I am facing a few seemingly unrelated issues (Unraid 7.0.1) that may be related to hardware problem, but unsure how to test/confirm.

 

List of issues:

 

Segfault in error log

Mar 25 17:16:18 Shelf kernel: inotifywait[14789]: segfault at 0 ip 000000000040288b sp 00007ffe20b84f90 error 4 in inotifywait[402000+2000] likely on CPU 5 (core 5, socket 0)
Mar 25 17:16:18 Shelf kernel: Code: 00 00 48 89 da 89 ee 48 8d 7c 24 70 e8 fe 0e 00 00 48 8b 54 24 28 48 89 de 89 ef 48 8d 4c 24 70 e8 7a 0f 00 00 48 8b 44 24 70 <48> 83 38 00 0f 84 6f 05 00 00 80 7c 24 4d 00 0f 85 fd 02 00 00 48

 

Windows VM randomly throwing runtime errors (may be just Windows thing, but it worked fine for months) - screenshot attached.

 

A few drives started reporting SMART errors:

Num  Test_Description    Status                  Remaining  LifeTime(hours)  LBA_of_first_error
# 1  Extended offline    Completed: read failure       90%       814         3971524688

Num  Test_Description    Status                  Remaining  LifeTime(hours)  LBA_of_first_error
# 1  Short offline       Completed: read failure       10%     65372         1563571408

 

At the same time, extended SMART tests keep hanging for hours (may be because of segfault error mentioned before). I tried changing cables as that was. easiest to debug but nothing changed. Could be that my LSI card is dying or my CPU is dyng or something, or it's just bad drives (still under warranty) so I kept the system down so I don't lose any data, since NAS part seems to be working right now, but I'm afraid to run the mover until I get the drives resolved to show no SMART errors at least.

 

I will run the MemTest first thing when I get back home, since I don't have JetKVM connected to this machine. but other than that, any idea on how to eliminate candidates and isolate what's wrong is welcome.

I will be able to check drives independently via any bootable Linux or even in a different machine one by one, just to be sure.

 

Diagnostics file attached.

My configuration is:

AMD Ryzen 5600G

Gigabyte AB350N-Gaming

24GB DDR4 RAM

4x 8TB, 3x 4TB, 1x 3TB (WD Red Plus)

LSI SAS2008 8i HBA card

shelf-diagnostics-20250325-1740.zip

Edited March 25, 20251 yr by brankko
Unraid version added

Any device that fails a SMART test need to be replaced, regarding the segfaults, try running memtest.

MemTest passed with multiple runs. Will check the disks one by one next and replace if failing extended SMART checks outside the Unraid box.

 

I am still confused by segfaults and buggy VM as those are running separately of HDDs (VMs are on NVMe cache disk).

9 hours ago, brankko said:

I am still confused by segfaults and buggy VM

Likely unrelated.

What happened here is not enough power from PSU.

Let me explain.

I started having Disk (read) errors on one of the larger drives. I replaced the drive, got the same thing after some time.

I changed the cables... the same... it works for days and then it detects some errors and disables the disk.

As I was adding the drives, I kept adding those SATA extender adapters and at some point it was too much for poor PSU.

I tried a different PSU and it all worked as a clock. Since I already started building second backup server, I reduced the number of drives in this machine and replaced smaller drives with larger. Now it works with the old PSU as well. It's weird because it's not obvious, but it makes sense. So people, be careful with your power load on a single branch of PSU.

Edited August 28, 2025Aug 28 by brankko