sl0pz

It's still happening but seemingly faster and faster. I don't think there is anything on the log, is there a different way to preserve the log that might show the error (that I'm not doing)? About to have time to deep dive into it, any other guides I can follow to try to figure it out? edit: I did update to 7.2.3 and also updated the bios to the latest one just in case Edit 2: Looking at glances I have /usr/libexec/unraid/shfs /mnt/user -disks 127 -o default_permissions,allow_other taking 70%+ of the cpu but also going over 100% edit 3: https://forums.unraid.net/topic/195401-cpu-being-hammered-by-shfs-and-other-errors-memory-related/#findComment-1593075 followed this post to move appdata to cache, increased the fuse file descriptors, ran fastusr to keep os in ram, and removed plexstreams plugin as well as unassigned devices looks like lower cpu usage, don't know if it will change the lockups tho. syslog

Thinking about how the issue comes out I would think it has something to do with the internal network crashing. I can't access the gui but the dockers are still accessible. Can't wrap my head around it but server is only staying accessible for 2-3 days at a time

Thanks Jorge! In my research before posting I checked that the bios was up to date a couple years past the 3900x release and set it to typical idle current and disabled the global c-states. I'm also not overclocking the ram

Gday I upgraded from a 1600x to a 3900x and have had issues accessing the server after a couple of days time. The only way to regain access that I've found is to hard shutdown. I believe the bios is up to date enough to handle the new cpu and have disabled global c states. Docker apps all work going in and out of the network but can't see each other (radarr can be connected to but can't access jackett) I have captured the logs via write to usb function and have uploaded to take a look at. Thanks for reading and anyone who wants to help me solve the puzzle! Cheers Jan26 log1

About 4 days of uptime at this point Anyone reading this thread thinking this sounds like me, my solutions were - cleaning up energy source (UPS + new power supply) - new flash drive - upgrading memory (could have been unnessecary) My guess is, because it was the last thing I tried fixing, that the home power going on and off messed with some part of the 8 year old power supply and was not giving clean power to the system.

The previous crash was the latest one, I changed the bios settings to typical idle current instead of auto as I found perhaps it was dropping out from failing to come back from a low power state. Current uptime since then is 23 hours

Ok So crashed again last night. not sure when, the last part of the log was 4pm and I found it dead at 11:25 There are no entries in the log between those two times. The nvidia log entries are gone after the removal gpustats plugin as well. So just looking for more tests to run to see why it's turning off syslog-192.168.0.200.log

It doesn't seem like it crashed last night but it is still giving me a machine check events error I guess we'll continue to wait to see if it comes up again? syslog-192.168.0.200.log

Ok memtest 4pass had zero errors (new memory from last week) I'm not sure why I can't get video output from the mb hdmi port when the gpu isn't attached, I believe I have before. I got into the bios no problem now and enabled iommu in (from auto to enabled) as I saw another person said this worked for them Removed GPU stats plugin to see if that gets rid of the log errors. Will post back if/when it crashes overnight!

While I can get to the server via networked computer, the video output on the MB isn't working with or without the GPU plugged in. Can't access bios or video output (memtest or unraid) via the hdmi ports... edit: got video output through the gpu and running memtest

syslog-192.168.0.200.log Ok so new psu is having more uptime, it did crash last night it seems but only once so maybe something has helped with that. I am still having issues showing video output when posting to get to mem test however. From looking over this log I'm guessing still try to find a way to get to memtest? or is it pointing at something else now

Couldn't post with new memory, old memory, or single dimms. Ordered a new power supply as maybe it got damaged in the power outages, not really sure what it could be other than that.

I just popped it in, booted it up, followed with a parity check. What do you think would be the most effective test?

So added a ups and switched to a new flash drive, server went down again an hour later. Got some new ram, everything started up and stayed up pretty well for 12 hours, now it's down again. What's the next thing I could check or replace? Are there different logs to look at?

Have a ups ordered and new flash drive as well, will swap over to new drive and ups will clean shutdown if there are power issues