[04:40] <mica> i am having dificuties using netplan network configuration on ubuntu 18.04 server LTS, is there any graphical way of setting this up to make sure i am not doing typo errors in config? centos has nmtui for example thank you in advance for helping
[04:44] <mica> no one here can help ?
[06:11] <lordievader> Good morning
[06:29] <dpb1> mica: patience
[06:40] <mica> dpb1:  i do have it but as you can see it is a bit slow here :)  managed in the meantime to get server 16 lts installed and configured, it hink ubuntu 18 is sthill nbaking and should not be used in production unless properly documented and tested.
[07:18] <dpb1> mica: there is no graphical way, no.  but, if you put up a pastebin of your netplan yaml, someone would be happy to check.
[07:18] <dpb1> mica: also, askubuntu is a good place to type up what you are trying to do
[09:01] <blackflow> mica: yeah but who's gonna do the testing :) bugreporting from users is very valuable.
[14:14] <cyphermox> mica: there's 'netplan try' if you're unsure of config, and we put a lot of examples of common things to do up on http://netplan.io/examples
[14:44] <gunix> does anybo have any idea how to connect two VMs without using bridge? so nic2 from vm1 connected to nic3 from vm2...
[14:46] <compdoc> you can use the built-in virtual lan
[14:53] <sdeziel> gunix: it's still kind of a bridge but you can use MACVLAN
[14:54] <sdeziel> gunix: but could you elaborate as to why you don't want a regular bridge?
[14:55] <gunix> sdeziel: to connect multiple switches together, to simluate a DC
[14:57] <sdeziel> gunix: why wouldn't that work with regular bridges?
[15:10] <gunix> sdeziel: because you don't get lldp
[15:11] <sdeziel> gunix: found this: https://the-bitmask.com/2017/08/04/fwd-lldp-frames-on-linuxbridge/
[15:12] <sdeziel> gunix: that said, I never had to deal with lldp so YMMV ;)
[15:12] <gunix> sdeziel: it's ok, 3rd party blogs is usually the last stuff i reffer to :D
[15:12] <gunix> i will try this first:
[15:12] <gunix> -netdev socket,id=mynet0,listen=:1234
[15:12] <gunix> -netdev socket,id=mynet0,connect=:1234
[15:17] <sdeziel> gunix: I expect that will be slower but probably good enough for an experiment
[15:20] <sdeziel> gunix: the same without the blog bits: https://www.kernel.org/doc/Documentation/ABI/testing/sysfs-class-net
[16:48] <teward> I have a server that has a bunch of Public IPs on it.  From a NAT perspective, containers on the system use IPs as 1:1 IP ranges.  Unfortunately, this causes some issues, because right now the individual IPs can't talk to each other, and the IPs sit only on the NIC on the host (and attempts to communicate between each other return 'Connection Refused').  Is there a way to make the IPs communicate with each other?
[16:48] <teward> I should know this and it's something simple likely but...
[16:50] <blackflow> teward: sounds like you need to NAT only when dst is not in the subnet of those IPs
[16:51] <blackflow> though from what you describe, I fail to see what exactly is going on. You have a bunch of public IPs, but then NAT to containers' subnet?
[16:51] <teward> blackflow: i figured it out, it's a "split DNS" problem
[16:52] <teward> in that i need to serve the IPs for the boxes internally separate from the public
[16:52] <teward> so it's a NAT problem but it's not a NAT problem too
[16:52] <blackflow> if it's DNS then you can't talk about IPs, but hostnames
[16:52] <teward> blackflow: it's "both"
[16:52] <teward> DNS resolves to public IP
[16:52] <teward> but the containers sit on the same bridge/subnet
[16:52] <teward> so they need to talk with the 'internal' IP that is on the same subnet
[16:53] <teward> i had this issue before but overlooked this >.>
[16:53] <teward> i can fix it now, I know what to do *stabs the postfix configs on these boxes to handle the local delivery route differently*
[16:53] <blackflow> I still don't get it though, you have public IPs, and then still do NAT?
[16:53] <teward> blackflow: too hard to explain without my diagram - TL;DR I figured out a solution, I'll fix the actual problem 'later' once I have free time :P
[16:53] <teward> and will probably be back here.  It's probably a NAT problem but meh
[16:53] <blackflow> mkay.
[18:45] <gunix> sdeziel: thank you for the advice
[20:45] <BLZbubba> i'm trying the 18.04 text mode server installer.  i already pre-created the partitions on disk how I like them.  what does it take for the new installer to use them?
[20:46] <BLZbubba> the only options appear to be to wipe the disk and automatically partition, or to wipe the disk and manually partition
[20:46] <BLZbubba> but the manual partition option doesn't let me set up things like efi and swap the way I like them
[20:47] <BLZbubba> can i use the old installer instead?
[20:48] <sdeziel> BLZbubba: the alternate installer is still officially supported
[20:49] <RoyK> the "alternate" installer should be the official
[20:49] <RoyK> the new one sucks
[20:51] <RoyK> whoever came up with the idea of ditching all sorts of options like re-using partitions, creating raid, lvm, encrypted drives, whatever, should be giving a public speach about why they did it on an LTS release and allow for at least three hours for questions
[20:51] <RoyK> s/speach/speech/
[20:52]  * RoyK wonders if the S in LTS now means Suffering
[20:59] <BLZbubba> haha still only 1% as horrible as 10.04 was though
[21:00] <BLZbubba> ubuntu 10.04 was the worst release of any distro in history
[21:01] <BLZbubba> hmm do i need a new iso to switch installers?
[21:01] <BLZbubba> or should I try "OEM"
[21:28] <mason> BLZbubba: The installer you want is its own ISO.
[21:29] <mason> BLZbubba: https://www.ubuntu.com/download/alternative-downloads#alternate-ubuntu-server-installer
[22:06] <gbc> hi all ... can anyone please help me figure out a boot issue? i suspect i may have screwed up my init scripts
[22:08] <gbc> ... when a server boots, does it transition right into its default runlevel? or does it start in runlevel 1 and then transition to runlevel 2?
[22:09] <sarnold> both upstart and systemd have some amount of runlevel-ish stuff as compatibility, but I don't think either one *really* thinks in terms of runlevels
[22:10] <gbc> i've got a VM running 14.04 ... it boots partially and then hangs ... i never get a command prompt ... since i can't login, i'm trying to deduce how far it gets in the boot process (or which job might be hanging)
[22:11] <gbc> ... i thought maybe if i could force it into singleuser mode, maybe i'd be able to login and selectively enable services
[22:11] <sarnold> good idea, can you get to the kernel command line and add 'single' to it? I *think* that ought to work..
[22:12] <gbc> problem is, i can't login to the VM ... so, i can't edit grub directly ...
[22:12] <gbc> i *can* mount the guest's qcow image in the host and edit its filesystem that way ...
[22:13] <gbc> but update-grub seems to assume that it reads/writes to specific location ...
[22:13] <gbc> can't figure out how to trick it to overwrite the guest VM's boot command
[22:13] <sarnold> if you've got console access to it you might be able to use a left-shift key to get to a prompt or menu or similar
[22:14] <gbc> it's a VM ... i can do "virsh console" to get to the serial console ... i see some services starting, and then it hangs ... never get a login prompt :(
[22:14] <gbc> ... network starts (i can ping it), but it doesn't get far enough to start sshd/telnet :(
[22:18] <gbc> sarnold... gtg in a minute...thanks for the time & sharing your thoughts
[22:19] <sarnold> gbc: good luck; sorry I don't know much about interacting with grub via serial .. never needed that
[23:03] <irwiss> ha, was about to complain how 18.04 slowed down ssh/tmux but apparently the vm had flakey connection issues that got resolved today, hail lazyness! :P
[23:05] <compdoc> 18.04 seems fast at all things to me