Hey guys - included Diagnostics download from console, and System log after a safe power off. 

 

Issue: Started after updating my Mobo Bios and running into issues. Bios update seemed to have failed - system wouldn't boot at all. Cleared CMOS, then used Asus Bios flashback to install BIOS straight from the USB. I could boot again and unraid booted fine.

 

However, i noticed after trying to use plex, or trying to move files from it to my Gaming PC - it would freeze. Both the transfers and the WebGUIs & Plex.

 

System:

Asus ROG Strix B550-F Bios 3611

AMD Ryzen 7 5800X3D

64Gb G.Skillz Trident Z Royal 3200Mhz Ram (with auto and XMP profile enabled)
Cache:

WDS100T2X0C

Parity:
1xST18000NE000
Array:
2xST14000NE0008
1xST10000VN0004

1xST16000NE000

 

I only have the following Dockers installed.

Binhex-Sabnzb
Binhex-Sonarr
Binhex-Radarr
Binhex-Plex

 

Community Applications
Dynamix File Integrity - Only added AFTER issue started
Unraid Connect

 

I have tried Safe Mode (seems fine), Regular boot with Array in Maintenance (best results).

 

Unraid Changes:
Docker changed from Docker Customer Network type from Macvlan to IPVlan
Ran DB Repair on plex (since plex triggers it at times)

BIOS changes have included:
SVM Enabled
Low Power Mode for PSU.
RAM Using Auto speed and XMP profile (yes i know its intels name)

 

 

Anything im missing? Looking at previous Sys log after i reboot, i don't see any issues...

houndtower-diagnostics-20250331-1449.zip Unconfirmed 230733.zip

11 hours ago, ChiefSpoonS said:

Asus Bios flashback to install BIOS straight from the USB

Is this the old or new BIOS?

4 hours ago, JorgeB said:

Is this the old or new BIOS?

New BIOS.

 

Digging through all my screen grabs to see what the old one was...

If you find it, try reverting.

1 minute ago, JorgeB said:

If you find it, try reverting.

I have my OLD USB stick - any chance its somewhere in the logs or back ups on it? the BIOS version that is.

Edited April 1, 20251 yr by ChiefSpoonS

Typically, yes, in system/motherboard.txt

Well found the BIOS version that came with the motherboard - I think that was the version I had. Gonna test that one.

36 minutes ago, JorgeB said:

Typically, yes, in system/motherboard.txt

Didn't help going back to the BIOS - i could try going back to MACVLAN vs IPVLAn?

Macvaln is no longer an issue since 6.12.11

So i've done the following.
Original Bios
Latest BioS
macvlan vs ipvlan

C-state Disable

Low Power Idle Disabled
PBO Curve +4 & +6
No RAM XMP profile

 

Digging through the web for others who had similar issues 

If I understood correctly, the problem started after updating the BIOS? If that's correct it won't be easy to find other with the same issue.

BTW, if the BIOS was reset, make sure you apply/reapply these:

https://forums.unraid.net/topic/46802-faq-for-unraid-v6/page/2/#findComment-819173

 

Did all that except for 1 part here: https://wiki.archlinux.org/title/Ryzen#Random_reboots

Install the amd-ucode package to enable microcode updates and enable it with the help of the Microcode page. These updates provide bug fixes that can be critical to the stability of your system. It is highly recommended to use it despite it being proprietary.

 

Is that important too? Either i did something wrong - or my unraid install is jacked, it could install it

Edited April 1, 20251 yr by ChiefSpoonS

Also tried adding this to my /boot/config/go file

/usr/local/sbin/zenstates --c6-disable

Seemed to last longer but ended up crashing as well

Also so - it could have been because my NVidia card never had drivers installed via the plugin - it could be causing it. So i added the plugin - binhex-Plex sees it (i think) still crashing.

latest diagnostics

 

I just don't get it - its been rock solid since April 2023

houndtower-diagnostics-20250401-1453.zip

It's possible that the BIOS reflash/reset is using different settings than before, and one of them is causing the issues.

I'm currently running a Long run Mem Test to validate my hardware is still good. I have been doing additional testing.  I have since tried the following.

 

Updated BIOS back to latest bios (no issues this update).

MACVlan to IPVLAN, back to MACvlan again.

Disable Global C-State
Low Power idle Disabled

PBO Curve +4 and +6 due to Ryzen and Linux Kernel issues. (https://forums.unraid.net/topic/46802-faq-for-unraid-v6/page/2/#findComment-819173)

No RAM XMP Profile.

Used the Tips and Tweaks plugin to set Flow Control to OFF and NIC Offload to OFF

Setting each Container to only use 10 GB

Disabling all Containers - keeping the Array running and transfer data to another network device

Tried setting rcu_nocbs on the syslinux config

Tried adding ppcie_aspm=off

 

 

The issue seems to be an I/O issues but i don't understand what's causing it from the Bios standpoint. Everything is running fine until you have to start utilizing the disk. I believe that, because As mentioned I can disable the Dockers - and just try doing a File transfer of a large file and it INSTANTLY freezes up.

 

For the logs I attached above - the server Was powered on, as soon as I could connect to the WEBGUI I did a file transfer and it froze. I walked back to my Unraid Tower, Rebooted via the Command line since this is working for me still. when it came back up i pulled these logs.

 

 

Any thoughts on BIOS that would impact I/O stuff?

BIOS can impact everything, but it's kind of a strange issue, do you see the same issue if you transfer to the NVMe device or the array?

9 minutes ago, JorgeB said:

BIOS can impact everything, but it's kind of a strange issue, do you see the same issue if you transfer to the NVMe device or the array?

If you can tell me how I can do that I can try.

I can trigger a download from SAB, it would right to CACHE then to the array - that should suffice right?

Assuming that's was sufficient- I pulled a 4GB file and a 75GB file. Wrote to Cache then moved to Array w.o issues 

 

Server is still up and running.

Let it run for a while like that to confirm it's stable, and if yes, you can set it up to always write to cache first, assuming it's large enough, the mover can then move to the array during off hours.

yeah thats how I have it setup, Write to cache, then in the morning hours a Schedule for Mover to kick off. I manually it to ensure that file transfer to the array wouldn't crash it. I do have my Plex Docker offline - just so no one tries to access it while im troubleshooting. 

 

But sitting idle for hours or using SAB/Radarr/sonarr hasn't trigger the issue. It was utilizing Plex or moving a file off the array.

Edited April 2, 20251 yr by ChiefSpoonS

Seeing as it's only when you transfer files.. Can it be the cables?
I've seen stranger things happen because of bad cables.

gonna see if I have any Sata cables laying around - i'd need 5 lol

Sata Cables did not help, Swapped all 5. Currenting running through Prime95 - but its been running for 10 minutes currently and everything is still working.

Dropped down to 1 Stick of Ram (even though MEM test for 24 hours was 100% pass.

 

Still have GPU plugged in, after all the CPU stress test is done I'll try without it.