DCPMM Cache Crash?

[iskyryan]

Can having too much DRAM cause stability issues with DCPMM? Currently have 4x128GB DRAM with 2x512GB DCPMM (2:1 DCPMM to DRAM ratio) in memory mode running analytics on a Python Darc library. The system works fine with an 80GB file but anything 100GB and up and the system crashes and reboots. We think there may be some sort of issue with the memory controller. We understand the "performance" recommendation is 4:1; will going beyond that cause problems?

[sscargal]

@iskyryan I would not expect any issues with that DRAM:PMEM configuration. Can you provide more information about the issue please:

• What error(s) do you see in or around the time of the unplanned reboot? Is it an OOM (Out of Memory) Killer reboot, a Machine Check Exception (MCE), or something else?
• What mode are you using? Memory Mode or AppDirect?
• What server make/model are you using?
• What BIOS version do you have? dmidecode -t bios
• What PMem Firmware do you have? ipmctl show -dimm
• What OS are you using? Linux or Windows + Version?