Complete webUI breakage after 6.12.0 update.

[tonyprindel]

Hi, everyone!

 

I recently updated to version 6.12.0. However, after updating, I seem to be completely unable to access my server's webUI. My local network generally has it sitting at 10.0.0.164:82 (I needed port 80 for something else). After this update, I'm unable to access it at either port 80 or port 82, including when I've started it in safe mode. I've also tried the GUI mode, both in normal and in safe mode, and I get a connection refused error from the built-in browser. This is all while I'm still able to ping out from the server when accessing the cli.

 

Given I'm unable to access the webUI at all, I'm not sure how to generate a debug log. Any guidance in the matter would be appreciated so I can provide more useful information. 

 

Thank you.

[JorgeB]

Type diagnostics in the CLI.

[bonienl]

Please also include the result of

# cat /etc/nginx/conf.d/servers.conf 

 

[tonyprindel]

Thanks for the quick replies. I'm attaching the requested logs and config.

 

blackstar-diagnostics-20230615-1009.zip servers.conf

[JorgeB]

@bonienlwill be better than me for this issue, but looks to me like port 445 set for SSL is already being used by something else:

 

Jun 15 10:08:43 Blackstar nginx: 2023/06/15 10:08:43 [emerg] 3876#3876: bind() to 127.0.0.1:445 failed (98: Address already in use)

 

[ljm42]

nginx: 2023/06/15 10:08:43 [emerg] 3876#3876: bind() to 127.0.0.1:445 failed (98: Address already in use)

 

nginx is unable to start SSL on port 445. When picking custom ports for things, it is best to use high numbers (over 1000, less than 64,000) that will never conflict with built-in ports.

 

I'd recommend carefully editing config/ident.cfg on your flash drive and changing:

PORTSSL="445"

to something higher, like:

PORTSSL="3443"

(assuming that isn't in use by a Docker container)

 

Once that is done you may be able to get the webgui by typing:

/etc/rc.d/rc.nginx reload

Or it may take a reboot.

[bonienl]

Ok. @ljm42 beat me.  

Please follow his advice and change the PORTSSL value.

 

[tonyprindel]

20 hours ago, ljm42 said:

nginx: 2023/06/15 10:08:43 [emerg] 3876#3876: bind() to 127.0.0.1:445 failed (98: Address already in use)

 

nginx is unable to start SSL on port 445. When picking custom ports for things, it is best to use high numbers (over 1000) that will never conflict with built-in ports.

 

I'd recommend carefully editing config/ident.cfg on your flash drive and changing:

PORTSSL="445"

to something higher, like:

PORTSSL="3443"

(assuming that isn't in use by a Docker container)

 

Once that is done you may be able to get the webgui by typing:

/etc/rc.d/rc.nginx reload

Or it may take a reboot.

 

This did it for me! That's a lesson learned, and I appreciate all of you for helping me out.

[Olegs]

On 6/15/2023 at 8:30 PM, ljm42 said:

nginx: 2023/06/15 10:08:43 [emerg] 3876#3876: bind() to 127.0.0.1:445 failed (98: Address already in use)

 

nginx is unable to start SSL on port 445. When picking custom ports for things, it is best to use high numbers (over 1000) that will never conflict with built-in ports.

 

I'd recommend carefully editing config/ident.cfg on your flash drive and changing:

PORTSSL="445"

to something higher, like:

PORTSSL="3443"

(assuming that isn't in use by a Docker container)

 

Once that is done you may be able to get the webgui by typing:

/etc/rc.d/rc.nginx reload

Or it may take a reboot.

It turns out that was indeed the culprit causing the problem. Thanks to your advice, everything is now functioning smoothly, and I can confidently reboot the server without any issues. I had initially mentioned the 445 port problem in another thread, and your input was invaluable in resolving it.

[DCJona]

i have the same issue (no access gui) i checked the plugin but it seems ok ... 

what can i upload to get help? 

 

found this 

Edited June 17, 2023 by DCJona

[bonienl]

Please start a new support ticket. This is a completely different problem you have.

Your USB device is corrupted.

 

[michaelmurfy]

On 6/16/2023 at 5:30 AM, ljm42 said:

nginx: 2023/06/15 10:08:43 [emerg] 3876#3876: bind() to 127.0.0.1:445 failed (98: Address already in use)

 

nginx is unable to start SSL on port 445. When picking custom ports for things, it is best to use high numbers (over 1000) that will never conflict with built-in ports.

 

I'd recommend carefully editing config/ident.cfg on your flash drive and changing:

PORTSSL="445"

to something higher, like:

PORTSSL="3443"

(assuming that isn't in use by a Docker container)

 

Once that is done you may be able to get the webgui by typing:

/etc/rc.d/rc.nginx reload

Or it may take a reboot.

Thanks for this. I had the same problem post upgrade however was getting the following:

Message from syslogd@Pikachu at Jun 18 01:54:05 ...
 nginx: 2023/06/18 01:54:05 [emerg] 20031#20031: bind() to [2406:]:80 failed (99: Cannot assign requested address)

(Note - masked IPv6 address) - the strange thing here is despite any "port in use" error I decided to update "PORT="80"" to something else, restart nginx and it worked. Changed it back to Port 80, stopped / started nginx and it is now working. I am not confident that it'll survive a reboot without manual intervention however but something is getting caught up here especially with servers with IPv6. But for this boot at-least it is a workaround to get the web interface up and running.

Edit - may be related to a potential wider issue:

 

Edited June 17, 2023 by michaelmurfy

[muihiuwev]

On 2023/6/17 at PM10点05分, michaelmurfy said:

谢谢你。升级后我遇到了同样的问题，但得到以下信息：

（注意 - 屏蔽的 IPv6 地址） - 这里奇怪的是，尽管有任何“端口正在使用”错误，我还是决定将“PORT =“80””更新为其他内容，重新启动 nginx 并且它起作用了。将其改回端口 80，停止/启动 nginx，现在它可以工作了。我不相信它会在没有手动干预的情况下重新启动后继续存在，但是这里出现了一些问题，特别是对于具有 IPv6 的服务器。但对于这次启动来说，至少这是一个让 Web 界面启动并运行的解决方法。

编辑 - 可能与潜在的更广泛的问题有关：

 

Some problem on 6.12.2

[Jake G]

I am running into similar issues after upgrading to 6.12.4. All plugins removed or updated. As soon as the array is started in Maintenance Mode, Normally, with docker on and off, etc. The only way I am able to access the webGUI after array start is within safe mode without plugins. I thought I'd held off long enough to upgrade to avoid breaking issues, but I guess not. 

I was able to get it working initially. I installed the new appdata.backup plugin. Halfway through the first backup is when issues began. Docker runs in safe mode. It is something in the plugins. I already attempted altering the SSL port suggested above. Can provide full syslogs if needed. Right now managing to get by without plugins but I can't stay this way. Shortly after Party tips & tweaks starts and you see remote shares mounting is the point where the WebGUI dies.

 

 

servers.conf ungootmedia-diagnostics-20231025-1507.zip ungootmedia-diagnostics-20231025-1510.zip

Edited October 25, 2023 by Jake G

[Jake G]

34 minutes ago, Jake G said:

I am running into similar issues after upgrading to 6.12.4. All plugins removed or updated. As soon as the array is started in Maintenance Mode, Normally, with docker on and off, etc. The only way I am able to access the webGUI after array start is within safe mode without plugins. I thought I'd held off long enough to upgrade to avoid breaking issues, but I guess not. 

I was able to get it working initially. I installed the new appdata.backup plugin. Halfway through the first backup is when issues began. Docker runs in safe mode. It is something in the plugins. I already attempted altering the SSL port suggested above. Can provide full syslogs if needed. Right now managing to get by without plugins but I can't stay this way. Shortly after Party tips & tweaks starts and you see remote shares mounting is the point where the WebGUI dies.

 

 

servers.conf 938 B · 0 downloads ungootmedia-diagnostics-20231025-1507.zip 136.09 kB · 0 downloads ungootmedia-diagnostics-20231025-1510.zip 118.94 kB · 0 downloads

It was the tailscale plugin. Even with it configured correctly for 6.12. Removing it allowed me to use my server normally, reinstalling it immediately crashed the webui. 

 

Guess I'll wait and see what happens next. 

[EDACerton]

FYI -- I released an update for the Tailscale plugin a couple days ago that fixes this issue. It took longer than I would have liked , but I was able to finally sort out what was causing issues for folks.