Hello folks,

I am desperate for your help and expertise. I'm on a 7.0,1 build, approx. 12 drives, have had no real problems other than having to force my boot to usb every time it boots (I know that's not an unraid problem.)

Server is doing it's thing, nothing going on, it's primarily a media server, and a place that I can test things out or try to improve my families digital encroachment in life.

Randomly, we cannot access the server - typically this is alerted in plex or trying to access sonarr and radarr. I am unable to get the server to respond with the keyboard hooked to it - no error messages at all on the attached lcd. I do just see the login prompt at the bottom of my screen, yet unraid is unreachable by name or ip, unpingable.

I believe I have the right logs setup - it truncated the prior log right at the error, and then started this new one. If it's not right, let me know and I'll try to correct that so the next time it happens I can post.

syslog-previous is what i believe is the log prior to the crash, the culprit is in here - but idk. syslog is the newest one since i restarted the server, maybe that will help.

You also will see some containers that are shut off - nginxproxymananger and things related to a reverse proxy I had setup but I broke something with it so I just disabled it all.

I am sincerely grateful for any assistance or ideas.

syslog-previous syslog

Edited June 29, 20251 yr by OnlyDads

Due to a serious data wiping bug I would advise you upgrade off v7.0.1 and into the lattest 7.0,4


are you runnign a reverse proxy?

Here's what stands out in the final portion of the syslog-previous before the system lockup:

rc.nginx restarting every 2 minutes:

Jun 28 01:08:39 Tower rc.nginx: Nginx server daemon... Started.

Jun 28 01:10:39 Tower rc.nginx: Stopping Nginx server daemon forcibly...

...

(this repeats every 2 minutes, consistently)

The nginx server is being forcibly stopped and restarted every ~2 minutes. This is abnormal and strongly suggests a script, container, or failing health check causing repeated restarts.

emhttpd: error: publish, 177: Connection reset by peer (104)

Jun 28 01:19:32 Tower emhttpd: error: publish, 177: Connection reset by peer (104): read

??? Is the unraid connect plugin installed? there have been some known issues with it and casuing this kind of behavior...

This means Unraid’s web UI (emhttpd) lost connection to its internal publishing mechanism — very likely due to the constant nginx restarts, or underlying network/socket issues.

No crash trace or kernel panic shown:
The log does not include a panic, OOM kill, or kernel error, which suggests the system froze in a hard state — possibly hardware or kernel deadlock rather than a soft failure.

but acts as the emhttp / https nginx system is breaking and thus the managment web interface is dead... youve reported that even at terminal it apears frozen and unable to take terminal caomdns.

The beginning of the new syslog shows a normal Unraid boot sequence:

Microcode and kernel version:

Jun 28 18:48:21 Tower kernel: microcode: updated early: 0xf6 -> 0xfc

Jun 28 18:48:21 Tower kernel: Linux version 6.6.78-Unraid

Normal RAM map and hardware detection:

BIOS-e820: ... usable/reserved memory regions ...

NX (Execute Disable) protection: active

Nothing unusual or faulty in the physical RAM mapping. No hardware or CPU flags are failing at boot. That I can see / find...

###########

Some Potential Root Causes

• Misconfigured reverse proxy:
  You mentioned Nginx Proxy Manager was disabled, but nginx is still running and forcibly restarted — you may still have services trying to talk through a broken proxy config.

• System resource exhaustion (socket/table limits, file descriptors, memory):
  If nginx or emhttpd are repeatedly crashing, it may indicate descriptor/resource limits being hit.

• Docker container script or cron task restarting nginx blindly on a loop.

• Hardware instability after extended uptime (45+ days):
  Could be overheating, memory corruption, or board-level instability (e.g. RAM or USB controller glitches).

If reverse proxy ...

check for nginx or docker loop:

• Look at /etc/cron.*, /boot/config/go, and any custom docker-compose or plugin that might auto-restart nginx.

• Disable nginx and nginx-proxy-manager fully:

docker ps -a

docker stop <nginx container>

docker rm <nginx container>

Inspect logs further up (before the nginx loop):
I can do that next if you'd like.

System health:

• Run a memtest overnight from Unraid’s boot menu.

• Review BIOS settings (disable secure boot, set USB boot priority, enable legacy mode).

• Review cooling/temps (sensors, IPMI if available).

Wow I just wanted to thank you for alerting me about the bug - I will go through the update process.

I've gotta get a few minutes to go over your post with a fine tooth comb but I need to figure out why is nginx running at all - so I will work my way through the outline provided.

I never considered motherboard/ram/platform bug, but I suppose this could be the case as well - although it's all brand new and ran fine for a few years (i know things break) - but once I eliminate this nginx gremlin I will wait and see if it happens again, and then if needed swap memory/cpu/mobo in that order (cost and likelyhood of failure - although I probably should do motherboard before the chip but IDK) Sometimes I think I woudl prefer it to be one of those problems I can just throw money at and it goes away - but I actually kind of look forward to this - definitely causes some more gray hairs!

If you have a paypal/venmo/etc you can DM me and I would like to buy you a beer or a 12 pack of pop or something. I really am very grateful for your time and willingness to look into this for me. I was extremely anxious about this process because I thought I was going to get literally murdered in the forums like I'm dialing into AOL Chat rooms lol.

Thank you!

by default unraid runs a nginx for emhttp for nginx use and ssl and https access to teh web managemnt... unraid also has a nginx proxy manger in tools section.