[02:23] <Ham62> I was wondering why I had such an old version of ubuntu on this system before
[02:23] <Ham62> for some reason the setup CD for any version past 12 hangs trying to access the disk
[02:24] <Ham62> 14.04.4 just fails out saying it can't mount the media, 14.04.1 almost makes it into the partitioning screen but then just hangs at a blank purple screen
[02:27] <sarnold> you might want to try 14.04.6, 16.04.6, or 18.04.2, all of which are significantly newer than 14.04.4
[02:28] <Ham62> I don't know where my blank CD-ROMs are those are the only versions I had laying around
[02:28] <Ham62> so I'm gonna finish this 12.04 setup then do thr upgrade to 16.04 through the update program
[02:28] <Ham62> this will also be good practice for if I want to upgrade my main server system from 14 to 16 at some point
[02:29] <Ham62> but I heard that's the last version with 32bit support so I'm afraid that might break everything
[02:29] <sarnold> that's fine I think all those images are too large to burn to a CD anyway, most folks use memory sticks if they're not pxebooting or similar
[02:30] <sarnold> 18.04 LTS appears to still support x86 http://archive.ubuntu.com/ubuntu/dists/bionic-updates/main/ -- note the i386 directories
[02:31] <Ham62> huh... I wonder what that was that I read about dropping i386 then
[02:31] <Ham62> because someone at uni just brought it up to me a couple days ago when I was talking about maybe updating my system
[02:32] <sarnold> I believe we make people type something to confirm if they want to upgrade their x86 18.04 machines to 18.10
[02:32] <Ham62> ah
[02:33] <Ham62> I just need to be really careful with that system because I even noticed some programs on the package manager wouldn't run on the system
[02:33] <Ham62> they were using SSE2 or something which my CPU didn't support
[02:33] <sarnold> those discussions were long enough ago that I've forgotten the outcome, but I'm pretty sure 18.04 LTS is the last LTS release we're going to do for x86
[02:34] <Ham62> and that was just a text editor that was using SSE2 for some reason... apparently someone messed up the Go compiler config
[02:34] <sarnold> ahhhh, that makes some ense
[02:35] <Ham62> someone told me to try recompiling it from source but the Go compiler package used the same instructions lol
[02:35] <Ham62> and I didn't really feel like recompiling Go from source
[07:11] <lordievader> Good morning
=== lotuspsychje__ is now known as lotuspsychje
[08:09] <vtq> Hi! I need to override some rules in an apparmor profile. That is, I want to create a profile that goes: #include <blah> (which includes deny /foo) ; then override that with an allow /foo. Can I do that? I've been trying and I can't seem to get it to work
[08:52] <lotuspsychje> !apparmor | vtq can this help?
[08:52] <ubottu> vtq can this help?: For information about the AppArmor security framework employed in Ubuntu (since Gutsy Gibbon), see https://help.ubuntu.com/community/AppArmor
[08:54] <vtq> ubottu: seen it already, yes
[08:54] <ubottu> I have no seen command
[09:06] <rbasak> kstenerud: looks like there's an issue between php7.2 and horde :-/
[09:06] <rbasak> http://people.canonical.com/~ubuntu-archive/proposed-migration/update_excuses.html#php7.2
[11:15] <blackflow> vtq: afaik no, once denied, you can't allow.
[11:21] <kstenerud> rbasak I've updated the MP for php7.2 to include a depends for libicu
[11:21] <kstenerud> https://code.launchpad.net/~kstenerud/ubuntu/+source/php7.2/+git/php7.2/+merge/363924
[11:30] <rbasak> kstenerud: replied in the MP, thanks.
[11:56] <kstenerud> One question: If I'm going to create a new MP with a new ubuntu version (ubuntu2 instead of ubuntu1), does that mean that the changelog requires a new entry (keep the ubuntu1 changelog entry as is, and add a new entry for ubuntu2)? Or do we just pretend ubuntu1 never happened and put everything together?
[12:00] <ahasenack> add a new entry for ubuntu2
[12:02] <rbasak> kstenerud: what ahasenack says, but note that you should be able to infer that from my MP comment and your question. We can't pretend ubuntu1 never happened - see my MP comment. And I did ask for a "new changelog entry"!
[12:34] <lufi> Hi. I have two nics. 1 dhcp and 1 static. My problem is there are 2 default routes with the static network being priority. How can I set the dhcp the priority?
[12:35] <lufi> the server is a vm with static as hostonly and dhcp bridged (connecting to internet). here is my initial configuration of netplan/50-cloud-init.yaml https://pastebin.com/8QAZX2fQ
[12:36] <lufi> running `ip route` gives this defaults https://pastebin.com/pGP0YQLM
[12:37] <ahasenack> cpaelzer: thanks for the careful on the complicated samba related MPs
[12:37] <qman__> lufi: the DHCP one already has a metric, so set the static one's metric higher
[12:38] <ahasenack> cpaelzer: the samba versioning has a precedent from the other time we went ahead of debian
[12:38] <ahasenack> where I added ~ubuntu to the +dfsg suffix
[12:38] <lufi> qman__ how do I do that? using `routes` in netplan?
[12:38] <ahasenack> the only reason is the dfsg big
[12:39] <ahasenack> had it been the upstream tarball directly, I would have used just +dfsg-0ubuntuN
[12:39] <ahasenack> but maybe my concern is invalid?
[12:39] <ahasenack> i.e., our tarball differing from whatever debian will create eventually
[12:39] <qman__> I don't know anything about netplan, but surely there must be a metric option
[12:39] <lufi> Something like this? https://pastebin.com/8zzMucqu
[12:39] <qman__> if you use the normal /etc/network/interfaces file, and have ifmetric installed, it's just "metric NNN"
[12:40] <lufi> oh ok. I've been trying that routes thing
[12:40] <lufi> yeah 18 has a lot going on
[12:45] <lufi> qman__ solved it by doing this https://pastebin.com/Pvqw3sNv
[14:41] <mike802> hello all!
[14:41] <mike802> i want to work on getting a publicly visible url for my ubuntu server, but i don't know where to start
[14:42] <mike802> my php/apache site looks fine over a lan, and i've reserved the domain name on godaddy.....
[14:43] <mike802> should i perhaps start in #ubuntu?
[14:43] <sdeziel> mike802: go to godaddy's panel, add the server's public IP under the name you want
[14:43] <sdeziel> mike802: here is the right place IMHO
[14:43] <mike802> alright
[14:45] <rbasak> cpaelzer: are you sure the git-ubuntu change doesn't count as a feature change?
[14:47] <mike802> how do i look up my public ip?
[14:47] <mike802> on windows i just type that into google
[14:48] <sdeziel> mike802: I'd first check if you server has it directly configured: "ip addr"
[14:49] <mike802> nope...
[14:50] <sdeziel> mike802: OK so that probably means it's configured on a router/firewall in front of it that does NAT to translate it to its private IP
[14:51] <sdeziel> mike802: if that's the case, you will likely need to punch a hole in that firewall/router to do a port forward in order for people on the Internet to be able to reach your web server
[14:52] <Ussat> I would be very carefull and make sure php is secure
[14:52] <Ussat> php is known for exploits
[14:52] <mike802> alright, sounds like i might not be ready for a full roll-out
[14:52] <mike802> some good points to consider though
[14:53] <sdeziel> running a public service require a bit more attention to security indeed
[14:53] <Ussat> yup
[14:53] <mike802> i know, i could post my link to the wrong channel and have all kinds of issues
[14:53] <mike802> :/
[14:54] <sdeziel> mike802: I'd recommend not relying on the "hidden" nature of the link in question
[14:54] <mike802> alright.....
[14:55] <mike802> as far as my php goes, i've barely even started scrubbing input for whitespace and reserved words, etc
[14:55] <mike802> anyway, good stuff to consider, thanks
[14:57] <mike802> alright, so i've barely used godaddy before, what an i looking for
[14:57] <mike802> *am
[14:58] <mike802> my account -> account settings -> domain reg defaults -> dns control - enter nameservers?
[15:03] <sdeziel> mike802: at this point, I'd suggest you stick with godaddy's provided nameservers
[15:04] <sdeziel> mike802: you likely want to add an "A" record
[15:04] <mike802> ?
[15:04] <tomreyn> mike802: the server you are planning on bringing online, is this hosted with godaddy, too, or is this at your home?
[15:04] <mike802> it is at my home
[15:04] <tomreyn> (or somewhere else entirely)
[15:04] <cpaelzer> rbasak: you meant uvtool right?
[15:04] <mike802> as i mentioned, over a lan it looks fine
[15:04] <rbasak> cpaelzer: yes sorry
[15:05] <mike802> i am trying to get away from using a 192-based url
[15:05] <cpaelzer> rbasak: well, it is both - a fix to avoid guests with issues on cirrus to work better - and a feature to have much more modern things
[15:05] <mike802> trying to find a starting point (not married to godaddy)
[15:05] <tomreyn> mike802: i see. then you need to find out whether your ISP does dynamic or static ip address assignment, i.e. does your public ip address change over time?
[15:05] <cpaelzer> rbasak: I'm fine having it 19.10 as we have adressed the majority of issues in xorg as well
[15:06] <mike802> alright, i'll come back later
[15:24] <rbasak> cpaelzer: OK, thanks. There should be a PPA available.
[15:24] <rbasak> I also want to add support for network-config.
[15:35] <leftyfb> Are there any downsides to removing cloud-init from an ubuntu server install?
[15:35] <leftyfb> Seems like for bare metal, not being managed by MAAS or juju or anything, it just gets in the way
[15:36] <rbasak> How does it get in the way?
[15:37] <rbasak> If there's no datasource I don't think it's even active.
[15:39] <leftyfb> rbasak: it's multiple steps to set a hostname
[15:39] <rbasak> leftyfb: and how does removing cloud-init change anything?
[15:39] <rbasak> You set /etc/hostname, and /etc/hosts, and you're done.
[15:39] <leftyfb> rbasak: good luck with that
[15:40] <rbasak> OK, well if you don't want to explain to Ubuntu developers what the problem actually is, what are you doing on this channel?
[15:40] <teward> rbasak: he *may* be running into the defaults for how cloud-init is
[15:40] <teward> let me pull up that subiquity bug..
[15:41] <leftyfb> rbasak: That in insufficient. You need to use hosnamectl set-hostname and you need to disable preserve_hostname: true in cloud.cfg. Otherwise cloud-init keeps the original hostname. Removing cloud-init allows you to just edit /etc/hostname
[15:41] <leftyfb> rbasak: sorry, I was still typing
[15:41] <teward> https://bugs.launchpad.net/ubuntu/+source/cloud-init/+bug/1780867
[15:41] <ubottu> Launchpad bug 1780867 in subiquity "hostname unchangeable / some daemon changes and resets /etc/hostname" [High,Triaged]
[15:41] <teward> leftyfb: rbasak: ^ read up, known 'issue'
[15:42] <leftyfb> hm, I'm not so sure it's invalid for cloud-init. Let me check
[15:42] <teward> re-reported by me in December.  cloud-init needs preserve_hostname in /etc/cloud/cloud.cfg to be set to "true"
[15:42] <teward> leftyfb: ultimately not a bug in cloud-init but a bug in the config file that Subiquity leaves as-is
[15:42] <teward> Subiquity could alter that file :P
[15:42] <sdeziel> I can relate to that. cloud-init does revert hostname changes on reboots by default
[15:42] <teward> ^ this
[15:43] <rbasak> Thanks
[15:43] <leftyfb> So shouldn't the bug be that cloud-init should set the default to not revert? As opposed to subiquity changing it at install?
[15:43] <rbasak> That's definitely a valid bug.
[15:43] <rbasak> But I'm confused as I didn't think cloud-init was active after a subuiqity install.
[15:44] <rbasak> If it is, and it's meant to be, then removing it probably isn't a good idea.
[15:44] <rbasak> Fixing the bug of course is :-/
[15:45] <leftyfb> https://photos.app.goo.gl/iQCS5LwWY4TqF1Zb7
[15:45] <leftyfb> cloud-init is definitely active on a fresh install
[15:45] <rbasak> I guess subiquity injects a data source then
[15:45] <rbasak> TIL
[15:46] <metabsd> Hi :)
[15:46] <metabsd> Can I use clearpart with curtin or preseed in recent ubuntu version ?
[15:48] <teward> rbasak: cloud-init *is* active, I can confirm that.  BUT it needs its defaults altered, IIRC that can be done as a subiquity "cleanup" task before the installer finishes, but that's a discussion for the Subiquity devs
[15:50] <teward> i'm not super familiar with the subiquity internals ;)
[15:53] <leftyfb> ok, so bottom line, removing cloud-init from a subiquity install has no adverse affect, correct?
[16:01] <teward> I can't verify that unfortunately leftyfb, YMMV.
[16:46] <coreycb> sahid: neutron-vpnaas 12.0.1 is now uploaded to the unapproved queue for bionic. i'll add the sru team to the bug now.
[16:47] <sahid> coreycb: ack thanks
[18:02] <Blueking> are there ways to find out what is where about finding what folders is where on multiple hdd's ?
[18:09] <metabsd> There a way to do templating storage layout and network config in MaaS ?
[18:49] <The_Actor> Hey guys
[18:50] <lordcirth__> Hello
[18:53] <The_Actor> I am trying to set up an LXE container to upload to my virtulization server for practicing web development. Is there a base LXE image for download, or do I have to create it?
[18:54] <ahasenack> you mean LXD?
[18:54] <ahasenack> if you mean lxd, you can easily launch almost any release of ubuntu, debian, centos, fedora, etc, like this:
[18:54] <The_Actor> no LXE as in a virtual container that can run on a virtulized environment as opposed to a KVM image
[18:55] <ahasenack> lxc launch ubuntu-daily:bionic
[18:55] <ahasenack> I don't know what lxe is
[18:55] <teward> The_Actor: there is no such thing as an "LXE" container
[18:56] <teward> unless you mean LXE *hypervisors*
[18:56] <teward> (which is similar to OpenVZ)
[18:56] <teward> but that's close to LXC/LXD in how it operates
[18:57] <teward> which is far easier to get 'working' than OpenVZ or similar systems
[18:57] <The_Actor> There are two popular types of virtulization, the first is KVM and the newest cool way of doing it is via LXE. LXE is not a full virtulization and adds the bennefit of shared resources such as free RAM.
[18:58] <ahasenack> it sounds like you are describing lxd
[18:58] <teward> The_Actor: you've just described LXD
[18:58] <The_Actor> I am looking for a base image to upload to my virtulization server and make it into a practice webserver
[18:58] <The_Actor> one second
[18:58] <ahasenack> the launch command I pasted above will download the image (if you haven't it already) and launch the container
[18:58] <teward> LXE is "not new" technology, neither is OpenVZ.  LX*D* is the 'newest' cool way of doing containerization and resource sharing on many host OSes, Ubuntu included.
[18:59] <teward> I think LXD is what you are actually looking at :P
[19:00] <The_Actor> I know it as LXC with the LXE file type extention
[19:00] <The_Actor> "LXC (Linux Containers) is an operating-system-level virtualization method for running multiple isolated Linux systems (containers) on a control host using a single Linux kernel."
[19:01] <The_Actor> Am I confused?
[19:01] <lordcirth__> I have never heard of LXE, and I use LXC containers lots
[19:02] <lordcirth__> What does a .LXE file do?
[19:02] <The_Actor> Ok well, I am running an EdgeLinux Server, I want to upload an LXC image to it. Do you know if there is a base image that is supported and adheres to the standards I can use to start?
[19:03] <ahasenack> do you have ubuntu installed on that server?
[19:07] <sdeziel> edgelinux seems to be a gentoo derivative that uses libvirt to drive KVM and LXC
[19:09] <lordcirth__> He's gone, anyway
[19:14] <ahasenack> I wasn't sure if he meant a server type, or an OS
=== arooni_team_b is now known as arooni
[20:22] <mason> I'm confused by something. I see a number of files in /usr/share/doc/libzfs2linux that aren't registered as belonging to any package, when I'd have expected them to belong to libzfs2linux.
[20:22] <mason> Maybe they're holdovers from when this box was 16.04.
[20:26] <lordcirth_> mason, possibly. You can use apt-file search to check
[20:27] <mason> lordcirth_: as with dpkg -S, it doesn't find anything - that said, I'll look on a Xenial box and see
[20:28] <mason> Not there either. Odd. /usr/share/doc/libzfs2linux/README.Debian is one of them.
[20:29] <sdeziel> mason: on Bionic, those files are all symlinks to files owned by libnvpair1linux, exception made of /usr/share/doc/libzfs2linux/copyright.
[20:29] <mason> sdeziel: Ah, hah. Good catch. I hadn't considered that they might not be regular files.
[20:30] <mason> And I guess I never realized that dpkg -S doesn't track symlinks.
[20:32] <sdeziel> it seems like it does though: https://paste.ubuntu.com/p/S8cMtS8pNV/
[20:33] <mason> Hrm. Wonder where my README.Debian came from then. I see it in the source package, but I don't see where it's being told to be installed.
[20:34] <mason> And you don't seem to have it in your paste.
[20:35] <mason> I don't see it in a fresh install of the library on a Xenial box, so that's not it.
[20:36] <mason> This is down a rabbit hole and probably not important, but I'm pondering using libzfs for an expansion of some stuff that's current sh.
[20:36] <mason> currently*
[20:37] <sdeziel> mason: I know there is/was a PPA for ZoL so maybe it left some files hanging around?
[20:37] <mason> So I looked for a man page, didn't see one, and then looked for plain docs, and noticed bits not registered as being part of a package.
[20:37] <mason> Maybe. I don't think I ever used it, but they came from somewhere.
[20:38] <sdeziel> on Xenial, I see no symlinks
[20:39] <sdeziel> FYI: https://paste.ubuntu.com/p/xb2Xpqfqx4/
[20:39] <mason> hrm