Azuracast: Streaming to an external icecast2-server fails with hostname but not IP

[MrSteelMiner]

I tried to connect and stream to an external icecast2-server but it's not working. What have I to do?
My external icecast2-server is running v2.4.2 on port 80.
Mount-Config:
<mount type="normal"> <mount-name>/dev1</mount-name> <username>dev1</username> <password> _HIDDEN_ </password> <relay-shoutcast-metadata>1</relay-shoutcast-metadata> <!--<max-listeners>300</max-listeners>--> <!--<intro>/intro.ogg</intro>--> <fallback-mount>/laut_sr</fallback-mount> <fallback-override>1</fallback-override> </mount>
Streaming from butt-software is working to this mount.

In azuracast-webinterface I use the following config:
Remote Station Listening URL: http://name.domain:80 Remote Station Listening Mountpoint/SID: /dev1 Broadcast AutoDJ to Remote Station: Yes AutoDJ Format: MP3 AutoDJ Bitrate (kbps): 320 Remote Station Source Port: 80 Remote Station Source Mountpoint/SID: _empty_ Remote Station Source Username: dev1 Remote Station Source Password: _HIDDEN_ YP-directory: false

I restarted the station multiple times (update-installation and so on), but azuracast isn't connecting to icecast.
Liquid-Soap-Config (Log Viewer):
output.icecast(%mp3(samplerate=44100, stereo=true, bitrate=320, id3v2=true), id="party_relay_2", host = "name.domain", port = 80, user = "dev1", password = "_HIDDEN_", mount = "/dev1", name = "StayRadio Party", description = "Dein Party-Sender", genre = "Party", url = "https://stayradio.de", public = false, encoding = "UTF-8", radio)

Note:
_HIDDEN_ is of course replaced by the real password, same for name.domain

Thanks in advance!

[MrSteelMiner]

Liquidsoap Log:
2019/03/14 14:07:49 [party_relay_1:2] Connection failed: Failure("inet_addr_of_string") 2019/03/14 14:07:49 [party_relay_1:3] Will try again in 3.00 sec. 2019/03/14 14:07:49 [party_relay_2:3] Connecting mount /dev1 for [email protected]... 2019/03/14 14:07:49 [party_relay_2:2] Connection failed: Failure("inet_addr_of_string") 2019/03/14 14:07:49 [party_relay_2:3] Will try again in 3.00 sec.

As said: Streaming through butt-software with this config is working

[MrSteelMiner]

seen this https://github.com/savonet/liquidsoap/issues/496
I will try using the IP and not the hostname

**EDIT
using the IP instead of hostname is working for me.

[Vaalyn]

This issue is related to the same underlying issue as #756 which was already fixed for the docker installations by an OPAM update a while ago. Around 3 days ago there was an update to the ansible scripts to use OPAM 2. Which version of AzuraCast are you currently using?

[MrSteelMiner]

latest version. Still not working with hostname

[Vaalyn]

Could you check which OPAM version you've installed?

[MrSteelMiner]

opam --version
2.0.2

[preterive]

Sam Problem, is a bug of liquidsoap i think. Error: inet_addr_of_string

It was already there in 1.3.3 and got fixed in 1.3.3-1 .

[preterive]

@MrSteelMiner is this a laut.fm Station too? Please watch my thread in the laut.fm Forum

[MrSteelMiner]

Yes, it's a laut.fm station. But it doesn't work with my own icecast either. So I use the IP as workaround and so I can live with it

[Vaalyn]

@preterive AzuraCast should already be using Liquidsoap v1.3.6 meaning this issue should be fixed already. If you have any additional information regarding this issue can you link that to us?

@MrSteelMiner What version of Liquidsoap is running on your installation? (liquidsoap --version)

[MrSteelMiner]

@Vaalyn 1.3.6

[preterive]

Hello,
i have the same problem like @MrSteelMiner since i updated to Version 1.3.6

[Vaalyn]

Okay, and you both are using Laut.fm, correct?

Can you test if this problem may be related to Laut.fm and if that feature works correctly with other external relays?

[MrSteelMiner]

@Vaalyn I tested it with a normal icecast and this also doesn't work (tested it 2 weeks ago)

[Vaalyn]

@MrSteelMiner Thank you for the information. Then it's pretty clear that the general functionality has a problem that is not caused by a single 3rd party.

[SlvrEagle23]

@MrSteelMiner @preterive What installation type are you all using?

This could be caused by a failure of either the host OS or container OS to resolve a domain name. With Docker, DNS resolution should always point to your host's DNS resolver configuration, which you may need to update to reference an external DNS resolver (i.e. 1.1.1.1 for CloudFlare or 8.8.8.8 for Google)

[MrSteelMiner]

the host (traditional installation) resolves the hostname correctly (ping).

[SlvrEagle23]

@MrSteelMiner @preterive @Vaalyn Figured this one out. When LS fixed it, the fix was actually with the cry OCaml library, not with LS itself. They published 0.6.1 as a fix, but we hard-lock our dependencies on OPAM to avoid unexpected update issues, and we had locked to 0.6.0, which didn't have the fix.

Updating both the Docker and bare-metal installations now to resolve that.

[SlvrEagle23]

@MrSteelMiner @preterive @Vaalyn Docker image is now rebuilding on the Docker Hub; if you're on bare-metal you can run the updater script now and it should bring you to the newest version, which should fix the issue.

[preterive]

Thanks! I will try it now.

[preterive]

Works, could be closed.

[SlvrEagle23]

Good to hear. New Docker image has finished building and is live as well.