Summary: problems with 4GB DIMMS in T5120



I had replies from Tony Cesaro, Bob Wickline, Changa Anderson, Bryan
Allen, Matthew Stier, and Francisco Roque. Several people have been
down this road, and even though Crucial (and other vendors) claim their
memory will work, people have not had (much) success. Bryan
specifically dealt with Crucial and ended up returning his memory.

Changa recommended looking closely at logs and show commands in ilom.
"I had this problem and what I found was that a log told me that the
simms weren't sun compatible. There is a flag for each dimm that can
tell you that." I don't see anything in the show report for the DIMM
that raises flags to me.

Francisco recommended updating the firmware. I already had the latest
firmware (now 8 days old). He also said he's had better luck with
dataram memory than with other third party vendors.

Matthew heard from Dataram that Sun does check the DIMM-ID's on memories
in newer systems.

I'm returning the memory; this sounds problematic at best.

thank you

Tom Lieuallen

Tom Lieuallen wrote:
We have an 8 core, 1.2Ghz Sun T5120 that is half filled with 2GB DIMMS
from Sun. The T5120's have 16 memory slots. The manual says this:

* There are a total of 16 slots that support industry-standard FB-DIMMs.
* Supported FB-DIMM capacities are 1 GByte, 2 GByte, 4 GByte, and 8 GByte.
* Valid quantities of FB-DIMMs are 4, 8, or 16.
* All FB-DIMMs in the server must be the same capacity.
* All FB-DIMMs in a branch must have the same part number.

We tried 4, 8, and 16x 4GB DIMMS purchased through Crucial, but the
machine won't recognize them. It seems to see something initially, but
then ends up disabling the memory controllers (I think). I don't seem
to have the full logs from this, but you get the idea. It disabled MCU0
then MCU2 and in the end, showed no memory available and powered off
again.

With 4x DIMMS, it just sits there and spins |\|/. We tried mostly
testing with 8x DIMMS.

Has anyone been here? Anything we're missing? There are no faults when
we start and in testing, we clear any faults that show up. We have
plenty of memory and are fairly sure that we're not dealing with
outright bad memory. These are the same specs as the original memory.

thank you

Tom Lieuallen

-----------------------

Chassis | major: Host has been powered on
Chassis | major: Oct 12 17:41:33 ERROR: MCU0 initialization failed: link
training, disabled
Fault | critical: SP detected fault at time Mon Oct 12 17:41:34 2009.
/SYS/MB/CMP0/MCU0 Forced fail (Link Training)
Chassis | major: Oct 12 17:41:38 ERROR: MB/CMP0/MCU1 unused because
MB/CMP0/MCU0 is not configured
Chassis | major: Oct 12 17:41:38 ERROR: MB/CMP0/L2_BANK0,
MB/CMP0/L2_BANK1 unused because MB/CMP0/MCU0 is not configured
Chassis | major: Oct 12 17:41:38 ERROR: MB/CMP0/L2_BANK2,
MB/CMP0/L2_BANK3 unused because MB/CMP0/MCU1 is not configured
Chassis | major: Oct 12 17:41:38 ERROR: Degraded configuration: system
operating at reduced capacity
Fault | critical: SP detected fault at time Mon Oct 12 17:41:38 2009.
Oct 12 17:41:38 ERROR: Operating with a degraded memory configuration.
Chassis | major: Oct 12 17:41:38 ERROR: Operating with a degraded memory
configuration.
Chassis | major: Oct 12 17:41:38 ERROR: System DRAM Available: 016384 MB
Chassis | major: Oct 12 17:41:39 ERROR: Only 4 cores, up to 32 cpus are
configured because some L2_BANKS are unusable
Chassis | major: Oct 12 17:41:48 ERROR: MB/CMP0/MCU1 unused because
MB/CMP0/MCU0 is not configured
Chassis | major: Oct 12 17:41:48 ERROR: MB/CMP0/L2_BANK0,
MB/CMP0/L2_BANK1 unused because MB/CMP0/MCU0 is not configured
Chassis | major: Oct 12 17:41:48 ERROR: MB/CMP0/L2_BANK2,
MB/CMP0/L2_BANK3 unused because MB/CMP0/MCU1 is not configured
Chassis | major: Oct 12 17:41:48 ERROR: Degraded configuration: system
operating at reduced capacity
Fault | critical: SP detected fault at time Mon Oct 12 17:41:48 2009.
Oct 12 17:41:48 ERROR: Operating with a degraded memory configuration.
Chassis | major: Oct 12 17:41:48 ERROR: Operating with a degraded memory
configuration.
Chassis | major: Oct 12 17:41:48 ERROR: System DRAM Available: 016384 MB
Chassis | major: Oct 12 17:41:50 ERROR: Only 4 cores, up to 32 cpus are
configured because some L2_BANKS are unusable
Chassis | major: Oct 12 17:41:59 ERROR: MB/CMP0/MCU1 unused because
MB/CMP0/MCU0 is not configured
Chassis | major: Oct 12 17:41:59 ERROR: MB/CMP0/L2_BANK0,
MB/CMP0/L2_BANK1 unused because MB/CMP0/MCU0 is not configured
Chassis | major: Oct 12 17:41:59 ERROR: MB/CMP0/L2_BANK2,
MB/CMP0/L2_BANK3 unused because MB/CMP0/MCU1 is not configured
Chassis | major: Oct 12 17:41:59 ERROR: Degraded configuration: system
operating at reduced capacity
Fault | critical: SP detected fault at time Mon Oct 12 17:41:59 2009.
Oct 12 17:41:59 ERROR: Operating with a degraded memory configuration.
Chassis | major: Oct 12 17:41:59 ERROR: Operating with a degraded memory
configuration.
Chassis | major: Oct 12 17:41:59 ERROR: System DRAM Available: 016384 MB
....
_______________________________________________
sunmanagers mailing list
sunmanagers@xxxxxxxxxxxxxxx
http://www.sunmanagers.org/mailman/listinfo/sunmanagers
_______________________________________________
sunmanagers mailing list
sunmanagers@xxxxxxxxxxxxxxx
http://www.sunmanagers.org/mailman/listinfo/sunmanagers



Relevant Pages

  • Re: Alocated Memory Error (SQLSERVR)
    ... memory for the process dropped from 990 to 330. ... to grab available RAM for performance reasons, ... configured alert threshold rather than an error in MSDE configuration. ... Server 2000 SP4, ...
    (microsoft.public.windows.server.sbs)
  • RE: Memory problems - WinDbg and SOS: Who recognizes this pattern?
    ... generally configuration error will cause a rapid & direct exception or ... resources are batch compiled so as to avoid potential heap memory ... Is there many COM interop or PInvoke code in your webservice, ... |> | ephemeral segment allocation context: ...
    (microsoft.public.dotnet.framework.performance)
  • Virtex4+PPC+ext. RAM: Problems generating ACE files (solved!?)
    ... containing both the FPGA and the software configuration. ... resulting ACE file loaded fine. ... Which is unfortunate as the start vector points to that memory type. ... something wrong but I am having trouble accessing the instruction ...
    (comp.arch.fpga)
  • Re: Alocated Memory Error (SQLSERVR)
    ... I had the same problem except, my memory hog was SBSMONITORING. ... critical and you don't want it grabbing all that RAM. ... configured alert threshold rather than an error in MSDE configuration. ... Server 2000 SP4, ...
    (microsoft.public.windows.server.sbs)
  • Re: PCIe Driver read problem
    ... regarding reality, I guess we'll have to disagree on what reality is. ... The customer could type in the burst size he wanted in the GUI. ... The memory available was a 32Kx 32 ... the GUI and user part of the operating system. ...
    (microsoft.public.development.device.drivers)