I've use AI to sum up the problems

my english is not my native language.

Hi everyone,

I'm looking for help diagnosing a persistent system freeze that I have been unable to solve after several weeks of troubleshooting.

Hardware

• Unraid 7.3.1

• Huananzhi X99 F8D Plus

• Dual Xeon E5-2697A v4

• 64GB ECC DDR4 (2x32GB SK Hynix RDIMM)

• Intel Arc B570 (used for Jellyfin hardware transcoding)

• 2 × LSI SAS9300-16i (SAS3008 IT mode)

• Multiple HDDs (24TB parity)

• Two onboard Realtek RTL8125 2.5Gb NICs (bonded in Active Backup mode)

---

The Problem

The server randomly freezes completely.

When it happens:

• WebGUI becomes unreachable.

• SSH stops responding.

• Tailscale reports the server offline.

• Nginx Proxy Manager returns 504 Gateway Timeout.

• Docker containers become inaccessible.

• The only recovery is a hard power reset.

Sometimes Jellyfin can still open its homepage briefly through the local IP, but playback never starts. Eventually everything becomes completely unresponsive.

---

When does it happen?

The freeze seems to occur during relatively heavy I/O activity, for example:

• Unbalance moving files

• Jellyfin playback / transcoding

• General file activity

However, it is NOT easily reproducible.

The server can sometimes run for over a day without problems, including normal Docker usage.

---

What has already been checked

Storage

• Parity check completed successfully.

• 24TB parity finished with 0 errors.

• Filesystems were checked and repaired previously where needed.

• No obvious disk errors found.

HBA

Both LSI controllers are detected correctly.

PCIe link status:

• PCIe Gen3 x8

• PCIe Gen3 x8

No mpt3sas reset storms appear in the logs.

---

Network

Using Realtek r8125 driver.

No evidence of:

• link down

• NIC reset

• transmit timeout

Bond status is healthy:

• Active Backup

• Link Failure Count = 0 on both interfaces

---

CPU / Memory

No overclocking.

Using ECC RDIMM.

Fix Common Problems reports Hardware Failure because the diagnostics contain repeated:

mce: [Hardware Error]: Machine check events logged

However, I cannot determine what hardware component is actually generating the MCE.

---

GPU

Intel Arc B570 is detected correctly.

Hardware transcoding works normally until the freeze occurs.

---

Additional observations

Earlier I was powering CPU1 and CPU2 from different PSUs because one EPS cable was too short.

I have since rewired the system so that each CPU receives dedicated power correctly.

After rewiring, a full parity check completed successfully (0 errors), but the random freezes still occur.

---

Diagnostics

Fix Common Problems reports hardware failure because of repeated Machine Check Events.

Diagnostics are attached.

---

Questions

1. Has anyone experienced complete freezes like this on a Dual Xeon X99 platform?

2. Could this be related to:

   • Huananzhi X99 motherboard?

   • PCIe subsystem?

   • SAS3008 controllers?

   • Machine Check Events?

   • BIOS settings?

3. Is there anything specific in the diagnostics that points toward CPU, RAM, motherboard, or PCIe instability?

Any suggestions would be greatly appreciated because I feel I've exhausted the usual troubleshooting steps.

Thank you!

tick-unraid-diagnostics-20260628-1057.zip

The first thing you will need to resolve are the hardware errors:

Jun 28 10:42:47 TicK-Unraid kernel: mce: [Hardware Error]: Machine check events logged

Jun 28 10:42:47 TicK-Unraid mcelog: CPU 0 on socket 0 received Bus and Interconnect Errors in Other-transaction

Jun 28 10:42:47 TicK-Unraid mcelog: Location: CPU 0 on socket 0

This is most likely a CPU or board problem.