Hello, 

I've been facing this issue with my unraid server quite a bit recently where out of nowhere, my server appears to lose the ability to communicate with any applications or display the WebUI. It started happening about a year ago but occurred maybe once a month, and now it's down to once every two hours. Hard restarts are required, as soft restarts never seem to finish doing pre-shutdown tasks, and I've attempted to follow almost all of the other topics here that even remotely relate to my problem I'm having, but we're back here again. I have attached what I believe are the diagnostics and the syslog files here for convenience. I will also note that while this connectivity issue is happening, I am unable to log in through the console or WebUI physically on the machine itself, as the console will time out after my attempt to log in, and the webui will show a blank black screen instead of the desktop. Any help would be massively appreciated.

helios-diagnostics-20241113-1219.zip syslog

per you logs. What dockers are you running and how are you handling there networking..

"http://localhost:8080/Dashboard"
Nov 13 12:17:31 Helios root: Fix Common Problems: Warning: Docker Application CrashPlanPRO has an update available for it
Nov 13 12:17:31 Helios root: Fix Common Problems: Warning: Docker Application dizquetv has an update available for it
Nov 13 12:17:31 Helios root: Fix Common Problems: Warning: Docker Application librespeed has an update available for it
Nov 13 12:17:31 Helios root: Fix Common Problems: Warning: Docker Application netdata has an update available for it
Nov 13 12:17:31 Helios root: Fix Common Problems: Warning: Docker Application qdirstat has an update available for it
Nov 13 12:17:31 Helios root: Fix Common Problems: Warning: Docker Application scrypted has an update available for it
Nov 13 12:17:32 Helios nginx: 2024/11/13 12:17:32 [error] 32248#32248: *629 FastCGI sent in stderr: "Primary script unknown" while reading response header from upstream, client: 127.0.0.1, server: , request: "GET /plugins/gpustat/gpustatusmulti.php?gpus={%22%22:{%22id%22:%22None%22,%22model%22:%22None%22,%22vendor%22:%22None%22,%22guid%22:%22None%22,%22panel%22:1}} HTTP/1.1", upstream: "fastcgi://unix:/var/run/php-fpm.sock:", host: "localhost:8080", referrer: "http://localhost:8080/Dashboard"

AS it loks like your fastcgi docer is running host.

If a docker runs in host mode it will replace unraids port. if a web server like fast cgi, nginx etc is using port 443 and or port 80 this would be why you lose access to unriads web ui.

Not enoth info in the diag to assist.

what docker are you running and in what dockere network fashion?

I am running some dockers that use traditional web server ports, I guess I never considered that putting docker's network into host mode by default could have potential conflicts with the port my WebUI runs on (8080)

I will attempt to clear any port assignments that conflict with Unraid WebUI or other management ports, however I don't know if that would resolve the issues pertaining to ssh and manual console login not working?

Okay, so I'm currently utilizing macvlans for docker, which I am aware had some critical issues recently. None of the ports assigned to my management access conflict in the LAN:IP category but there is one conflict on the container port side, which I don't think should affect my webui, correct me if i'm wrong.

Any web service application show be running in ipvlan/macvlan (I prefer macvlan myself) with its own IP address via the network option [eth0,bond0,br0].
per your diag you were ruining 7 beta 3. which has the macvlan patch. you can run bridge mode.

atm it apears you have br0 briding and bond0 bonding off....
eth0 [192.168.0.110/24] wg0 [10.253.0.1/32] wg1 [10.253.1.1/32] wg2 [10.253.2.1/32]

Gotcha, I will change all web service applications to utilize bridge rather than host mode.

1 hour ago, dnglchlk said:

I am running some dockers that use traditional web server ports, I guess I never considered that putting docker's network into host mode by default could have potential conflicts with the port my WebUI runs on (8080)

I will attempt to clear any port assignments that conflict with Unraid WebUI or other management ports, however I don't know if that would resolve the issues pertaining to ssh and manual console login not working?

in terms of the console that due to new implementations. Setting > console Setting local screen blank time...



I'm not seeting any other panic or issues in the logs that would cause this outside of a web service docker breaking ports using host.

I will change those settings, and continue to monitor for any additional problems. If I find anything else that appears to be related I'll head back here with my findings, thanks for the assistance!

Well, unfortunately after I spun up the docker service again I had the issue occur again while attempting to open the GUI of a docker container that had just started. I have updated syslog attached. Is the 

sd 4:0:3:0: Power-on or device reset occurred

anything I should be worried about? I'm seeing it quite frequently 

syslog

Edited November 14, 20241 yr by dnglchlk
added file

I will power the system back on and keep all non-essential containers running until a fix is found