Random server restarts - MDS CPU bug present? (SOLVED)

[MrGeeza]

Hi People!

 

Relatively new to Unraid, I moved from FreeNAS a few months ago, it's true that you get what you pay for as I'm loving it!

 

I have however one problem thats only recently started, lets say a week or so. Short and simple, my server appears to be randomly restarting and looking through the logs there are various things highlighted but I'm no expert and thought one of you might be able to see whats going on. 

 

The servers fans are set to high performance and IPMS is reading good temperatures.

 

Any further questions please don't hesitate to ask and I look forward to your reply

 

Thanks

 

Tobes ✌️

rex-diagnostics-20200713-1054.zip

Edited August 1, 2020 by MrGeeza
Update title

[JorgeB]

Try this to see if it catches something, post it after a crash.

[MrGeeza]

1 hour ago, johnnie.black said:

Try this to see if it catches something, post it after a crash.

Ok sorted. I setup remote logging to another machine and it appears to be working. It's created a directory with multiple log files inside it, each one referring to a different part of Unraid. So i'll leave it until it happens again, then check back in!

Edited July 13, 2020 by MrGeeza

[MrGeeza]

OK. I managed to get all the logs to dump into one file rather than on a per application basis 😀 I've managed to capture a reboot of the server and all the logs pre and post restart. I've attached the log file. 

 

 

Rex.log

[MrGeeza]

Looking at my CCTV feeds the shutdown occurred around 23:43 on 13/07/2020, the last entries before that time:

 

2020-07-13T22:56:12+01:00 Rex webGUI: Successful login user root from 192.168.1.141 (I LOG IN TO SEE IF IT HAD RESTARTED AFTER I RETURN FROM WORK)
2020-07-13T23:19:19+01:00 Rex afpd[13342]: transmit: Request to dbd daemon (volume Apple Time Capsule) timed out.
2020-07-13T23:19:19+01:00 Rex afpd[13342]: afp_openvol(/mnt/user/Apple Time Capsule): Fatal error: Unable to get stamp value from CNID backend
2020-07-13T23:19:19+01:00 Rex afpd[13342]: dsi_stream_read: len:0, unexpected EOF

 

Now, looking at the above I have an AFP share for a Apple Time Machine Backup and it's complaining about something. I've switched off AFP on Unraid now and cancelled the backups.

 

So next, after that time:

 

2020-07-13T23:55:00+01:00 Rex avahi-daemon[6601]: Joining mDNS multicast group on interface docker0.IPv4 with address 172.17.0.1.
2020-07-13T23:55:00+01:00 Rex avahi-daemon[6601]: New relevant interface docker0.IPv4 for mDNS.
2020-07-13T23:55:00+01:00 Rex avahi-daemon[6601]: Registering new address record for 172.17.0.1 on docker0.IPv4.
2020-07-13T23:55:00+01:00 Rex rsyslogd: action 'action-2-builtin:omfwd' resumed (module 'builtin:omfwd') [v8.1908.0 try https://www.rsyslog.com/e/2359 ]
 

I'm guessing these lines are where the server has rebooted, and is now reconnecting to the network and reestablishing a connection with the syslog server?

 

There is this line further along but could it be a red herring as it's starting up?

 

2020-07-13T23:55:16+01:00 Rex kernel: L1TF CPU bug present and SMT on, data leak possible. See CVE-2018-3646 and https://www.kernel.org/doc/html/latest/admin-guide/hw-vuln/l1tf.html for details. 

Edited July 14, 2020 by MrGeeza

[MrGeeza]

Now I've changed nothing else apart from switching off AFP on the server, I was using this for TimeMachine but I understand its been superseeded now so that's fine. Since switching off AFP, the restarts appear to have completely stopped. Uptime on the server is currently 1 day, 13 hours, 14 minutes. I will leave it a couple more days, check back, then close the thread.  

 

Any idea why this might have been?

 

[MrGeeza]

This can be marked as solved