Nick2Smith

I tried running with different set last month and crash still happened. I'll try only one slot.

Server crashed again today. What else should I try changing?

Could it be this issue even though I see no log messages? Global C-state is already disabled, I just changed Power supply idle current from auto to the recommended. Also added the positive 4 curve optimizer adjustment.

For the past 9 months I've been trying to track down the reason my server will randomly restart. Before the restart there doesn't seem to be any log warnings or issues, just normal operating and then the startup log messages. It is an unclean shutdown, but usually the array wont start because one of the NVME disks isn't found. I need to power off the computer for a few moments and then turn it back on for the NVME to be detected again. I've run SMART checks on the NVME and nothing seems to be wrong with it. I've tried replacing the PSU and PSU cables, I've run memtest86 for days and replaced the RAM anyways, I've run OCCT system stability test for 12 hours with no problems. Uptime can be anywhere between 12 hours to 45 days (when I restarted for an update). Does this seem like a MB issue? Any other things I can do to diagnose and test? bacchus-diagnostics-20250505-1239.zip syslog-192.168.1.69.log

Damn its less than a year old and second one in 2 years. Thanks!

pretty sure its usb unless the "()" isn't accurat.

This morning docker was down and the docker page just said docker failed to start. A reboot brought docker back but the GUI was empty. Removing some errored plug-ins and restarting again brought the gui back but I noticed this error in logs. sdb is my flash drive and I didn't find any other errors in previous logs. Apr 28 11:53:58 Bacchus kernel: I/O error, dev sdb, sector 2309696 op 0x1:(WRITE) flags 0x800 phys_seg 1 prio class 2 Apr 28 11:53:58 Bacchus kernel: sd 1:0:0:0: [sdb] tag#0 device offline or changed Apr 28 11:53:58 Bacchus kernel: I/O error, dev sdb, sector 2635 op 0x1:(WRITE) flags 0x0 phys_seg 1 prio class 2 Apr 28 11:53:58 Bacchus kernel: Buffer I/O error on dev sdb1, logical block 587, lost async page write bacchus-diagnostics-20250428-1229.zip 4 28 25.txt

Woke up to server GUI unresponsive but docker still working. On Tdarr I saw system memory usage was 100% but I couldn't access the gui or cli since there was no video output on the server gpu. At some point the server just restarted itself. My first thought was a docker fork bomb but the number of PIDs per docker seem fairly stable and reasonable currently. I checked individual docker logs and none o f them seem to be spamming stuff. syslog-192.168.1.69 (1).log

Shinobi seemed to be making a lot of PID so I stopped that and changed the USB. That seemed to have fixed those two issues, but now I have other errors in the logs. For future reference to check PIDs by docker use "docker stats" syslog-192.168.1.69.log

Opened log and saw all those errors again. Changing the USB. Any idea how to check for the docker fork bomb?

Thanks for the reply! Interesting, how do I check if that was the issue? Sorry I'm a noob. as mentioned in the topic you linked, do you think the flash dropping was due to a docker fork bomb or is there reason to believe it was some other thing? Noticed this using htop, any idea what it is?

Woke up and went to log into unraid but nothing was loading well and at some point it said no flash drive. I removed the flash drive and used chkdsk on windows both /f and /r which found no errors. After putting the usb back into the server I could no longer access the GUI or any dockers using the ip address, but I could still play plex files. Hooked up a monitor and keyboard to server to try and shut down but the "powerdown" command had some error that I forgot to note. Also tried unplugging the UPS so that auto shutdown would start but that didn't work. Finally I hit the power button which seems to have triggered the UPS timeout shutdown. This is a new MB install and first unraid boot was on the morning of July 26, after running a day of memtest86. In the log it seems like the error started around 2am July 27th but i could be wrong. Thanks for helping! syslog-192.168.1.69.log