X9SCL and 32gb ram

[chappys4life]

I have a X9SCL with a E3-1230v2 (updated bios) and trying to use 2 of the crucial 16GB kits CT2KIT102472BD160B. The machine powers on but after a few days cuts off, it appears to be a memory issue.

Any thoughts? Is there a different ram I should use?

 

[RTM]

As far as I can tell, your RAM should be compatible, Crucial also claims that with their "Will it work" function on their website, but of course the only sticks that are supported are the ones on "Supermicros Tested memory list".

I would do the usual RAM incompatibility/stability checks, which for me are (in no particular order):
- Run memtest86+, if it detects any errors try to isolate the stick or slot that is at fault.
- Reseat all sticks and remove dust from sticks/memory slot with canned air etc.
- Update BIOS (newer BIOS could contain memory compatiblity and stability fixes)
- Perform overheating trouble shooting, heat could be affecting the memory sticks directly or the CPU.
- Use a different powersupply, if the PSU you are using is bad (all PSUs can be affected) you can get all kinds of weird problems.
- Reseat CPU (in the E3v2 series the memory controller is located in the CPU, a bad connection to the MB could cause connection problems to memory sticks).

I may be forgetting something, but atleast if you do these things you will be occupied for a while

EDIT: A couple of things i forgot:
- Remove any hardware that is not absolutely essential for booting system (no PCIe cards, only boot disk, etc.)
- Try to run your system with a clean OS (Centos 7 or whatever might fit the bill), you haven't specified exactly how you come to the conclusion that it is a memory problem, but in theory software problems could disguise themselves as hardware problems. It would probably be beneficial if the OS is different than the one you are using now.

 

[chappys4life]

This is box I just put together and has not been running more than 2 weeks.

-I tried another power supply
-Latest bios version
-Cpu is fine heat wise (H55 cooler)

 

[EffrafaxOfWug]

When I've seen this behaviour before (i.e. server boots fine and then turns off after some indeterminate period of time), it was always down to either one of the DIMMs being busted or (more commonly) not being seated properly. Check in the IPMI logs and see if it mentions any memory errors (in particular uncorrectable ECC) which will usually trigger a power-off from the BMC, and if so reseat the memory modules and give each slot a blow with some compressed air. Then fire up memtest86+ and run it through a few loops and see if the problem repeats itself.

At a guess I would say that your server is taking a few days to fill up its memory enough to get to the "problematic" memory region, and as such you'll get a certain period of problem-free performance. memtest86+ should be able to trigger the same behaviour within a few minutes; if you're running debian you should be able to install it and get it added to your boot menu which makes things handy.