[6.12] Unraid webui stop responding then Nginx crash

6.12.9, crashed too.

On 4/5/2024 at 8:36 AM, Tahuwa said:

I can confirm, 6.12.10, also crashing nginx, server unresponsive and the only way to get it back to life is a hard reboot, which then triggers a parity scan

Came here to say literally the same thing - same experience, same steps, same results.

I've had the same error for two days and it's driving me crazy. I deactivated all Docker except for a few that have been installed for ages because I thought that was the reason. Now I came across this thread after studying the syslog and am even more frustrated to see how long the error has existed. Especially since I just upgraded from plus to pro for $69. I'm running the latest stable version. All my smart home applications run in unraid, now my house is stupid and the heating stays cold, luckily it's going to be summer 🤣 .

Paying customers get no support on a question/issue for over a year.   Nice.

 

I can attest to this issue myself.  6.12.10. 

OK, I think I was able to solve the problem for me. I replaced the USB stick (including license transfer), transferred the config from the old stick to the new one and uninstalled NerdTools. Since then the system has been running without any problems. I'll wait how long. 😅

I updated to 6.12.10 and started seeing the below filling up my logs a couple of days later.

 

Attached is a diagnostic file. 

 

I don't seem to have lost the GUI but couldn't check while it was happening.

 

A minor note is I noticed when I requested the diagnostics file from my windows client it left the downloaded file named .crdownloaded for whatever reason.  I renamed it .zip and took a look and it seems ok.  I was able to reproduce this. Maybe a client side issue?

 

May 21 22:36:30 raid-01 nginx: 2024/05/21 22:36:30 [alert] 9145#9145: worker process 15029 exited on signal 6
...
May 22 04:16:37 raid-01 nginx: 2024/05/22 04:16:37 [alert] 9145#9145: worker process 30337 exited on signal 6

raid-01-diagnostics-20240522-1654.zip

9 hours ago, dabl said:

I updated to 6.12.10 and started seeing the below filling up my logs a couple of days later

See if this helps with that error:

https://forums.unraid.net/topic/86114-nginx-running-out-of-shared-memory/?do=findComment&comment=1417198

 

7 hours ago, JorgeB said:

See if this helps with that error:

https://forums.unraid.net/topic/86114-nginx-running-out-of-shared-memory/?do=findComment&comment=1417198

 

Thanks for the reply and info, greatly appreciated! 

 

Looking over my notes I realized I had seen this before and it was possibly exacerbated by having 2 browser tabs both open to raid-01, one to the syslog (http://<servername>/Tools/Syslog) and/or an unraid browser tab becoming 'stale'? I have been doing that since the update to look for issues in the syslog.  Will discontinue that and check the log in a different way in addition to pursuing the nchan_shared_memory_size thing. 

Edited May 23 by dabl

50 minutes ago, dabl said:

it was possibly exacerbated by having 2 browser tabs both open to raid-01, one to the syslog (http://<servername>/Tools/Syslog) and/or an unraid browser tab becoming 'stale'?

That can also cause issues, and for some users, even having just one browser window opened to the GUI can do it.