0

I apologize if this question is not appropriate to ServerFault, but I have been Googling for hours to no avail and Sun doesn't really exist anymore and I don't have a support contract with Oracle so I can't really ask them for directions...

Anyways, I was recently given a Sun Fire X4100 (not the M2), with 2 x AMD Opteron 254 and 4GB of PC-3200R DDR1 ECC RAM and I intend to use that for ESXi but I needed to upgrade the RAM first.

After checking the documentation I was able to find online, I decided to purchase 8 x 2GB DIMMs. Now, these are HP DIMMs since I couldn't find SUN ones, and they are from Micron, as opposed to the the Samsung ones currently in the server, but did a lot of research on DIMM compatibility and was not able to find anything saying these shouldn't be compatible, so I went ahead with the purchase.

They arrived today, I did everything the documentation instructed me to do (Power off the server, replace the DIMMs, clear CMOS, etc...) but they just wouldn't work. I power up the server, it shows a RAM error (the amber lights on the RAM slots) and doesn't even POST.

So I checked the ILOM logs and this is what I got:

1102    IPMI    Log critical    Wed Nov 30 20:03:46 2016    ID = 3ce : 11/30/2016 : 20:03:46 : Memory : BIOS : Memory Device Disabled
1101    IPMI    Log critical    Wed Nov 30 20:03:46 2016    ID = 3cd : 11/30/2016 : 20:03:46 : Memory : BIOS : Uncorrectable ECC Node 0 DIMM 0
1100    IPMI    Log critical    Wed Nov 30 20:03:46 2016    ID = 3cc : 11/30/2016 : 20:03:46 : System Firmware Error : BIOS : No usable system memory
1099    IPMI    Log critical    Wed Nov 30 20:03:42 2016    ID = 3cb : OEM record e0 : 01000000044004000000000000
1098    IPMI    Log critical    Wed Nov 30 20:03:41 2016    ID = 3ca : OEM record e0 : 00000000041308084801a067f6
1097    IPMI    Log critical    Wed Nov 30 20:03:41 2016    ID = 3c9 : 11/30/2016 : 20:03:41 : OEM sensor : BIOS : Hyper-Transport Sync Flood Error
1096    IPMI    Log critical    Wed Nov 30 20:03:41 2016    ID = 3c8 : 11/30/2016 : 20:03:41 : Memory : BIOS : Memory Device Disabled
1095    IPMI    Log critical    Wed Nov 30 20:03:41 2016    ID = 3c7 : 11/30/2016 : 20:03:41 : Memory : BIOS : Uncorrectable ECC Node 0 DIMM 1
1094    IPMI    Log critical    Wed Nov 30 20:03:40 2016    ID = 3c6 : 11/30/2016 : 20:03:40 : System ACPI Power State : sys.acpi : S0/G0: working
1093    IPMI    Log critical    Wed Nov 30 20:03:23 2016    ID = 3c5 : 11/30/2016 : 20:03:16 : Physical Security : sys.intsw : General Chassis intrusion
1092    IPMI    Log critical    Wed Nov 30 20:03:23 2016    ID = 3c4 : 11/30/2016 : 20:03:16 : Voltage : mb.v_bat : Lower Non-critical going low : reading 2.61 < threshold 2.69 Volts
1091    IPMI    Log critical    Wed Nov 30 20:03:15 2016    ID = 3c3 : 11/30/2016 : 20:03:15 : Entity Presence : ps0.prsnt : Device Present
1090    IPMI    Log critical    Wed Nov 30 20:03:14 2016    ID = 3c2 : 11/30/2016 : 20:03:14 : System ACPI Power State : sys.acpi : S5/G2: soft-off

I tried setting 2 of the old DIMMs in NODE 0 and SLOT 0 and 1 and 2 of the new DIMMs on NODE 1 on SLOT 0 and 1 which made the system POST but it would only recognize 2GB of RAM installed on NODE 0, even though ILOM shows that the 2 x 2GB DIMMs are indeed installed. And the faulty RAM lights were on for NODE 0 SLOT 1 and NODE 1 SLOT 1, which seems pretty odd to me.

I have gone through all of the basic troubleshooting I could think off, cleaned connectors, swapped slots, reset CMOS but nothing worked.

According to the manuals I was able to find, this system does indeed support 2GB DIMMs and the BIOS is updated to the latest version, so is the SP Firmware. So I'm out ideas, I'm hoping someone here has or has had this same system and is able to help me out with this issue...

Extra info on the labels that might be helpful: Old DIMMs:

Model: Sun (Samsung) PC3200R-30331-C3 1GB DDR PC3200 CL3 ECC Sun P/N: 371-0072-01 Single-Rank DIMM

New DIMMs:

Model: HP (Micron) PC3200R-30331-Z 2Gb DDR 400 CL3 ECC REG HP P/N: 373030-851 Dual-Rank DIMM

Thanks in advance for the help!

1 Answers1

0

Sun equipment is picky on certain things. Make sure that your mem is installed in the correct slots, the board requires equivalent pairs. Weird errors if you try to mix and match. Try populating the mem without the old mem and see if you post. On those servers, if I recall, you need to have the same amount of mem for both processors. If they are of different layouts, they can report a slight difference in size and the system sees that as an error/bad mem.

Pierre.Vriens
  • 1,159
  • 34
  • 15
  • 19
Kevin
  • 56
  • 4