System freezes with no log entry

[ralieghnet]

Currently, I'm running 6.3.5 for the system. It will run for about 12-18hrs before it locks us. I've turned on Troubleshooting mode. Here are the last logs that was reported through FCPsyslog_tail.txt

 

Nov  1 04:24:31 Tower root: Fix Common Problems: /var/log currently 2 % full
Nov  1 04:24:31 Tower root: Fix Common Problems: rootfs (/) currently 4 % full
Nov  1 04:34:31 Tower root: Fix Common Problems: Capturing diagnostics.  When uploading diagnostics to the forum, also upload /logs/FCPsyslog_tail.txt on the flash drive
Nov  1 04:34:33 Tower root: Fix Common Problems Version 2017.10.28
Nov  1 04:34:34 Tower root: Fix Common Problems: /var/log currently 2 % full
Nov  1 04:34:34 Tower root: Fix Common Problems: rootfs (/) currently 4 % full
 

I have turned off all VM's for now

Docker containers running are 

apache     cpu lock = 1

couchpotato

letsencrypt   cpu lock = 1

muximux 

mysql           cpu lock = 1

nextcloud    cpu lock = 1

nginx            cpu lock = 1

plex              cpu lock = 2-5

sabnzbd

sonarr

 

System

Ryzen 1700x

Asus Prime x370-pro

32GB Kingston Ram

 

I have run the memtest for over 24 hrs with no errors

 

Any suggestions on what I can do next. 

 

[JorgeB]

19 minutes ago, ralieghnet said:

Any suggestions on what I can do next. 

 

Upgrade to v6.4, there's known issues with Ryzen on v6.3.x, though latest 6.4 is missing an important setting, see here:

 

 

[ralieghnet]

Thanks, Ill give it a try

[ralieghnet]

The system was up and running for at least 12hrs. I'm not sure when it froze again. The last log entry was from the mover around 330am. I tried to access the system through direct access and there was no response. The only thing I can think of is any time there is a load placed on the system from either the mover, streaming, or file transfer is causing the problem. I have moved all containers and vm away from the first cpu so unraid can operate without competition for the processor. Also, it was noticed that after upgrade to 6.4 newest release one of my 4TB drives showed full. I did the recommended fix of copying a large file and this fixed the problem. But, after the system freeze the drive now shows full again. Any other suggestions?

[ralieghnet]

While actively monitoring the log I noticed these errors

 

channel 10: open failed: connect failed: Connection refused
channel 10: open failed: connect failed: Connection refused
channel 7: open failed: connect failed: Connection refused
 

[ralieghnet]

I've upgraded to the newest 6.4.13 and the system still locks up any other suggestions or logs that I can post to help get this working?

[JorgeB]

Have you tried disabling C-states? Though I don't use Ryzen and IIRC some users can only run it stable with C-states disable, more info on this thread, on can just read the last couple of pages:

 

https://forums.lime-technology.com/topic/55150-anybody-planning-a-ryzen-build/

 

[grither]

I had a similar problem months ago, when my mover was set to hourly.  changed mover to weekly, and have been up for 4 months without a lockup.  depends a bit on how big your cache drive is

 

 

[ralieghnet]

I did at one point have the mover set to hourly when I was first building the system, but I have now set it back to the default schedule. I through this would help but it didn't. I do notice that when I'm playing a lot through Plex the system seems to lock up more often. I'll give the C-state a try tonight. Thanks for all the help.

Edited November 22, 2017 by ralieghnet

[ralieghnet]

It has been since the 22 of November when I changed the C-state setting. The system has not frozen once. Thanks for the help.