Install script fails - maybe nameserver/dns
-
I have been trying to install cloudron on my homeserver for a while now without luck, please help.
My homeserver is running linux and I have setup a quickemu ubuntu-server vm for running cloudron.
When I run the install/setup script it just looses internet connection. ping no longer works, dns resolve also stops working.
The issue is reproducible quite easily. I will add below the log output.
Same issue with the full ubuntu 22.04 distro in the same setup.Thanks
Link: https://upload.disroot.org/r/8uWyAJKq#CwiwLtiMqIQj+bx+oIymlaUADpO52IJEe7UaVCRSMqA=
echo "==> Install collectd" ==> Install collectd # without this, libnotify4 will install gnome-shell apt-get install -y libnotify4 libcurl3-gnutls --no-install-recommends Reading package lists... Building dependency tree... Reading state information... libcurl3-gnutls is already the newest version (7.81.0-1ubuntu1.15). libcurl3-gnutls set to manually installed. Suggested packages: gnome-shell | notification-daemon The following NEW packages will be installed: libnotify4 0 upgraded, 1 newly installed, 0 to remove and 0 not upgraded. Need to get 20.3 kB of archives. After this operation, 67.6 kB of additional disk space will be used. Ign:1 http://be.archive.ubuntu.com/ubuntu jammy-updates/main amd64 libnotify4 amd64 0.7.9-3ubuntu5.22.04.1 Ign:1 http://be.archive.ubuntu.com/ubuntu jammy-updates/main amd64 libnotify4 amd64 0.7.9-3ubuntu5.22.04.1 Ign:1 http://be.archive.ubuntu.com/ubuntu jammy-updates/main amd64 libnotify4 amd64 0.7.9-3ubuntu5.22.04.1 Err:1 http://be.archive.ubuntu.com/ubuntu jammy-updates/main amd64 libnotify4 amd64 0.7.9-3ubuntu5.22.04.1 Temporary failure resolving 'be.archive.ubuntu.com' E: Failed to fetch http://be.archive.ubuntu.com/ubuntu/pool/main/libn/libnotify/libnotify4_0.7.9-3ubuntu5.22.04.1_amd64.deb Temporary failure resolving 'be.archive.ubuntu.com' E: Unable to fetch some archives, maybe run apt-get update or try with --fix-missing?
-
@OrezBodajoy this happens if unbound is not able to make DNS requests directly from the server. Can you please check if
systemctl status unbound
andhost www.cloudron.io 127.0.0.1
work? If not, that is the source of the problem.We are aware of this problem, it's not entirely clear why this happens though. Are outbound DNS requests blocked in your home network?
A workaround :
- Use the forward everything configuration in https://docs.cloudron.io/troubleshooting/#unbound
- Check if
host www.cloudron.io 127.0.0.1
works after - does
cloudron-setup --redo
work after above?
-
-
Thanks for your reply @girish
unbound was active and running and the host command was not working.
I did the forward everything setup but that did not fix the issue. btw only after the forward everything change did I start to get the SERVFAIL responses to the host commmand; before that it was just the same reply as when unbound is not installed at all.Fortunately I just tried deactivating my VPN on the host machine and it worked! The installation was able to proceed.
In my case I will need to re-activate the vpn so it might not be a complete solution -
Please let me know if you have any ideas on what settings on the VPN I should look into.
Split tunnelling is not really an option but I will also investigate that.
Thanks! -
@OrezBodajoy said in Install script fails - maybe nameserver/dns:
Fortunately I just tried deactivating my VPN on the host machine and it worked! The installation was able to proceed.
In my case I will need to re-activate the vpn so it might not be a complete solutionYou have network related problems when installing and omit THAT information?!
-
@necrevistonnezr yeah sorry, at first I thought I had described everything and then I was sure it couldn't be the VPN as it was running fine before the cloudrun install. I didn't know there was a name server in cloudrun.
-
The setup works fine with split tunneling on the host machine for the cloudron VM.
My understanding is that right now my ISP can see all my cloudron traffic.Do you know a way to avoid that?
Thanks! -
@OrezBodajoy with https, ISP cannot see Cloudron traffic. Or did you mean something else?
-
I was thinking of hiding even the hosts (and IP ) part of the traffic (with a vpn) but it's ok I will check that later on.
Right now I setup a reverse proxy to send the 80 and 443 ports to a vps on ports 80 and 443.
I can access the cloudron setup dashboard using the vps IP address, but on the domain setup page, when I enter my domain I get an error:
Configuration error: Domain resolves to ["XXX.XXX.XXX.XXX"] instead of IPv4 XX.XX.XX.XX
where XXX is my remote vps server IP address and XX is my home internet's IP address.
I think the setup is ok but that check is failing and I don't know how to move forward.[The goal here was to host a large amount of data that's why I want to host it locally instead of directly on the vps]
Thanks.
-
I wanted to let you know how I got everything to work, it might help other users.
The initial issue was about DNS resolution. The quick fix is indeed to use the forward-everything option in unboud but with the previously used dns server. To find that value you just need to do a
resolvectl
and look at the value of the Current dns servers. In my case it was10.0.2.3
(the one that was linked to my vpn on the host machine of the server). That is enough to fix the installation script losing internet.For the
Configuration error: Domain resolves to ["XXX.XXX.XXX.XXX"] instead of IPv4 XX.XX.XX.XX
the quick fix in my case (home server not directly exposed to the internet) was to install an openvpn server on the vps server (https://www.cyberciti.biz/faq/ubuntu-22-04-lts-set-up-openvpn-server-in-5-minutes/) and an openvpn client on my home server so that my home server gets the vps's IP address. In this setup you I also need a reverse proxy to expose the 80 and 443 ports from the home server to the vps. -