My server seems to be operating just fine, but the chassis warning light is lit and upon a reboot, the board (S5520HCV) throws out three beeps. This indicates RAM is bad. Is there any way to tell 1) if the RAM is actually bad, and 2) how I can tell which of my 6 DIMMs is actually the bad one?
If the OS sees the correct total memory capacity, I wouldn't worry about the beeps. See http://www.intel.com/support/motherboards/server/sb/CS-031644.htm Intel® S5500 and S5520 Server Boards: Not all beep codes signal an error.
For the warning LED, you can check the http://downloadcenter.intel.com/Detail_Desc.aspx?agr=Y&DwnldID=17933 SEL first.
The beep codes are not the USB devices initializing, they are BIOS beep codes. Specifically three long ones that indicate memory issues.
The SEL shows: CRITICAL event: Mmry ECC sensoe reports uncorrectable error. There has been an uncorrectable memory error for the memory module CPU_2. Channel = D. DIMM = 1.
The OS (Windows SBS 2008) does show the full memory amount.
So, I should assume that D1 from the diagram below is defective?
This is weird as three long beeps should lead to disabled DIMMs or even no boot...
Anyway, is that memory error a single instance? If it's repeated you can easily test it by swaping DIMM in D1 with DIMMs in other slots.
Hmmm, good point on the disabled RAM or no boot. There are many instances of this happening in the log (every reboot), so it's consistent that way.
I'll swap the DIMMs around this weekend; great idea. Maybe update the BIOS/FW as well.