-10600E (also tried Crucial 4GB & Hynix 2GB)
IBM M1115 (IT mode using 9211-8i P20 firmware. No BIOS)
Rosewill Quark 550w Power Supply

 

[Spartacus]

What happens if you take out the ram in slot B with the 1115 card in?
Any pins on the mobo bent? have another power supply you can test? Are all the power cables seated fully?
Those are the only things I can think of why that would cause that, doesn't seem logical why that would do that in any sense though.

 

[neb50]

It is probably an issue with the SMBUS pins on the HBA card. You may be able to tape it off.

See video on it on an Dell card that can cause the same issue.

 

[Spartacus]

I thought that was only with the dell cards? (worth a shot though)

 

[Mithril]

I thought that was only with the dell cards? (worth a shot though)

I've had an HP card that caused bootlooping on a motherboard until I tapped the pins.

 

[Markess]

And the winner is.... @neb50 . At least sort of. Tape on the pins worked to get the RAM recognized with the card installed.

But, on the first boot the BIOS lost all its settings, and now it keeps crashing with what I think(?) is a kernel panic every few minutes. Unfortunately, whatever is wrong is happening before it gets written to the Unraid system log and I only see it flash by on a connected monitor. I've enabled Syslog Server in Unraid to see if that captures what's going on.

Previously, the system was up for several days, it just wasn't registering all the RAM. So either taping the pins did something, or most likely I screwed something up removing and reinserting hardware.

This silly build has been all kinds of trouble. I'm seriously thinking I should just pull it apart and try again with something else.

 

[Markess]

Well, in the end I think that the M1115 issue may have been a symptom, and not the problem. Even with the card out of the machine, it crashes almost immediately now. The crash messaging in Unraid seems to be saying that the clocksource is unstable, and the clock_watchdog is really unhappy. The motherboard's BIOS log showed a number of ECC errors, so I swapped the memory (again) and now it doesn't boot at all. Flashing a different USB stick didn't help either.

Going to try dropping in a different motherboard/CPU/RAM next. Thanks to those who posted with advice!