Hi All,

 

Hoping someone might be able to assist as my knowledge of Unraid isn't the best. My server has been running happily for months until recently when I wouldn't be able to connect in. The system is still running and accessible via IPMI but the OS itself has crashed and the IMPI log indicate a bluescreen crash but nothing further. A reset brings the system back up but then it happens again with a 1-3 days. The system shows no drive errors or failures and I've completed a memtest with zero faults. I was running  6.9.2 when the crashes started and tried upgrading to 6.10.0-rc2 in the hope it might sort the problem but it hasn't. Any advise would be appreciated. I have attached my unraid diagnostics logs to assist.

 

Many thanks

kronos-diagnostics-20220106-1644.zip

2 minutes ago, rees2050 said:

IMPI log indicate a bluescreen crash

Don't know what that might mean with Unraid, it doesn't go to a blue screen like Windows does.

 

Setup Syslog Server

Hello,

 

Ok maybe my IPMI logs are just giving a generic reason then. I will follow your link and turn on Syslog.

 

Thanks

All setup and I'll await the next crash......

So just gone to access server and it's not responding. I've had to power cycle from the IPMI access in order to reboot it and bring unraid back online. I've opened up the syslog and from what I can see something happened at Jan  9 04:00:29 as the next entries are me rebooting the server this afternoon. I've attached the syslog and would appreciate if anyone is able to to decipher for me.

 

Many thanks

syslog

Edited January 9, 20224 yr by rees2050

On 1/6/2022 at 12:20 PM, rees2050 said:

memtest

Was this from the Unraid boot menu?  It won't find any errors because you have ECC memory.

 

You want to run a couple of passes with a separate stick set up with https://www.memtest86.com/ 

 

5 minutes ago, Squid said:

Was this from the Unraid boot menu?  It won't find any errors because you have ECC memory.

 

You want to run a couple of passes with a separate stick set up with https://www.memtest86.com/ 

 

Hi Squid, yes it was with the memtest boot option. I don't have access to any other memory, is the syslog suggesting that its a memory issue??

 

Do you mean a separate USB stick with memtest on?

 

Thanks

Edited January 9, 20224 yr by rees2050
error

You need to create another bootstick with the program from the above link (it can't be distributed with the OS due to licencing restrictions).  It will adequately test your RAM

 

Most motherboards with ECC will halt everything if it detects a memory error that cannot be corrected.  (Which is part of why you bought components that support this)

Thanks Squid, I've done that and I've left the test running. I will update you with the results....

So memtest complete and zero errors. What next?

 

 

Anyone else have any idea what might be causing the problem? Thanks

43 minutes ago, rees2050 said:

Anyone else have any idea what might be causing the problem?

Are you trying to use the iGPU in your 2146G for hardware transcoding in Plex, Emby, HandBrake, etc.?

 

If so, you might have this issue.  It is present in 6.9.2 and the 6.10.0 rc with the Linux 5.x kernels.

On 1/12/2022 at 12:47 AM, Hoopster said:

Are you trying to use the iGPU in your 2146G for hardware transcoding in Plex, Emby, HandBrake, etc.?

 

If so, you might have this issue.  It is present in 6.9.2 and the 6.10.0 rc with the Linux 5.x kernels.

Hi Hoopster,

 

Thank you for your reply. I have been using the iGPU for hardware transcoding. After reading through you post I've followed the steps in an attempt to remedy so will monitor and see if I encounter any more crashes.