Hello,

 

I just recently made the switch from OMV to Unraid on my small NAS.

 

I'm still in the progress of adding all services to OMV but already hit a road block early on. I observed daily parity checks caused by "unclean shutdown".

 

Looking at the diagnostic logs, I did not see any cause for a reboot / errors and I suspect that more than one reboot might have happened, so the "last" logs mirrored to flash are also not useful. So this morning, I set up the local syslog server and made Unraid log to itself, persisting in appdata.

 

Judging by the uptime, the last reboot must have happened at April 13 at 11:28.

 

During that time, only the line "wsdd2[3431]: 'Terminated' signal received." seems a little suspicious to me:

Quote

Apr 13 10:02:12 Tower emhttpd: read SMART /dev/sdd
Apr 13 10:17:13 Tower emhttpd: spinning down /dev/sdd
Apr 13 11:29:36 Tower root: Delaying execution of fix common problems scan for 10 minutes
Apr 13 11:29:36 Tower root: LXC: Waiting 10s for autostart from container(s) in background
Apr 13 11:29:36 Tower unassigned.devices: Mounting 'Auto Mount' Devices...
Apr 13 11:29:36 Tower emhttpd: Starting services...
Apr 13 11:29:36 Tower emhttpd: shcmd (55): /etc/rc.d/rc.samba restart
Apr 13 11:29:36 Tower wsdd2[3431]: 'Terminated' signal received.
Apr 13 11:29:36 Tower wsdd2[3431]: terminating.
Apr 13 11:29:36 Tower winbindd[4555]: [2024/04/13 11:29:36.259740,  0] ../../source3/winbindd/winbindd_dual.c:1950(winbindd_sig_term_handler)
Apr 13 11:29:36 Tower winbindd[4555]:   Got sig[15] terminate (is_parent=0)

.. but the following logs after that indicate a startup, rather than a shutdown.

 

The syslog server logs indicate that my assumption of multiple reboots was wrong. I have no clue in which direction to investigate so I have attached the syslog server file and diagnostic logs.

 

For obvious reasons the system is not used and not under any kind of load when it happens. Any pointers or tips for debugging this would be greatly appreciated.

tower-diagnostics-20240413-1325.zip syslog-20240413-1325.log

Edited April 22, 20242 yr by buxel

I just realized I had "mirror to flash" accidently disabled. I have turned it on now and will provide a new set of diagnostic logs after the next crash/reboot.

 

EDIT: here we go, another reboot happened at 14:23. It did not log much after I enabled mirroring to flash:
 

Apr 13 14:09:01 Tower rsyslogd: [origin software="rsyslogd" swVersion="8.2102.0" x-pid="12322" x-info="https://www.rsyslog.com"] start
Apr 13 14:16:00 Tower root: Fix Common Problems Version 2024.03.29
Apr 13 14:16:05 Tower root: Fix Common Problems: Warning: Syslog mirrored to flash
Apr 13 14:16:05 Tower root: Fix Common Problems: Other Warning: Unassigned Devices Plus not installed ** Ignored

that's it 🤔

tower-diagnostics-20240413-1425.zip

syslog-127.0.0.1.log

Edited April 14, 20242 yr by buxel

Just to be sure,

- i have reset al BIOS settings and re-plugged all cables on the mainboard/drives.

- I also ran two memtest passes for ~3h with no errors.

- made sure i'm not using macvlan after reading the latest changelog

- used a different USB port for the Unraid flash drive

 

 

Still getting these crashes. Is there anything I'm missing to gather more information about the root cause?

Edited April 22, 20242 yr by buxel

Server rebooting on its own, vs. crashing/hanging, is almost always a hardware problem, but it can be caused by different things, like PSU, RAM, board, etc.

Hi Jorge, thanks for chiming in.

That's what I assumed as well but the timing is just too odd. This behavior started exactly the day i made the switch from OMV to Unraid 😒

It is hard for me to verify this after swithcing, since the former boot drive is now the ssd cache and all HDDs are in the array. Even if i boot some live distro, it would not make any use of the HDD.

 

Could it be something driver related? My system is nothing fancy, just an old Skylake PC.

Edited April 23, 20242 yr by buxel

10 hours ago, buxel said:

Could it be something driver related?

It's possible but I would guess very unlikely, if you have multiple sticks try with just one, if the same try with a different one, that will basically rule out bad RAM.