Hi All,

 

My Unraid Server has been experiencing intermittent crashing over the last few months. Sometimes the server will last only a couple hours, sometimes up to a week but never longer.

 

When the crashes occur, the server GUI becomes unresponsive, can't stop array, reboot, or interact with it (ie. to stop parity check etc). At this stage, the server will be unable to be pinged and unable to be SSH'd into. A hard-reset or power-cycle is then needed to restore functionality. Server activity at this time seems to be none so I am yet to incur any corruption or data loss but is obviously not great for the server in general.

 

For reference since the issues have started occurring the following significant changes have been made to the Server, but am not sure if any are contributing to the issue.

 

1. Upgraded the Motherboard, CPU and RAM (ASUS ProArt Z790, Intel 14900K @ stock settings, Corsair 2x48GB DDR5 RAM @ without overclock)

2. Upgraded Array drives and Pool drives to ZFS

3. Changed from Docker Disk IMG to Docker Directory format

4. Upgraded from Unraid Stable Release 6.12.13 to Pre-Release Version 7.0.0-beta.4

5. Deleted ZFS Docker Directory and recreated using Overlay2 driver as recommended per Pre-Release Version 7.0.0-beta.4 patch-notes. 

 

It may have felt like there were slightly less crashes before migrating from XFS to ZFS and also before upgrading from Stable to Pre-Release, but this may also just be in my head and the crashes were still occurring nonetheless. 

 

A look through the syslog spanning at least a few of these occurrences doesn't seem to show anything particular but I have attached incase someone more Unraid-literate than I can see something I cannot. My best guess is to chalk this up to a hardware issue?

 

As always thanks in advance,

Andrew

syslog-192.168.0.34.log

Unraid driver is crashing, this is usually a hardware problem, try running memtest, could also be the Intel CPU stability issue.

Thanks JorgeB,

 

I have taken the following steps:

1. Upgraded Motherboard bios to latest version.

2. Reseated RAM and changed from slots A1/B1 to A2/B2 as per manufacturer recommendations.

3. Ran memtest, passed with zero errors.

4. Reseated CPU.

5. Re-adjusted cooler mount pressure.

6. Adjusted the following bios settings (as per recommended elsewhere on the Unraid forums):

- All settings as per Intel recommended defaults.

- Set ASPM from hardware-controlled to software-controlled.

- Disabled power saving.

- Disabled Fast Boot.

- Disabled C-States.

 

The server has now been running no issues for almost a week which looks promising.

 

The motherboard in this server has been replaced 3 times, CPU twice and RAM twice, with similar outcomes. So it seems the culprit is likely some funny business with power-saving states, likely software-side, as this was a similar erratic behavior that was caused by installing powertop, ultimately leading to me uninstalling it.

 

Unfortunately, the downside is the server now has a substantially higher power-draw even at idle which is obviously suboptimal, and will be looking to tweak back in some power-saving settings slowly over the coming weeks. Will update with any major changes.

 

Thanks again for the help.

Edited December 4, 20241 yr by Irithor