Why does AssetUPnP listen on port 80?

[riwi]

Hello,
I am using AssetUPnP under slackware Linux without major issues for a few years already. Now recently I found that when the Linux server is rebooted, the Appache httpd complains it can't connect with port 80 because another process is listening on port 80. This process is AssetUPnP.
I have now upgraded from 5.1.1 beta to version 6.7

* netstat -ntulp
Active Internet connections (only servers)
Proto Recv-Q Send-Q Local Address Foreign Address State PID/Program name
tcp 0 0 192.168.1.11:80 0.0.0.0:* LISTEN 2831/AssetUPnP

root@riwixeon:/etc/rc.d* ./rc.httpd start
(98)Address already in use: AH00072: make_sock: could not bind to address [::]:80
(98)Address already in use: AH00072: make_sock: could not bind to address 0.0.0.0:80
no listening sockets available, shutting down
AH00015: Unable to open logs

The httpd starts late because the www folder is mounted quite late. So AssetUPnP is already started before httpd is.
I can fix this manually by killing AssetUPnP

* ps -ef | grep Asset
root 2482 1 0 20:05 ? 00:00:00 /home/richard/asset_upnp/bin/AssetUPnP/AssetUPnP --ipaddress 192.168.1.11
root 2489 2482 0 20:05 ? 00:00:00 /home/richard/asset_upnp/bin/AssetUPnP/AssetUPnP --fork-is-webconfig --ipaddress 192.168.1.11 --control-pipe 5
root 2490 2482 0 20:05 ? 00:00:00 /home/richard/asset_upnp/bin/AssetUPnP/AssetUPnP --fork-is-umedialibrary --assetid 0 --control-pipe 11
root 2831 2482 0 20:05 ? 00:00:00 /home/richard/asset_upnp/bin/AssetUPnP/AssetUPnP --fork-is-upnp --assetid 0 --ipaddress 192.168.1.11 --control-pipe 9
root 18011 4366 0 20:15 pts/0 00:00:00 grep Asset

kill 2482
kill 2490

Then I can start httpd :
* /etc/rc.d/rc.httpd start

And then AssetUPnP :
* /home/richard/asset_upnp/bin/AssetUPnP/AssetUPnP --ipaddress 192.168.1.11 &
[1] 17247
root@riwixeon:/etc/rc.d* Starting Asset web interface at http://192.168.1.11:45537
pid:17248 Entering mainloop
pid:17251 Entering mainloop
pid:17250 Entering media library mainloop

Why is AssetUPnP taking port 80? When I point a browser at is it will only show a blank page.
What is the best way to "kill AssetUPnP" when I need/want to? Just the kill on the 2 pid's?. Or is there a nicer shutdown script/command?
Can I prevent AssetUPnP from taking port 80?

Regards, Richard

[PeterP]

Thanks for the detailed bug report, the problem is being investigated.

In meanwhile, you can gracefully shut Asset down with --shutdown switch.

What happens if you open the spurious port80 server in a web brorwser?

Which exact Asset build is this? x86-64 or other?

[riwi]

Thanks for the detailed bug report, the problem is being investigated.

In meanwhile, you can gracefully shut Asset down with --shutdown switch.

What happens if you open the spurious port80 server in a web brorwser?

Which exact Asset build is this? x86-64 or other?

Thanks. I will use the --shutdown option in the future
root@riwixeon:~* /home/richard/asset_upnp/bin/AssetUPnP/AssetUPnP --shutdown
Asset shut down successfully.


Yes this is 64bit x86

root@riwixeon:~* uname -a
Linux riwixeon 4.4.227 *2 SMP Thu Jun 11 15:07:16 CDT 2020 x86_64 Intel(R) Xeon(R) CPU D-1518 @ 2.20GHz GenuineIntel GNU/Linux

OK, so I stopped httpd, started Asset again

root@riwixeon:~* netstat -ntulp | grep Asset
tcp 0 0 192.168.1.11:80 0.0.0.0:* LISTEN 2625/AssetUPnP
tcp 0 0 192.168.1.11:45537 0.0.0.0:* LISTEN 2622/AssetUPnP
udp 0 0 0.0.0.0:1900 0.0.0.0:* 2625/AssetUPnP


Now when I use a browser and connect with 192.168.1.11:80 it shows a blank page (not a timeout or 404 message)

[riwi]

I am not sure if my reply was lost. I don't see it here.

It is x86_64
uname -a
Linux riwixeon 4.4.227 *2 SMP Thu Jun 11 15:07:16 CDT 2020 x86_64 Intel(R) Xeon(R) CPU D-1518 @ 2.20GHz GenuineIntel GNU/Linux


/etc/rc.d/rc.httpd stop
/home/richard/asset_upnp/bin/AssetUPnP/AssetUPnP --shutdown
/home/richard/asset_upnp/bin/AssetUPnP/AssetUPnP --ipaddress 192.168.1.11 &

netstat -ntulp | grep Asset
tcp 0 0 192.168.1.11:80 0.0.0.0:* LISTEN 2625/AssetUPnP
tcp 0 0 192.168.1.11:45537 0.0.0.0:* LISTEN 2622/AssetUPnP
udp 0 0 0.0.0.0:1900 0.0.0.0:* 2625/AssetUPnP


pointing a browser at 192.168.1.11:80 just gives an empty white screen. No error or 404 message.

[PeterP]

Thanks for the details.
Is anything on your system taking port 26125? Because that's what Asset is supposed to be using, but apparently it's using port 80 instead for you for some reason.

[PeterP]

Please try:
https://www.dbpoweramp.com/beta/Asse...nux-x64.tar.gz

Apparently there was a bug with --ipaddress on systems that did not specifically disallow listening on port 80.

[riwi]

Thanks for the details.
Is anything on your system taking port 26125? Because that's what Asset is supposed to be using, but apparently it's using port 80 instead for you for some reason.

When I have http started first then Asset takes port 26125 instead :

root@riwixeon:/home/richard/scripts* netstat -ntulp | grep Asset
tcp 0 0 192.168.1.11:26125 0.0.0.0:* LISTEN 11381/AssetUPnP
tcp 0 0 192.168.1.11:45537 0.0.0.0:* LISTEN 11378/AssetUPnP
udp 0 0 0.0.0.0:1900 0.0.0.0:* 11381/AssetUPnP

[riwi]

Please try:
https://www.dbpoweramp.com/beta/Asse...nux-x64.tar.gz

Apparently there was a bug with --ipaddress on systems that did not specifically disallow listening on port 80.

I have tried 6.8. beta but It still has the same result.

While httpd is not running :
root@riwixeon:/home/richard/asset_upnp/bin/AssetUPnP* netstat -ntulp | grep Asset
tcp 0 0 192.168.1.11:80 0.0.0.0:* LISTEN 27375/AssetUPnP
tcp 0 0 192.168.1.11:45537 0.0.0.0:* LISTEN 27372/AssetUPnP
udp 0 0 0.0.0.0:1900 0.0.0.0:* 27375/AssetUPnP

Asset UPnP: riwixeon
Library Contains 2376 Tracks, 203 Albums
Version: R6.8 beta 1 [Linux-x64]; Functionality: Premium Registered

When I start httpd first it will take port 80. Then Asset takes 26125

root@riwixeon:/home/richard/asset_upnp/bin/AssetUPnP* netstat -ntulp | grep Asset
tcp 0 0 192.168.1.11:26125 0.0.0.0:* LISTEN 13104/AssetUPnP
tcp 0 0 192.168.1.11:45537 0.0.0.0:* LISTEN 13101/AssetUPnP
udp 0 0 0.0.0.0:1900 0.0.0.0:* 13104/AssetUPnP

I need to specify --ipaddress because this PC also functions as my router/firewall and has multiple interfaces. eth0 and eth1 towards the ISPs and eth2 (10Gbe) to my internal lan.

[riwi]

I made some sort of mistake with twestnig 5.8B1. Not sure what happened becasue I still saw 80 being used.
I did the INSTALL.sh again (while asset was not running. Then I moved the folder away from /root to where I want it.

* ps -ef | grep httpd
root 12693 1 0 20:46 ? 00:00:00 /usr/sbin/httpd -k start
apache 12694 12693 0 20:46 ? 00:00:00 /usr/sbin/httpd -k start
apache 12695 12693 0 20:46 ? 00:00:00 /usr/sbin/httpd -k start
apache 12696 12693 0 20:46 ? 00:00:00 /usr/sbin/httpd -k start
apache 12702 12693 0 20:46 ? 00:00:00 /usr/sbin/httpd -k start
root 21660 379 0 21:14 pts/2 00:00:00 grep httpd
root@riwixeon:/home/richard/asset_upnp/Asset-R6.8b1-Linux-x64* /etc/rc.d/rc.httpd stop

ps -ef | grep Asset
root 16772 1 0 21:11 pts/2 00:00:00 /root/bin/AssetUPnP/AssetUPnP
root 16775 16772 0 21:11 pts/2 00:00:00 /root/bin/AssetUPnP/AssetUPnP --fork-is-webconfig --control-pipe 7
root 16776 16772 0 21:11 pts/2 00:00:00 /root/bin/AssetUPnP/AssetUPnP --fork-is-umedialibrary --assetid 0 --control-pipe 14
root 16777 16772 0 21:11 pts/2 00:00:00 /root/bin/AssetUPnP/AssetUPnP --fork-is-upnp --assetid 0 --control-pipe 12
root 28231 379 0 21:18 pts/2 00:00:00 grep Asset
root@riwixeon:/home/richard/asset_upnp/bin/AssetUPnP* ./AssetUPnP --shutdown

* ./AssetUPnP --ipaddress 192.168.1.11 &
[1] 29302
root@riwixeon:/home/richard/asset_upnp/bin/AssetUPnP* Starting Asset web interface at http://192.168.1.11:45537
pid:29303 Entering mainloop
pid:29306 Entering mainloop
pid:29305 Entering media library mainloop

root@riwixeon:/home/richard/asset_upnp/bin/AssetUPnP* netstat -ntulp | grep Asset
tcp 0 0 192.168.1.11:26125 0.0.0.0:* LISTEN 29306/AssetUPnP
tcp 0 0 192.168.1.11:45537 0.0.0.0:* LISTEN 29303/AssetUPnP
udp 0 0 0.0.0.0:1900 0.0.0.0:* 29306/AssetUPnP

This time it worked OK and it took 26125

Sorry for the confusion. Thanks for the quick update

[PeterP]

Thanks for verifying. This was one interesting bug - apparently only manifesting itself if --ipaddress was specified and the system allowed Asset to take port 80, which wasn't the case on any of my test machines.
If port 80 was busy or unavailable, everything worked normally - 26125 was taken like it should be.