Hi,

 

For a while now I've been having random lock ups or reboots every few weeks, typically around the 3 week mark, which has stumped me.

I have run Memtest, many passes, the version on the USB and a separate latest version, no errors in the RAM. I did initially have RAM errors that suddenly appeared after a BIOS update, but the kit was RMA'd and the new kit tests fine.

Stress testing the CPU with tasks doesn't seem to trigger it. In fact I can't seem to ever trigger it, it usually just happens when the machine is near idle, although once happened whist watching something via Plex.

Symptoms I've noticed:

• Nothing at all in the syslog (I did have it logging to USB for ages, now disabled but have it remotely going to syslog now). It will all look normal until nothing and then the next messages will be from the reboot.
• It would typically lock up with a sudden spike in power usage reported by the power monitoring plug connected to it, as if the processor were at full tilt. I don't regularly have a screen attached but one time I did connect it there was a funky RGB scatter pattern over the output. Other times I tried it would just be displaying the "Login" CLI prompt but be unresponsive.
• I recently replaced the motherboard (Asrock Z790 Pro RS to ASUS Prime Z790-P), figuring it was more likely that than the CPU failing, but a few days ago during a parity check it did exactly the same again with the exception that instead of hanging outright it rebooted. Again nothing in syslog. The previous motherboard did occasionally exhibit weird behaviour, not POSTing and hanging, but then would be fine again.
• I did try to enable power saving features in the BIOS, but disabled them all/reset to defaults to eliminate low power states potentially causing hangs.
• I've had a look through and all I could see weird and potentially related is the i915 driver. I am a little out of my depth here (ok, a lot) but it seems to be both modprobe in the go file, which I gather loads the module, but also blacklisted. I did have this in an i5 3rd gen and it's perhaps a holdover from there combined with something else I did. It's the only config weirdness I can spot and there was mention years back of this module causing issues. I did wonder if a "fresh" reinstall of Unraid may help but without any indication I didn't want to have to set everything back up again.

 

I'm at a loss. Generally. My remaining suspects are the CPU somehow having a glitch, perhaps in the iGPU (based on the weird display output I saw), or the other suspect is the PSU, but it's a fairly solid Corsair (can't quite remember which at the moment!).

Any suggestions/things to try? I'm tempted to buy a 14th gen i3 (i3-14100) to replace the i5-12400 with, but don't want to throw more money at it if it's something else.

 

Thank you for reading this far!

 

Edited April 8, 20251 yr by lilpete
Removed unnecessary diagnostics file

12 hours ago, lilpete said:

or reboots

If you mean the server reboots by itself, this is almost always a hardware issue, since memtest is only definitive if it finds errors, if you have multiple sticks try using the server with just one, if the same try with a different one, that will basically rule out bad RAM, the CPU or the PSU would then be my next suspect.

 

 

The rebooting by itself is a new one since the replacement motherboard, before it would just hang.

I figured it was most likely hardware, just hoped it wasn't. I've ordered a replacement CPU so will give that a go and wait 3 weeks. I understand your point on Memtest, if it's still glitchy I'll give it a shot!

So, if anyone is interested, I replaced the memory, replaced the PSU, still with either reboots or random hangs and a flashing blue/red corrupted visual output. At a loss (because at this point I've replaced everything but the drives) I recreated the boot USB on a different USB stick.

Where it was down to crashing every few days, so far it's up to 14 days with no hangs. So, despite all indications being hardware, it must have been a failing USB. Not known a failing USB stick to cause this kind of thing before, but, worth trying as a cheaper option then replacing literally the entire server to try and troubleshoot!

Let's hope it sticks! I am having almost identical issues and have replaced all the hardware (except HDDs) at least once, but still it randomly crashes/reboots anywhere between 3 and 45 days. I replaced the USB stick but it did not help.

 

Sometimes when I replaced hardware the system was stable for a few weeks, and as soon as I started to think everything was fine it crashed again. I've found no way to reproduce it manually, so I just have to wait and see...

As much as I wanted you to be wrong, with nearly a month of uptime it randomly rebooted this morning.

Have you had any more joy since?

For what it's worth, I have since taken out the smart plug (Kasa KP115) that I had been using for energy monitoring and ever since it's been stable. It was linked to Home Assistant and didn't ever show any changes in state/being toggled. All I can think was something momentarily breaking the connection causing a brown-out. Either way without the plug it's been stable since, which is the longest it's been stable in a while. Will update if this changes.

Last week I built your build, well sort of. I have a Asrock Z790 Pro RS with i3-14100. I am getting crashes multiple times a day. Best I've had is 16 hours. I also saw "weird display output".

Rest of the build is:

16GB Corsair Vengeance DDR5 RAM

2x 20TB Toshiba N300 3.5 HDD (Array)

2x Crucial T500 500GB SSD (cache pool)

Thermaltake PF3 850 PSU

I've tried:

>Memtests - all good.

>Updated BIOS.

>Swapping USB boot drives - using USB 2 slot, USB 3.2 Slot, USB-C slot.

>Change SATA cables.

>Plugged into and not into UPS.

>Changed docker from macvlan to ipvlan.

>Set up logs to persist and nothing is logged at time of crash.

>Uninstalled all plugins and containers.

>Monitoring temps, nothing getting very hot.

Most resent (30mins ago) I've swapped the RAM from slot B2 to A1.

How is your stability since your last message?

Edited June 11, 20251 yr by jhbunraid
typos

6 minutes ago, jhbunraid said:

Last week I built your build, well sort of. I have a Asrock Z790 Pro RS with i3-14100. I am getting crashes multiple times a day. Best I've had is 16 hours. I also saw "weird display output".

Rest of the build is:

16GB Corsair Vengeance DDR5 RAM

2x 20TB Toshiba N300 3.5 HDD (Array)

2x Crucial T500 500GB SSD (cache pool)

Thermaltake PF3 850 PSU

I've tried:

>Memtests - all good.

>Updated BIOS.

>Swapping USB boot drives - using USB 2 slot, USB 3.2 Slot, USB-C slot.

>Change SATA cables.

>Plugged into and not into UPS.

>Changed docker from macvlan to ipvlan.

>Set up logs to persist and nothing is logged at time of crash.

>Uninstalled all plugins and containers.

>Monitoring temps, nothing getting very hot.

Most resent (30mins ago) I've swapped the RAM from slot B2 to A1.

How is your stability since your last message?

Sorry, I can't read. Saw you mentioned the i3-14100 but you're actually thinking about getting that one. I wouldn't based on my build.

Stability has been fine still, guessing it was the smart plug somehow! I went for an i5-14400 in the end.

That's good to hear. Changing the RAM slot didn't help me. Then I started more specific research about the motherboard. I found a few reports of all Intel 13th and 14th gen CPUs having issues. Then found the Asrock support page for the motherboard said there a specific setting for Intel suggested settings if you have instability. Well, the screenshot of the settings is not that of v17+ firmware but I found a setting at the bottom of CPU config section which was disabled. Having enabled it my server's been up for its longest time now and I've been running parity and downloading constantly to cache and moves to array and streaming Plex all to try and stress it. It's having no problems so far.

On 6/12/2025 at 2:06 AM, lilpete said:

Stability has been fine still, guessing it was the smart plug somehow! I went for an i5-14400 in the end.

I stumbled upon this thread have having the exact same issues for a while now. Religiously, around the 45 day mark my server would reboot. Nothing in the logs. Just an unclean reboot.

Same steps as OP; changed my RAM timings, C-States, ASPM then fully changed RAM, motherboard, CPU. Same thing. Started to wonder whether it was the ConnectX-3, LSI HBA or 3060Ti but couldn't find anyone mentioning these as triggers for reboots.

What I do have in common with OP is the TP-Link KASA smart plug! I'm using a EP25 which is basically the same as the KP115. This would explain the lack of logs and the reboot. It would never have occurred to me had OP not mentioned it. I agree with with him that there must be something in the smart plug causing a momentary brown out which seems to happen every 45 days. I've removed and it will see if it reboots again but I have to wait for the 45 days. I'll update if it reboots again.

OP, do you have the smart plug between the server and a UPS? That's my current set-up so maybe it's a smart plug / UPS interaction which is the true culprit?

No UPS here, the smart plug sits just between the PC's power supply and an extension lead. Still not a single hang or reboot since I removed the smart plug so I'm now pretty certain it was my problem all along. It is sat in the corner thinking about what it's done.

On 7/11/2025 at 9:05 AM, lilpete said:

No UPS here, the smart plug sits just between the PC's power supply and an extension lead. Still not a single hang or reboot since I removed the smart plug so I'm now pretty certain it was my problem all along. It is sat in the corner thinking about what it's done.

I had the same crashes with the 12400 cpu. They vanished after I limit the C-states to C7. Before crashes from 1-14 days regularly … I guess problem with the microcode of the 12400 which I found in other forums. CPU change may indeed help.

thanks