[08:54] <jo-erlend> I'd like to know what exactly must be reset on a new VM image that's derived from a base image. I can't find any documentation on it. In addition to obvious stuff like hostname, I've found that /etc/machine-id and /var/lib/dbus/machine-id has to be changed. Is there more?
[10:17] <SupaYoshi> help
[10:17] <SupaYoshi> how do i reinstall netplan from iso
[10:18] <SupaYoshi> i removed netplan by accident, and ccannot resolve apt-get
[10:19] <mybalzitch> have you tried configuring the network manually then using apt get?
[10:20] <SupaYoshi> i noticed that netplan isnt recognized.
[10:20] <SupaYoshi> im on ubuntu 18.04.x lts
[10:20] <SupaYoshi> mybalzitch, how can i do so?
[10:20] <mybalzitch> otherwise you'd need to find the netplan .deb on the iso and install it with the dpkg command
[10:21] <mybalzitch> ifconfig and route and possibly an entry in /etc/resolv.conf
[10:22] <SupaYoshi> will these work without netplan yea?
[10:22] <SupaYoshi> lets test!
[10:22] <SupaYoshi> yh ifconfig is working
[10:23] <SupaYoshi> mybalzitch: can you please guide me through these steps?
[10:23] <SupaYoshi> my ethernet adapter is called ens18
[10:23] <SupaYoshi> https://danielmiessler.com/study/manually-set-ip-linux/#ifconfig
[10:23] <SupaYoshi> that good?
[10:27] <SupaYoshi> https://prnt.sc/vsr6dc k soo
[10:27] <SupaYoshi> i got it back online and installed netplan
[10:27] <SupaYoshi> but its saying this
[10:32] <mybalzitch> err
[10:32] <mybalzitch> you want netplan.io
[10:34] <SupaYoshi> ipcheeers
[10:35] <SupaYoshi> installing
[10:35] <SupaYoshi> can i just type netplan apply now yh?
[10:35] <SupaYoshi> i checked my config
[10:35] <SupaYoshi> and reboot
[10:49] <juliank> cpaelzer:
[10:49] <juliank> Preparing to unpack .../13-qemu-system-gui_1%3a5.1+dfsg-4ubuntu2_amd64.deb ...
[10:49] <juliank> cp: -r not specified; omitting directory '/var/run/qemu/Debian'
[10:49] <juliank> dpkg: warning: old qemu-system-gui:amd64 package pre-removal script subprocess returned error exit status 1
[10:49] <juliank> dpkg: trying script from the new package instead ...
[10:49] <juliank> dpkg: error processing archive /tmp/apt-dpkg-install-ltGa5Q/13-qemu-system-gui_1%3a5.1+dfsg-4ubuntu2_amd64.deb (--unpack):
[10:49] <juliank>  there is no script in the new version of the package - giving up
[10:50] <juliank> trying to remove it:
[10:50] <juliank> Removing qemu-system-gui:amd64 (1:5.1+dfsg-4ubuntu1) ...
[10:50] <juliank> rm: cannot remove '/var/run/qemu/Debian': Is a directory
[10:51] <juliank> I could now rmdir the dir manually and remove&reinstall it
[10:51] <juliank> but this is weird stuff
[11:28] <cpaelzer> juliank: it is weird stuff, but fixed - https://bugs.launchpad.net/ubuntu/+source/qemu/+bug/1905377
[11:28] <cpaelzer> juliank: yet we can't fix the past - so anyone on ubuntu1 needs to do the rm manually to get through
[11:28] <cpaelzer> not a problem eventually for the release, but for anyone running -dev right now it is
[11:30] <cpaelzer> juliank: this initially happened to qemu-block-extra but the issue in qemu-system-gui should be the same - never the less in bug 1906245 I'm working with a report sorting out if there might be something else to fix
[11:34] <juliank> cpaelzer: Can you write that to ubuntu-devel list so people know what's going on?
[11:35] <juliank> Also you could fix it I suppose, as dpkg said, if it had a prerm script in the new package it would use that
[11:36] <juliank> though the bug is about postrm not prerm
[11:37] <cpaelzer> juliank: are you saying if the new PKG would have an "empty" prerm then it would use that new prerm and not the one of the old package ?
[11:37] <juliank> old-prerm upgrade new-version
[11:37] <cpaelzer> I always read "upgrading" in https://wiki.debian.org/MaintainerScripts as if it would always start with the "old" prerm being called
[11:37] <juliank> If the script runs but exits with a non-zero exit status, dpkg will attempt:
[11:38] <juliank> new-prerm failed-upgrade old-version
[11:38] <cpaelzer> uh thanks for bringing that to my mind, I never hit that little detail
[11:38] <cpaelzer> that will help
[11:39] <juliank> you can see that in the graph for upgrading where it goes right first, and then can go back left
[11:39] <cpaelzer> sure I see it in the graph, it just wasn't in my mind-cache
[11:40] <cpaelzer> that also explains why it isn't an issue for qemu-block-extra as there we still need some prerm and that won't RC!=0 for failed-upgrade
[11:40] <cpaelzer> I can fix things up
[11:40] <cpaelzer> and once working will send a mail
[13:04] <jo-erlend> I'd like to know what exactly must be reset on a new VM image that's derived from a base image. I can't find any documentation on it. In addition to obvious stuff like hostname, I've found that /etc/machine-id and /var/lib/dbus/machine-id has to be changed. Is there more?
[18:43] <jaguardown> Could someone help me? Following the Ubuntu server documentation, I'm trying to create this simple bridge for a KVM VM: https://pastebin.com/iWTG8tuU this is exactly how they tell you to do it. It works perfectly after running "sudo netplan apply" but if I reboot the KVM host it breaks the internet connection. The only thing it can do is ping the gateway. Other LAN and WAN IP's are unreachable.
[18:44] <jaguardown> I have to delete the bridge to restore connectivity
[18:45] <jaguardown> sorry had a mistake in the namservers, here's the correction: https://pastebin.com/qRhKAYC8
[19:34] <SupaYoshi> best way to delay docker till a certain service has started?
[19:34] <SupaYoshi> (mergefs)
[20:15] <sarnold> jaguardown: you don't put nameservers in the 'search', you put domain names in the 'search list'
[20:16] <sarnold> jaguardown: eg you put 'jaguardown' in there, and then when a user types in 'www.google.com', the dns service will first look for 'www.google.com.jaguardown.' and if that fails, then it will look for 'www.google.com.'
[20:48] <jaguardown> sarnold: roger, I'll fix that, do you think that's related to my current problem?
[20:48] <sarnold> jaguardown: no idea there :)
[20:49] <jaguardown> sarnold: it works until I reboot. I can ping gateway but no other IP addresses. If I ping a site like google.com with IPV6, that works. If I delete the bridge with "sudo networkctl delete br0" and then "sudo netplan apply" it can ping IPV4 IP addresses but I still can't ping LAN IPs and I can't ssh to it. I have to delete the bridge and leave it to resume full connectivity. So weird!
[20:50] <sarnold> jaguardown: one common issue with bridges is that you assign an IP to the *bridge* and when you add a nic to the bridge, the nic stops niccing, it starts bridging... I don't understand the whole thing, but I see it come up from time to time
[20:52] <jaguardown> hmmm
[20:54] <jaguardown> yeah I did notice in dmesg that the ethernet interface enters promiscuous mode which I thought was really strange. I don't want it to do that lol
[22:39] <jaguardown> Okay guys got it figured out
[22:41] <jaguardown> The problem was with dhcp vs static ip and it was partially because I didn't really understand how a bridge works. I searched and found a better tutorial that explained you need to assign a static ip to the bridge and turn off dhcp for the ethernet interface, and it also explained how to configure KVM/QEMU to use the bridge. All good now!
[22:44] <sarnold> jaguardown: excellent! :D
[23:16] <jaguardown> sarnold: :D