Hello. I appreciate a little help. Recently I got a notification that there was an error with the server. Fix Common Problems reported that it found Machine Check Events detected on your server. And that it "detected hardware errors"

 

I recently took apart my server, cleaned all the dust bunnies out, and then reassembled the server. Everything seemed ok. I ran a parity check and it finished without problem. I did have a drive error but it cleared without repetition so far.   I did change out the CPU cooler because the fan was a dust bunny hideout spot. I also changed some cables.

 

I looked at the syslog. And these are the errors it reported since the server was cleaned ...

 

 

Quote

Mar 11 19:48:08 Tower ntpd[1746]: kernel reports TIME_ERROR: 0x41: Clock Unsynchronized

Mar 11 19:49:36 Tower kernel: BTRFS error (device sdi1): incorrect extent count for 44044386304; counted 3956, expected 3955

Mar 11 19:57:28 Tower ntpd[75412]: kernel reports TIME_ERROR: 0x2041: Clock Unsynchronized

Mar 16 20:12:35 Tower kernel: mce: [Hardware Error]: Machine check events logged
Mar 16 20:12:35 Tower kernel: [Hardware Error]: Corrected error, no action required.
Mar 16 20:12:35 Tower kernel: [Hardware Error]: CPU:0 (15:60:1) MC2_STATUS[-|CE|MiscV|AddrV|-|CECC|-|-]: 0x9c4340900010010a
Mar 16 20:12:35 Tower kernel: [Hardware Error]: Error Addr: 0x0000000000000480
Mar 16 20:12:35 Tower kernel: [Hardware Error]: MC2 Error: L2 Tag ECC error.
Mar 16 20:12:35 Tower kernel: [Hardware Error]: cache level: L2, tx: GEN, mem-tx: GEN

Mar 17 02:40:21 Tower root: mcelog: ERROR: AMD Processor family 21: mcelog does not support this processor.  Please use the edac_mce_amd module instead.

Mar 17 19:39:36 Tower root: mcelog: ERROR: AMD Processor family 21: mcelog does not support this processor.  Please use the edac_mce_amd module instead.
Mar 17 19:39:36 Tower root: CPU is unsupported

 

I assume Fix Common Problems is complaining about the last couple of lines about AMD Processor not supported, which makes me scratch my head since it is the same hardware that was running before and after the cleaning. I don't see a reason why this error is suddenly coming up. And from what I see this is really just a badly worded notice that Unraid is using edac_mce_amd module instead of the normal one.

 

Should I just tell FCP to ignore this error?

 

I'm attaching the diagnostic zip file if that helps.

 

Any help appreciated.

tower-diagnostics-20250317-2145.zip

I'm getting the same error... Would like to know what to do please.

12 hours ago, tnorman said:

Any help appreciated.

 

This happened because your CPU's on die cache had an error (and was corrected).  Could be a on-off (high energy solar radiation particle happened to hit the CPU at that time) or could be a sign of the CPU having issues.

 

Don't have FCP ignore it, but instead reboot the server and then see if it comes up again

3 hours ago, leejbarker said:

I'm getting the same error... Would like to know what to do please.

You'd need to create a new thread and include your diagnostics.  Highly unlikely your MCE error is the same as the OPs

Thank you Squid. I rebooted and ran FCP. no issue came up. Thank  you for the help. I'll try to remember to reboot and see if the error stays in the future.

Cheers @Squid will do.