Server freezes after upgrade to Electric Eel

After upgrading my system from Dragonfish to Electric Eel it’s freezes after a while. It may works for 5 hours, sometimes 2 hours and then it dies, I can’t connect to webUI, only power off helps. Before the upgrade all works flawlessly, 24/7 without any problems.

CPU: AMD Ryzen 7 1700
16 Gb Ram
Motherboard: Gigabyte A320M-H

Not much to go on here but those older Ryzens were known to be unstable at idle, perhaps EE is better at letting it idle and thus crash.

Get the latest BIOS update, install it and then look for “Power Supply Idle Control” and set it to “Typical”. If you can’t find it, try disabling at least C6-sleep states in the BIOS and see if it helps.

If it doesn’t help, run memtest overnight.

As said, before the update everything worked without any issues. I have the latest BIOS, didn’t have any issues with power and memory before. All started immediately after updating my system to Electric Eel.

You could spend hours upon hours looking at logs trying to find something that looks off, or you could spend 3 minutes rebooting the system to set the recommended Idle setting for that line of CPU. The memtest will take more than 3 minutes to set up but requires no babysitting after it gets going, chugging along while you sleep.

I view as a form of time management.
You may end up looking at those logs in the end, but at least by then you know it’s likely warranted.

1 Like

I have a similar issue. Server ran fine before, now after about a day it freezes and I need to force reset it. Did you resolve the issue mentioned with the idle settings?
I’m on a Ryzen 9 3800x
128gb ECC RAM
Gigabyte Aorus Master x570

Same here. Before update no random reboots.

After update to EEl, my system reboots randomly ~3 times per day. No log information, no email

Apparently, no hardware issues.

Still, please add detailed information about your hardware.

Also affected by this. Had to roll back to Dragonfish 24.04. Though, before upgrading to Electric Eel, I previously would experience similar crashes at really random intervals, like sometimes after a week ,sometimes after 2 months. I believe it is related to the Ryzen idle bugs especially with the first generation, despite changing BIOS settings to avoid that.

Ryzen 5 1600X
Asus Prime B250-Plus
32 GB ECC RAM
Intel Pro 1000 4 Port NIC

If you’ve updated the BIOS and set Power Supply Idle Control to Typical and seen no change you could try also disabling C6-sleep states. Just to see if it helps.

If not, I suspect your crashes are due to a different problem, possibly bad RAM, PSU or maybe CPU.

Start testing the RAM with a thorough memtest, let it run for at least 6 hours, ideally a full day, and double check what you actually have since I can’t find anything that confirms that your board actually supports ECC RAM. In fact, the manual specifically says it supports non-ECC. It makes no mention of ECC functionality.

Intel Docs say the B250 Chipset does not support ECC.

I always assumed ECC memory does not work in non ECC boards. Maybe it is different wether it is unbuffered or registered?

Sorry, that was a typo on my part. Its a Asus Prime B350 Plus

From the manual:
Memory
AMD Ryzen™ processors:
4 x DIMMs, max. 64GB, DDR4 3200(O.C.)/2933(O.C.)/2666/2400/2133 MHz, ECC
and non-ECC, un-buffered memory

C-states have been disabled along with any eco settings in the up-to-date bios, and I had previously done a memtest after this issue started. Memory was fine, but could never get the crashes to fully go away. This was a very occasional problem before I put TrueNASon the system. Still, doesn’t quite explain why Electric Eel is causing several crashes a day, even 10 minutes after boot.

Thanks for looking at it.