Hi, 

 

my Server crashes currently at least once a week. 

The server:

Unraid Version: 6.12.10

CPU: AMD Ryzen Threadripper 1920x
MB: ASROCK Taichi x399 MB
Memory: G.Skill F4-3200C16D-32GVK 2x16GB
Array: 6x 10TB (2 as parity) + 1x 4TB drive

Storage Controller: LSI Logic SAS 9207-8i Storage Controller LSI00301 (with Mini SAS to 4 SATA connector cables)

Cache: 2x 1TB NVME

Case:  Inter-Tech IPC 4U-4424 

GPU: Zotac GeForce GT 730

 

I mainly use the server to serve nextcloud and calibre (still in setup) and host my reserve-proxy via swag. 
So I only use docker containers and no VMs.

The cache and the 4 TB drive are formatted to zfs. The rest of the drives is xfs. 
Nextcloud and calibre libraries are directly stored in the array and not cached.  

Attached you can find the diagnostics workstation-diagnostics-20240805-1303.zip (unfortunately after the reboot, because the server gets unresponsive completely) and the syslog syslog-199.2.22.21.log.
I also attached fotos of the screen after the server stops working. 
I already updated the bios and ran a memory test. 

 

Maybe somebody can explain the errors in the syslog and point me into a direction so I can proceed properly. 
Currently I think that maybe one of the storage controllers is broken, or maybe it is just a loose connection to one of the hard drives. 
But I am not sure and looking for help! 

Thanks in advance!
Philipp

 

Edit: Server run without problems for 2years. These crashes started to happen 2month ago. 

Edited August 5, 20241 yr by PhilippF1992

Make sure this has been taken cared of:

 

https://forums.unraid.net/topic/46802-faq-for-unraid-v6/?do=findComment&comment=819173

 

Thanks for the fast reply. 

 

The bios says, that it is a dual-rank memory.
So I changed the frequency from 2133 to 1866. 

I will update if it crashed again. 
 

Hi, 

 

unfortunately the server still crashes and is completely unresponsive. 

The screen showed the same error message regarding the kernel panic. 

 

Do you have any other idea? 
 

Have you traced the PCI ports throwing the "Corrected Errors" back to the actual devices?

Looks to be this...

 

00:01.1 PCI bridge [0604]: Advanced Micro Devices, Inc. [AMD] Family 17h (Models 00h-0fh) PCIe GPP Bridge [1022:1453]
    Subsystem: Advanced Micro Devices, Inc. [AMD] Family 17h (Models 00h-0fh) PCIe GPP Bridge [1022:1453]
    Kernel driver in use: pcieport

 

Have you searched and found/read this?

Also update to 6.12.11, since that looks like a macvlan call traced, and those should no longer happen with the latest release.