Random Restarts

[GingerSnaps35]

I'm getting random restarts on the server. Can anyone please help me figure out what it is, because the logs don't make sense to me. =(

federation-diagnostics-20200912-0228.zip

[GingerSnaps35]

I think, and I could be COMPLETELY wrong, that transcoding enabled was doing some of the restarts. At least the last two I had were as someone started a stream from my plex docker. I turned off transcoding and it is still running. Earlier today it happened a couple of times but no one was streaming my plex server.

[GingerSnaps35]

This time it did a pfsense VM restart only.

federation-diagnostics-20200912-0259.zip

[GingerSnaps35]

Ok, so right after I made the previous comment, I turned off the Plex docker and the entire server restarted.

federation-diagnostics-20200912-0304.zip

[ChatNoir]

Your last diag is 0 bytes.

Since you have random errors, did you try to do a memtest ?

[GingerSnaps35]

yes. Memtest came back good. I ran pfsense on this box for a year before installing unraid a week ago. I just went through an entire process to narrow down what could be wrong:

 

*Deleted Plex docker and vdisk; Same with pfSense VM

*Disabled VM's and Docker

*Ran Mover

*Restarted unraid

*Started array

*ran mover again

*enabled VM/Docker

*re-installed pfSense (had to do it three times to get it to recognize the quad port Intel NIC on passthrough)

 

Finally am on the interwebs. This Diag is after pfSense was installed correctly and up and running just now.

federation-diagnostics-20200912-0410.zip

[Squid]

Take this for what it is, as it's not necessarily the problem.

 

You're attempting to use hardware passthrough with a CPU that was released in 2012.  Hardware passthrough is highly dependent upon proper support of IOMMU from the cpu, chipset, and BIOS.  

 

Historically, this support while being technically available on consumer level products has been present, it has not worked very well until the later generations of cpus / chipsets.  This is especially true with regards to AMD instead of Intel.

 

No software can overcome hardware limitations, and unRaid is no different in that respect.

 

What you need to do is see if the restarts happen without the VM service even being enabled.  If they don't, then the passthrough support from AMD on that generation is entirely to blame.  If they do continue, then it's something else.

[GingerSnaps35]

So I did that Saturday and both ways it restarted, so I reinstalled the unRaid on the ISB drive Saturday afternoon, did the full setup again, and this time I just set up pfSense As a VM with the Quad Port NIC being pass-thru. Ever since then, it has been running without any issues or restarts. I did have to restart the system after the VM was made so that it saw the NIC, but once I did, it was smooth as butter and no issues. Fortunately I’m only holding less than 2TB so I can offload data and rebuild if I need to. 

[GingerSnaps35]

I chose to do unRaid because I had read that several people had my same setup and were running it as a Plex server and pfSense firewall which is what my intentions were. I guess I need to buy my AMD EPYC 7351P SOC board and build my server so I get proper Hw pass through.