capsloth

As requested. x15-diagnostics-20260708-0323.zip

I am looking for general Unraid guidance on repeated Machine Check Events. This is on a new server, and I am trying to determine the best way to capture more detail from Unraid/Linux. System details Server: 45Homelab / 45Drives X15 Motherboard: Gigabyte MW34-SP0 rev 1.1 CPU: Intel Core i7-14700 RAM: ECC UDIMM OS: Unraid 7.3.1 Kernel: 6.18.33-Unraid CPU microcode: 0x133 Issue Fix Common Problems alerted with: Machine Check Events detected on your server The kernel has logged repeated generic MCE notifications. Before a RAM swap, I saw 9 MCEs over about 59 hours: [Wed Jul 1 02:23:46 2026] mce: [Hardware Error]: Machine check events logged [Wed Jul 1 03:08:54 2026] mce: [Hardware Error]: Machine check events logged [Thu Jul 2 04:49:11 2026] mce: [Hardware Error]: Machine check events logged [Thu Jul 2 18:50:03 2026] mce: [Hardware Error]: Machine check events logged [Thu Jul 2 21:27:54 2026] mce: [Hardware Error]: Machine check events logged [Fri Jul 3 00:10:19 2026] mce: [Hardware Error]: Machine check events logged [Fri Jul 3 00:12:09 2026] mce: [Hardware Error]: Machine check events logged [Fri Jul 3 05:29:56 2026] mce: [Hardware Error]: Machine check events logged [Fri Jul 3 12:23:03 2026] mce: [Hardware Error]: Machine check events loggedI swapped RAM and rebooted. The MCEs still recurred afterward: system boot 2026-07-06 03:32 [Tue Jul 7 03:34:32 2026] mce: [Hardware Error]: Machine check events logged [Tue Jul 7 06:00:25 2026] mce: [Hardware Error]: Machine check events loggedSo the RAM swap did not resolve it. What I have checked so far mcelog: - /usr/sbin/mcelog exists - /dev/mcelog exists - mcelog --client returned no decoded records - Direct read from /dev/mcelog returned no queued records - Briefly starting the mcelog daemon did not recover historical detail rasdaemon: - Not installed / not available on this Unraid install IPMI/BMC: - No matching entries in the BMC SEL / event log PCIe AER: - AER is enabled - No corrected / uncorrected / fatal PCIe AER errors shown in dmesg - Only normal boot-time AER enablement lines appear EDAC: - EDAC modules are present - igen6_edac loads - edac_core loads - No memory controller is registered - /sys/devices/system/edac/mc has no mc0 Relevant output Linux x15 6.18.33-Unraid #1 SMP PREEMPT_DYNAMIC Mon May 25 10:59:51 PDT 2026 x86_64 Intel(R) Core(TM) i7-14700 GenuineIntel GNU/Linux model name : Intel(R) Core(TM) i7-14700 microcode : 0x133 EDAC state: Loaded modules: skx_edac_common igen6_edac edac_core /sys/devices/system/edac/mc: no mc0 present Current interpretation At this point I only have the generic kernel line: mce: [Hardware Error]: Machine check events logged I do not have decoded MCE bank/status/source data, so I cannot tell whether this is CPU core/cache, memory-controller/RAM, PCIe/root-complex, firmware/microcode, or something else. The system has not crashed or rebooted from the MCEs. Questions 1. What is the recommended way on Unraid 7.3.1 to capture decoded MCE detail on newer Intel platforms? 2. Is mcelog still expected to be useful here, or is rasdaemon preferred? 3. Is there a clean way to install/use rasdaemon on Unraid for this kind of issue? 4. Has anyone seen “mce: [Hardware Error]: Machine check events logged” with no decoded payload and no EDAC/IPMI/AER detail? 5. Are there specific diagnostics or logs I should collect before rebooting or changing BIOS settings? I have already opened a vendor support ticket, but I am also looking for Unraid/Linux-specific guidance on how to capture better MCE detail if this recurs.

Looks like 45Homelabs R&D confirmed it will work during the build.

Trying to figure out if swapping the default Intel X550-T2 Dual 10G Ethernet for the Broadcom P210p 10Gb SFP+ Dual Port NIC will work with Unraid 7.3.1 and if LACP is configurable?