(SOLVED) Two unclean restarts within 2 weeks

[Pete0]

Hello,

 

today when I tryed to access SMB on my unRAID server I have found out that the array was offline but the server running. This happened to me before. One day after I installed the 6.12.3 update. I didn't pay much attention and blamed it on a storm we had that night. The second time was today, after the server had an uptime of about 8 or 9 days.

 

I'm not saying that 6.12.3 is the issue, it was just the first time this happened.

 

I am not 100% sure but if I remember correctly my ASUS B550-I mainboard is configured to stay turned off after power loss. So if this would be a PSU issue and the PSU fails and the power is gone, I believe it should stay off, right?

 

There are a couple of things I have changed in the near past, besides updating to 6.12.3. I changed my cache pool to use ZFS and I also changed the frequency scaling to use AMD pstates.

 

I have attached diagnostics, but I am afraid that they won't be of much help because I think the most interesting parts were removed after the restart. But maybe somebody would still find a clue - I hope.

 

Not sure what to do next. It's not easy just to start replacing hardware, everything became so expensive. A simple PSU would cost me already 100€ or more.

 

One more question: what would be the best way to enable persistent logging? Do I have to check daily and make sure that my USB drive with unRAID on it won't be completely full?

Nevermind this. I will setup a Raspberry Pi as syslog server.

 

Thank you.

 

unraid-diagnostics-20230727-2018.zip

Edited February 19 by Pete0

[JorgeB]

On 7/27/2023 at 7:39 PM, Pete0 said:

Nevermind this. I will setup a Raspberry Pi as syslog server.

Do that or mirror the log to the flash drive, then post that after a crash.

[Pete0]

I have. unRAID is now logging to a different server. Verified that it is working and now I am waiting for the next crash.

[Pete0]

This morning at 8am my server shutdown and rebooted again.

 

We had some storms in the past days and nothing happened on the server. Once I even saw the light flicker but the server was still running. This morning there were no storms. At least not visible to the eye that could cause any electrican interference. Until this morning the server was again running for about 2 or 3 weeks without any issue. It's just confusing to me.

 

I looked through the log file and didn't find anything special. Maybe someone else does. It just needs a clean up of the many autofan messages.

The server is now again doing a parity sync

 

 

unraid-diagnostics-20230827-1117.zip 2023-08-27-unraid.log

[JorgeB]

Unfortunately there's nothing relevant logged, this usually suggests a hardware issue, first thing take a look here:

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

[Pete0]

@JorgeB thank you for the link.

 

I did a CLEAR CMOS and loaded optimized defaults and went through the BIOS again.

There is no such thing as "Power Supply Idle Control" although I could have sworn that in the past I have seen this setting. Maybe it was removed after a BIOS update.

There are other settings though, ERP Ready, Energy Star Ready, CEC Ready. Tha later two I leave usually disabled. But I enable ERP Ready.

I also disabled now global C-States. Will this mean that my power drop of the server will go up? Will it go up considerably?

 

Also my ECC RAM is running on 3200 MHz. This is not a D.O.C.P Profile, in fact, there is no profile stored on the RAM. Can I just lower the frequency in the BIOS for the RAM an expect no problem? As far as I know the highest supported speed is 3200 MHz. Not sure if lowering it would give me any benefit.

Edited August 27, 2023 by Pete0

[JorgeB]

3200MT/s is OK for your setup.