Hello all!

I'm pretty new into Unraid, but figured to use Unraid for my next NAS build. I built a new system and setup Unraid and I was confident to buy the license. I like the features and the apps and plug ins. I was almost ready to move all my data to the new build, when I found an issue with my NVMe's.

 

First here is my hardware setup:

Mainboard: ASUS Prime H770-Plus

CPU: Intel Core i5-14400

RAM: Crucial Pro DIMM Kit 32GB, DDR5

NVMe: 2x Kingston FURY RENEGADE SSD 2TB, M.2 2280 PCIe 4.0 x4

HDD: 2x Western Digital WD Red Pro 16TB

 

I installed the latest version of Unraid, used the two Kingston NVMe's for cache pool and the two WD HDDs for the array. The system was up and running for more than a week, everything stable and working. Than I wanted to take a look on power consumption and temperature. I found that the NVMe's were realy hot (around 75 Celsius) and also the power draw was hefty.

 

I tried a lot of different things (see a post in the German Sub Forum). Long story short: For comparison I booted from a live linux (Ubuntu) with the exact same hardware setup. Here the NVMe's only reached 30 Clesius and the power draw was significantly lower.

 

When I disconnect the SATA HDDs and boot Unraid, the temperature of the NVMe's was 30 Celsius too. But as soon as I connected the HDDs back, the temperature of the NVMe's rises to above 70 Celsius within minutes after booting.

 

Another weird thing I noticed was the "Error Information Log Entries" value of the NVMe's. I retrieved that through smartctl on the cli.

 

When I boot the system from my live linux (Ubuntu) this value does not increment at all. When I boot Unraid without the SATA HDDs connected, this value does not increment at all. Only when I boot Unraid with the SATA HDDs connected this value is incrementing continuosly. Increamenting roughly by 30 every minute. See the screenshot above. It apears to me as if the operating system is trying to repetively send varios commands the NVMe with an invalid field. Unfortunately I couldn't find any information what is doing that or what invalid field there could be.

 

I am at my wit's end and have no idea what else I can do to get to the bottom of this. I can provide more information about the setup and the system if it would help anyone. Otherwise I am desperate for any tips. Especially since the Unraid license is about to expire and I would only buy Unraid if I can get my setup to work.

Unraid only reports what the device is reporting regarding temps, looks more like BIOS/firmware issue.

Of course Unraid only reports what's going on on the device. Same does the smartctl (SmartMonTool). My take on that is, that it must have something to do with the OS, as it doesn't show at all in an Ubuntu evnironment.

 

I would guess, that the OS continuosly sends commands to the device, what prevents the device from going into powersaving states and therefore creating the high temperature. By the way, the device doesn't show any other activity in Uraid or smartctl. The value "Data Units Written" nor "Data Units Read" aren't changing where as the "Error Information Log Entries" keeps going up.

Was Ubuntu using the same kernel as Unraid?

I don't have knowledge about that. I simple tried it with Ubuntu Desktop 24.04.2 LTS because it was the first thing what came to my mind. How can I find out about the kernel?

This sounds to me more like a BIOS/firmware issue, but it could only happen with a specific kernel series, assuming you were using Unraid 7.0, which is based on kernel 6.6, try for example TrueNAS 24.10, since it's based on the same kernel, you can also first try Unraid 7.1 (kernel 6.12), just to see if there's any difference.

You were right about me using Unraid 7.0. I did an update to 7.1.0-rc.1. The NVMe behaves still the same. I'll try TrueNas as well.

So I have an update on the situation. After installing TrueNAS-SCALE-24.10.2.1 and setting up the storage 2x HDDs and 2x NVMes. The issue does not show!

 

Summary:

Unraid 7.0      -> NVMe getting hot

Unraid 7.1      -> NVMe getting hot

Ubuntu Desktop 24.04.2 LTS         -> NVME no issue

TrueNAS-SCALE-24.10.2.1         -> NVME no issue

 

On 4/11/2025 at 7:09 PM, JorgeB said:

This sounds to me more like a BIOS/firmware issue, but it could only happen with a specific kernel series, assuming you were using Unraid 7.0, which is based on kernel 6.6, try for example TrueNAS 24.10, since it's based on the same kernel, you can also first try Unraid 7.1 (kernel 6.12), just to see if there's any difference.

 Could you elaborate what kind of Bios/firmware issue you're thinking of? Any tipps which settings I can try changing? Or what can only happen with a specific kernel series?

 

I assume the HDDs are still connected? Also, does the log error issue persist with TrueNAS or not?

On 4/11/2025 at 10:54 AM, Thomas0080 said:

"Error Information Log Entries" keeps going up.

After some searching, seems to be a common issue with those devices, though unclear why it doesn't happen with every OS, or just with disks connected with Unraid:

 

https://linustechtips.com/topic/1607748-random-nvme-error-logs/

 

https://askubuntu.com/questions/1460668/nvme-faulty-drive-smart-error-information-log-entries-fastly-increasing

9 hours ago, JorgeB said:

I assume the HDDs are still connected? Also, does the log error issue persist with TrueNAS or not?

Yes HDDs connected all the time. For TrueNas I added a SATA SSD for the installation too. The error issue occurs only under Unraid, but it must have something to to with the state of the array, as I sometimes see the counter rising and sometimes the number is stable.

 

3 hours ago, JorgeB said:

After some searching, seems to be a common issue with those devices, though unclear why it doesn't happen with every OS, or just with disks connected with Unraid:

 

https://linustechtips.com/topic/1607748-random-nvme-error-logs/

 

https://askubuntu.com/questions/1460668/nvme-faulty-drive-smart-error-information-log-entries-fastly-increasing

 

I have seen this posts as well. Unfortunately they don't come to a conclusion...

 

I bought two cheap NVMe drives (Patriot P300, 128GB). I'm going to test these drives with Unraid. Probably tomorrow.

Test with Parot P300 did not show the temp issue so far. I have to run the parity process right now, but the NVMes behave as expected. 

 

Still I would be curios, what the OS does (e.g. what command is it sending to the NVMe) to keep them in this high Power Status. I've spent 300€ for NVMes which are apparently not compatible with Unraid. If I now buy two Samsung SSD 990 PRO I am not sure if it will operate as expected. Very unsatisfing. Is there a compatibilty list available from UNraid team?

6 minutes ago, Thomas0080 said:

I've spent 300€ for NVMes which are apparently not compatible with Unraid.

I don't think it's just the device and the OS, possibly also related to the board or other hardware, I have a Kingston KC3000, which, I believe, is basically the same as the Fury Renegade, and the model from one of the links above about issues, and it works fine with Unraid, that server uses an AMD board.

@JorgeB first of all, I want to thank you for your time and advise. It is much appreciated!

 

I tried to choose my setup with a lot of thought (e.g. the Kingston NVMe's have one of the highest TBW (2PB) compared with other similar products), but now I have to face the pontential outcome, that there is nothing I can do to get my setup running with Unraid. And besides this forum (where you were the only one trying to help me) I have no idea who else I can talk to.

 

Again, with all the different tests, I would conclude, that the hardware setup by itself is not an issue. I would love to learn, what's going on under the hood, (e.g which command are causing the Error Infromation Log Entries, what invalid field there is, or what is causing the NVMe to stay permanently in the highest power state), but there is apperently nobody around to help analysing the problem. Not sure if the Unraid Team is interested at all.

 

That leaves me two possible ways forward. Change the hardware setup (on my own costs) or change the OS.

I suspect it may be a combination of SSD and the board, since one of the other users also has a 14th Gen Intel CPU, this may be resolved in a future kernel, but for now, I would replace the SSDs, since it's the easiest and more certain way to resolve the problem.

I've ordered a Samsung 990 Pro 2TB NVMe. Should arive next week. Will perform some test with Unraid and post the results.

Update: Installed the two Samsung SSD 990 Pro NVMe. So far no issues. Temp is 32 Clesius. No Error Information Counts.

Currently the parity check is running. I post another update in a few days.

 

Still a bit disapointed about the whole circumstances but I'm goging to continue with Unraid anyways.