obowan

I wanted to give you a big thanks for this guide! It worked like a charm. Now to work up the courage to update the authentik container itself

Zero errors on the second run. Hoping this was due to the failed upgrade or an unclean shutdown. Thank you for the help!

Will do. Thank you!

Hello, I've been trying to make my server more efficient and recently switched over from an Asrock B660m Steel Legend to a Gigabyte H610i motherboard to try and lower energy usage. That part has been successful and the server feels like it is performing well but upon running my first parity check, over 2000 errors were corrected. I didn't think much of it, assuming some data had been moved incorrectly during the rebuild process but I recently started a new check, and it once again found over 2000 errors. I am running a check with corrections on and am planning to run again with corrections turned off but was hoping someone could review my logs and see if the culprit is obvious. I'll add that I attempted to upgrade from 7.1.2 to 7.1.3 but had to roll back due to some networking issues. I also ran Memtest for about 3 hours and no issues were detected. Thanks for the help! tower-diagnostics-20250616-0731.zip

@pOpYRaid would you mind sharing what CPU you have? I have an i5-14600 with a Gigabyte H610i mobo and Pure Power 12 M 550W Gold power supply. I also purchased a new I226-V NIC to replace the onboard I219. Without the NIC installed, I can hit C8 and see power draw lows of 18W. With it installed, I can only hit C6 and typically will see lows in the 22-24W range. Not a huge penalty but was curious if you had any advice. I've enabled Native ASPM, confirmed that L1 is enabled across all peripherals and manually set the C-State limit to C10. I'm just itching for that sub 20W range! Appreciate it

@leheI haven't checked in on this in a while but after reading your comment, I hit C8! Really appreciate the info. An update: after forcing ASPM, my network speed on my server tanked. I have a 2.5Gbe hardwired connection and verified with iperf3 that my connection speed went down by about 5x. I reverted the change and saw my speeds return.

Thank you @bmartino1 for your response. Whenever I ask my voice assistant a question that requires Ollama, I see the 6 logical cores of my server spike. I tried to pin all logical and virtual cores to the docker container but it didn’t really change anything. I also attempted to pass my iGPU through to the container too but I don’t think it’s able to utilize it at this point. Thanks again!

Hello Unraid folks! I have been tinkering with Home Assistant's Voice Assitant feature and after watching the below video from Network Chuck, I decided to run Ollama 3.2 on my Unraid server for a better AI experience. I have everything up and running using the Ollama container found in the Community Store but when my assistant uses Ollama, there is about a 20 second delay until I get an answer. My Unraid server specs are below. Is this about what I can expect or are there ways to speed up the response time? I tried the tinyllama model in lieu of 3.2 but it's not as capable. Any guidance/information is greatly appreciated! Server specs: Intel i5-14600 32GB DDR4 3200mhz RAM Seagate Firecuda 530 NVME

Well I thought this was done but just made a discovery: out of curiosity, I disabled Native ASPM in the BIOS on my mobo and upon rebooting my server found that ASPM shows as enabled on all my devices! I am still using the native Realtek driver, R8169, and see no loss in speeds in docker or in iperf3. That being said, I'm still stuck at C3 once my drives spin down. I then installed the R8125 driver plugin, added the r8125.conf file to /boot/config/modprobe.d, and the speed issues persist. I am marking this new entry as the solution because ultimately, I will need an updated Realtek driver for the lower C states to be possible.

I am closing this topic because the issue is related to the onboard Realtek R8125 ethernet adapter. I went so far as to purchase another motherboard, with an onboard Intel i225v-3 controller, which caused me even more issues. It seems almost impossible to find a combination of performance, power efficiency and stability with the current gen motherboards out there. I will be following the Realtek plugin page in the hope that a solution will be found at some point.

Thank you for sharing this. While this did effectively disable EEE, which the previous r8125.conf file did not, my speed issues still persist in Unraid.

@PeeFeSwhen you implemented the r8125.conf fix, did you notice any speed issues? Any information you can share is appreciated.

I think you are right, it's just odd to me that iperf shows full performance while in practice, speeds are reduced. I'm wondering if it's similar to this Bridge issue, just without the issue visible in iperf tests?

@crushjzWhat do you see when you run spci -vv | awk '/ASPM/{print $0}' RS= | grep --color -P '(^[a-z0-9:.]+|ASPM )' ? I have the same NIC as you and in Unraid the stock driver does not support ASPM. If you install the R8125 plugin, you can reach higher C states but for me at least, my speed becomes drastically reduced in my docker containers. One of my PCIE devices states "ASPM disabled" until I load the Realtek plugin, so that might be part of what you're seeing also. Hope this helps!

@SzeneI'm having a similar issue but the weird thing is my iperf tests show full speed. With the stock driver, I get no ASPM and am stuck at C3 but get full speed in my docker containers. When I enable the driver, I get ASPM and can hit higher C states but my speed drops drastically in my containers. I even tried enabling ASPM on the stock r8169 driver but see the same issue.