Hi all,

My Unraid server started repeatedly crashing over the last few days. I've been trying to figure it out myself, but haven't had much luck. The crashes are usually a total lockup - no webui, no SSH connections, no response to ping, and not even seeming to recognize a keyboard plugged directly in (no change in numlock/capslock status lights). The last crash was this morning, though it didn't completely lock up this time - I was able to SSH in and see some odd behavior on the command line. I saw tons of rpcd_lsad processes in top (first page attached here, I didn't think far enough ahead to grab the whole thing) and saw this was related to Samba, so I tried to kill that through rc.samba which didn't work. I manually killed the processes with pkill, but other processes kept the server (mostly) locked up. I wasn't able to get diagnostics to complete, but ran it after a reboot and have attached the results here. Some other things that may/may not be relevant:

• The server is running Unraid 7.2.2

• I've had intermittent crashes before (maybe 1-2 a year), but not enough for me to really try tracking it down

• This system has been running for about 5 years now

• I have run 4 passes of Memtest86+ with no issues

• The main system hardware is a Ryzen 5 2400G / 16G DDR4 / Asrock B450M Pro4 / Dell H310 LSI 9211-8i / Mellanox ConnectX4-LX / Super Flower Leadex III 550W

• The CX4-LX was recently added (last week or so). I bought it here on eBay: https://www.ebay.com/itm/267446143250

• The primary network hasn't changed and is still through the built-in Realtek 1G NIC. The CX4 is running a separate subnet (10.2.1.x for Realtek, 10.2.2.x for CX4) and a DAC cable direct to another server with an identical CX4 for file sharing only.

• I only have one m.2 NVMe slot on my motherboard and limited PCIe slots, so the second NVMe cache disk is on a PCIe x1 to m.2 adapter. This SSD was on a PCIe x4 to m.2 adapter previously, but was moved to accommodate the CX4-LX.

• There are kernel issues showing up in the logs - kernel null pointer dereference and other things I don't really understand.

• I've noticed some BTRFS corruption warnings on my cache drives. I don't know if this is an issue with the drives themselves or is just caused by unsafe shutdowns when I restart the server after it hangs like this.

The logical thing seems to be that somehow the new network card (CX4-LX) or transfer of the SSD to a new adapter card and PCIe slot is to blame since that's the only recent configuration change. I can't figure out how a bad network card would bring the whole system down like this though. For now I'm planning to bring the server back up in safe mode and keep the new hardware config, but if it crashes there I'll revert the hardware changes and try again. Any insight into what might cause this or potential fixes would be a huge help!

blackbird-diagnostics-20251208-1055.zip samba_issues.txt top_output.txt

Thanks for the quick response! I've set that in the BIOS and rebooted, and will keep my fingers crossed. If that's the root cause it's certainly an easy fix.

I've gone 7 days without any crashes, so I've marked this as solved.