[00:01] <xnox> treehug88, you say rolling, but compiling on trusty?!
[00:01] <xnox> using acient ruby?
[00:01] <xnox> and then launch docker on trusty, to building something in bionic
[00:02] <xnox> treehug88, why not ask travis to provide bionic in dist: straight away, instead of trusty?
[00:02] <sarnold> is that possible?
[00:02] <sarnold> I see much grumbling about trusty on travis ..
[00:04] <xnox> ah based on https://github.com/travis-ci/travis-ci/issues/7260
[00:04] <treehug88> xnox I've been adding features to and maintaining the bash scripts of git-secret itself. The test/build system isn't something I've worked on much. We'd definitely welcome PRs
[00:04] <xnox> they did make `dist: xenial` work and then took it away
[00:07] <xnox> treehug88, this is extremely painful to read, as the logs are not at all structured it seems
[00:07] <xnox> but i think gnupg documentation fails to build on bionic atm
[00:07] <xnox> Makefile:1185: recipe for target 'gnupg-module-overview.pdf' failed
[00:10] <treehug88> xnox nice catch. Can we tell what's missing?
[00:10] <xnox> treehug88, i got cross-eyed
[00:10] <xnox> treehug88, if you can generate buildlogs with newlines that would be helpful
[00:10] <xnox> or like try configuring/building gnupg without docs
=== not_phunyguy is now known as phunyguy
[00:20] <treehug88> good idea, thanks xnox
[04:41] <patz0r> hey all, is there a way to install ubuntu server (18.04) without the cloud init stuff, or is there a better way to install if it I just want a very minimal installation without any extra stuff?
[04:42] <patz0r> maybe i'm not understanding it correctly but it doesn't seem useful to me, like resetting hostname every reboot etc.
[04:51] <cpaelzer> good morning
[04:51] <cpaelzer> patz0r: if you do not provide a config to cloud-init it does almost nothing on first boot, and nothing at all later on
[04:52] <cpaelzer> also it isn't big in terms of disk space, so there usually is really no need to go an extra mile to install without it
[04:52] <patz0r> cpaelzer, thanks for that, i just found that when i change my server hostname, it changes back every reboot thanks to cloud-init scripts
[04:53] <patz0r> I would rather not have these kind of scripts making changes, if you know what I mean
[04:54] <cpaelzer> patz0r: I understand, but the question is who is providing the config so that it does so - do you run this in a cloud which provides metatdata?
[04:56] <cpaelzer> I think we can still disable it by telling it in a local config file to ignore everything else, I need to check where exactly
[04:56] <cpaelzer> but first lets make clear we understand where your config is from - so what service are you running on if I might ask?
[04:58] <patz0r> i think it can be disabled with sudo touch /etc/cloud/cloud-init.disabled
[04:59] <cpaelzer> patz0r: yes https://cloudinit.readthedocs.io/en/latest/topics/boot.html
[05:00] <patz0r> I guess I am finding a lot of changes in 18.04 which I'm not used to, I prefer a more traditional installation
[05:01] <patz0r> i'm using other tools like ansible to manage my configs
[05:01] <patz0r> so when i change something with ansible, and then cloud-init changes it back, that's annoying
[05:02] <patz0r> maybe i just need to change how I am doing things
[05:06] <cpaelzer> patz0r: if you are in a cloud that provides metadata cloud-init will do what it is told (e.g. set the hostname), cloud-init and ansible are not doing exactly the same things so I think they are fidn to live alongside for different purposes. yet obviously if there are conflicting configs both have to be modified to not fight about the setting
[05:31] <patz0r> cpaelzer, thanks for that. I am not using ubuntu in any kind of cloud, and I don't have metadata servers or anything like that, which is why I don't want cloud-init
[05:31] <patz0r> ubuntu has specific openstack images/installers, it makes sense it it's included in those, but not the default bare metal server installer
[05:34] <cpaelzer> patz0r: the new installer uses the same mechanisms locally as in clouds so they can benefit of the same bugfixes and feature development, but it is mostly curtin they share https://curtin.readthedocs.io/en/latest/topics/overview.html and much less cloud-init features
[05:36] <cpaelzer> personally I also like that I'm able to throw such images at bare metal as well and using cloud-init/curtin (through https://maas.io/ most of the time) will make it a usable "instance" without ahving to touch the installer ever
[05:38] <cpaelzer> with LOC being known as lines of code, I doubt I can use LOL as "lines of log" :-/
[05:39] <patz0r> thanks cpaelzer - i'm just a noob trying to manage my small fleet of servers and begin automating, i will just have to keep working through it and peel back the layers which are causing my problems as I discover them
[05:39] <patz0r> i found a simple fix to disable cloud init from changing my hostnames on boot, but it made me wonder if I needed it at all or if it would be simpler not to have it
[05:39] <cpaelzer> patz0r: I thought about the hostname resetting, and I agree that if you installed through the new installer it should be a one shot setting and no reoccuring one
[05:40] <cpaelzer> patz0r: would you mind filing a bug at https://bugs.launchpad.net/cloud-init or https://bugs.launchpad.net/subiquity
[05:40] <cpaelzer> ping me when you have opened it, I'll make sure it has both bug tasks so the devs of it can consider making it a one shot action (or not, but with reasons)
[05:41] <patz0r> Yea it seems the new installer created a config that is being applied at boot, rather than a static config I can modify
[05:41] <patz0r> I had to go into /etc/cloud/cloud.cfg and change preserve_hostname: from false to true
[05:41] <patz0r> I will open a bug there shortly if you think it is indded a bug
[05:42] <patz0r> thank you for your help and advice
[05:44] <cpaelzer> it is a community thing patz0r, so many people benefit from either having it being faster (than old install) or using its new features - it is only normal that there are rough edges that should be discussed
[05:46] <patz0r> I can definitely see the advantages, but for me personally I think it would be nice to have a 'normal' installer and a 'cloud' installer for cloud images
[05:53] <patz0r> I think I found my problem, maybe i should just be using the debian installer rather than the subiquity installer
[05:53] <patz0r> https://blog.printk.io/2018/04/ubuntu-18-04-lts-bionic-beaver-server-installer-differences/
[05:53] <cpaelzer> patz0r: if you "just want the old style" yes
[05:54] <patz0r> I actually normally prefer to use the minimal ISO installer, so I think that will solve my problem for now :)
[05:55] <patz0r> i'm in the process of upgrading a lot of older machines from 14.04 and 16.04 to 18.04 and trying to maintain consistency, but it's hard with things like netplan and systemd changing
[05:55] <patz0r> so I'm trying to keep it as simple as possible
[05:56] <patz0r> well, I consider myself a linux noob and so I am finding it to be challenging :)
[05:56] <cpaelzer> you are clearly not a noob, and challenges let us grow :-) I hope the further upgrading goes well
[05:59] <patz0r> thanks cpaelzer nice chatting to you :)
[05:59] <cpaelzer> yw
[10:30] <zzarr> hello! I am trying to run a lxc, but I get "lxc-start: command not found"
[10:32] <TJ-> zzarr: is package lxc-utils installed?
[10:33] <zzarr> I'll have a look
[10:33] <zzarr> no, and I can't find it with apt search lxc-utils ether
[10:34] <zzarr> it's a brand new installation of Ubuntu server 18.04
[10:34] <TJ-> !info lxc-utils bionic
[10:34] <ubottu> lxc-utils (source: lxc): Linux Containers userspace tools. In component universe, is optional. Version 3.0.1-0ubuntu1~18.04.2 (bionic), package size 374 kB, installed size 1260 kB
[10:34] <TJ-> zzarr: You'll need to enable the "universe" component
[10:35] <zzarr> universe, thanks
[10:35] <TJ-> zzarr: "sudo add-apt-repository universe"
[10:36] <zzarr> thanks :)
[10:36] <zzarr> nice
[10:39] <zzarr> I got "Failed to load rcfile" when running lxc-start
[10:40] <zzarr> where should the rcfile be located and what should be in it?
[11:26] <a_ok> So I have set up NUT the way I want it with upssched. However the shutdown command that I have runs as the nut user. Unfortunatly that user has no permissions to terminate the system. What is a good way to do this?
[11:27] <blackflow> polkit trickery or passwordless sudo for a script that only contains the shutdown call
[11:28] <blackflow> maybe 'pkexec poweroff' even suffices
[11:32] <a_ok> blackflow: I'm afraid that polkit needs a host of things like a logind session. Sudo seems to be the easiest route
[11:32] <a_ok> thanks!
[11:32] <blackflow> yw
[12:12] <ahasenack> good morning
[12:19] <ahasenack> cpaelzer: is it ok to have a Bug-Fedora in a dep3 header?
[12:19] <ahasenack> looks fine to me, but I haven't seen that in a debian or ubuntu package yet
[12:24] <cpaelzer> hi ahasenack
[12:24] <cpaelzer> I think it is fine
[12:24] <cpaelzer> it might be out of spec, but personally I think useful+reasonable are the only tag guidelines there
[12:25] <cpaelzer> unless you want to have them processed by scripts, but that isn't important for bug-fedora
[12:25] <ahasenack> the dep3 spec says Bug-<Vendor>
[12:25] <ahasenack> so yeah
[12:26] <ahasenack> cpaelzer: another question, I don't remember our discussion from the sprint. Should I refresh a patch that applies as-is from upstream, but a) with offset; b) with "noise" (--show-c-func style)?
[12:26] <ahasenack> rbasak IIRC likes to be able to compare the patch with upstream in these cases and expected a 100% match. But this is a small patch, easily verified by visual inspection
[12:27] <cpaelzer> ahasenack: if it applied with dquilt push --fuzz=0 then keep it as is
[12:27] <cpaelzer> if forced to modify for this (forbidden fuzz) or anything else, then full refresh with "the good" quilt config
[12:27] <cpaelzer> that is what I remember
[12:28] <ahasenack> it applies
[12:28] <ahasenack> ok
[12:28] <cpaelzer> TL;DR: keep them as-is if working
[12:28] <rbasak> I think the conclusion was that usually we'd want quilt patches to be refreshed with the standard parameters, with the exception that if it applies directly from upstream, then leave as-is
[12:28] <rbasak> Yeah
[12:29] <cpaelzer> and we said that in most cases this would match tags "upstream, link" would be as-is and "backport, link" would be with refresh
[12:31] <ahasenack> cpaelzer: is there a place to record why forwarding wasn't needed?
[12:31] <ahasenack> maybe the long description?
[12:32] <ahasenack> I would love "forwarded" to allow for a short explanation after the "no, ", or "not-needed, "
[12:32] <ahasenack> that was probably Karl's intent with the second Description field
[12:33] <cpaelzer> in the long desc
[12:33] <cpaelzer> I sometimes added Forwarding-Info: <txt>
[12:33] <cpaelzer> but that is out of spec as much as everything else
[12:33] <cpaelzer> and since it could one day collide with something the description is probably better
[12:33] <cpaelzer> I put notes there since a while
[12:34] <ahasenack> ok
[12:36] <cpaelzer> I put notes there since a while
[12:36] <cpaelzer> sorry, key+enter in wrong chan
[15:04] <DenBeiren> i'm having troubles mounting a share from one server into another :s
[15:05] <DenBeiren> i can't get the rights as they should be
[15:07] <DenBeiren> on the serverside (nfs ubuntu) the rights are 777, on the client side (nfs ubuntu) 750
[15:08] <DenBeiren> i've found so many tutorials i can't see the trees trough the forest
[15:13] <ahasenack> does anybody know how a2query (from apache2) know the difference between a module that's just not enabled, versus one that was disabled via a2dismod?
[15:14] <ahasenack> specifically these two outputs: https://pastebin.ubuntu.com/p/mbZnqfrxVY/
[15:14] <ahasenack> it's either "no module matches <name>"
[15:14] <ahasenack> or "no module matches <name> (disabled by site administrator)"
[15:15] <ahasenack> in both cases, the symlink in /etc/apache2/mods-enabled/ does not exist
[15:15] <ahasenack> so I don't know where else it could be recorded
[15:22] <TJ-> ahasenack: maybe /var/lib/apache2/module/ ?
[15:39] <kierank> rbasak: after yesterday's preseed change we have this error when trying to install
[15:39] <kierank> https://askubuntu.com/questions/1076247/after-clean-install-of-ubuntu-14-04-i-get-shim-signed-error
[15:39] <kierank> The fix suggested works but it's a bit strange
[16:02] <rbasak> kierank: can you try again in a few hours please, and/or after a remirror?
[16:02] <rbasak> There was a broken time window earlier today. It's been fixed I believe (and tooling changed to try to prevent a regression).
[16:02] <kierank> rbasak: sure
[16:23] <Zahovay> Hello guys, can any of you help me? we just made a new website, I would like to check if my server is protected against known attacks
[16:29] <teward> Zahovay: if you have no objections to me pointing OpenVAS and Nessus at your site I'll be happy to generate a report for you
[16:30] <teward> but note that that has some risks - you're trusting someone else to identify whether your site is 'safe'
[16:30] <teward> and you're trusting they won't destroy your site :P
[16:31] <Zahovay> teward: well its better if we know this now, instead of after launch
[16:42] <DammitJim> what does this do: Unattended-Upgrade::Allowed-Origins {
[16:42] <DammitJim>         "${distro_id}:${distro_codename}";
[16:42] <DammitJim> on /etc/apt/apt.conf.d/50unattended-upgrades ?
[16:43] <DammitJim> it seems I have security, updates, proposed and backports commented out, but not that first line
[16:43] <DammitJim> I have been instructed to disable all automatic updates, but first wanted to understand what that line meant
[17:15] <ahasenack> TJ-: yep, it's in there, thanks!
[17:26] <plm> TJ-: hey =D
[17:30] <rbasak> DammitJim: it's the release pocket. It's active during development of a release (ie. before the release date) and inactive after.
[17:38] <TJ-> plm: hiya. script isn't ready yet; got other things ahead of it
=== tomreyn_ is now known as tomreyn
[19:01] <plm> TJ-: all right, not problem. The important is that you have success running ubuntu armhf on LXD using qemu-sttic, right? Now just missing the script =D
[19:02] <plm> TJ-: do you think that until next sunday the script will be ready?
[19:02] <TJ-> plm: I'm going to spend an hour on it tomorrow my time (in the UK)
[20:02] <ahasenack> cpaelzer: I won't be around tomorrow to ping you within your timezone, sorry. Were you planning to also handle https://bugs.launchpad.net/ubuntu/+source/samba/+bug/1792400 for trusty? It's the same version of samba that is in xenial, and you updated that one
[20:02] <ubottu> Launchpad bug 1792400 in samba (Ubuntu Trusty) "smbd failed in host when both lxd container and host have smbd" [Undecided,Triaged]
[20:04] <DammitJim> do you guys know why when I try to mkfs.xfs for a new logical volume that is in a software RAID 1 configuration, my system crashes?
[20:04] <DammitJim> any known issueson Ubuntu 16.04 that would cause this?
[20:05] <sarnold> do simpler operations like dd over the logical volume show the same issue?
[20:06] <sarnold> can you use mkfs.xfs on simpler block devices? individual block devices? files?
[20:06] <DammitJim> I could try that
[20:06] <_KaszpiR_> what do you mean by 'crashes'?
[20:07] <DammitJim> the putty session freezes after the mkfs.xfs /dev/arraylv/activelv
[20:07] <DammitJim> and when I go to the console (this is a VM), there is a crash dump of addresses and stuff
[20:11] <DammitJim> maybe a different question... is there such a thing as setting up LVM on top of a software RAID 1?
[20:11] <blackflow> of course. with mdadm usually. LVM can do raid too but I never tried that.
[20:12] <blackflow> it'd be interesting to see that dump, top of it more than the bottom. pic even if you can't pastebin.
[20:13] <DammitJim> let me try
[20:13] <DammitJim> I'm having to see if I can get support from Canonical and pay for this problem since I need it resolved right away
[20:13] <DammitJim> *sigh*
[20:14] <blackflow> that'd be swell.
[20:15] <sarnold> I think I'd expect lvm to be able to stack on top of mdadm..
[20:15] <DammitJim> do you think I can get support right away or is this something where one needs to have a support plan in place?
[20:15] <DammitJim> sarnold, apparently that's normal (in CentOS)
[20:15] <sarnold> but you might be better served by zfs instead https://pthree.org/2012/12/04/zfs-administration-part-i-vdevs/
[20:15] <DammitJim> but for some reason things aren't working for the company that is doing this install on Ubuntu
[20:16] <mason> Eh? mdadm is fine underneath lvm - what's breaking?
[20:16] <mason> DammitJim:
[20:16] <sarnold> hey mason :)
[20:16] <mason> sarnold: o/
[20:17] <DammitJim> when I try to mkfs.xfs /dev/arraylv/activelv
[20:17] <DammitJim> the system crashes
[20:17] <mason> DammitJim: That's not supposed to happen.
[20:17] <DammitJim> correcto
[20:18] <mason> DammitJim: What's the underlying storage?
[20:18] <DammitJim> 2 fusionIO cards
[20:18] <DammitJim> ESXi on the host
[20:18] <DammitJim> the server I'm configuring is a VM
[20:18] <_KaszpiR_> why on earth you want to make a raid1 on vm?
[20:18] <blackflow>  +1 for the ZFS recommendation.
[20:18] <DammitJim> because of redundancy of the fusionIO cards
[20:19] <blackflow> _KaszpiR_: why not, the storage of a VM can be actual physical hardware
[20:19] <mason> DammitJim: So, does it also bomb out with something other than XFS? I'd vary that since you can.
[20:19] <_KaszpiR_> just use proper SAN?
[20:19] <blackflow> it's usually like that in modern clouds where compute nodes are separate from storage, iscsi galore
[20:20] <DammitJim> ext4
[20:20] <mason> DammitJim: It bombs out with ext4?
[20:20] <DammitJim> yes ma
[20:20] <DammitJim> mason
[20:21] <DammitJim> gosh, I'm torn... I don't know if I need to get support or spend the time troubleshooting this myself
[20:21] <mason> DammitJim: Can you get a vmcore? Do you see a stack trace?
[20:21] <DammitJim> I'm sure Ubuntu Advantage is not cheap
[20:21] <_KaszpiR_> what kind of service you want to host on that vm?
[20:21] <DammitJim> IBM DB2
[20:21] <DammitJim> mason, vmcore?
[20:22] <DammitJim> there is a stack trace on the console... how can I grab it? is it all saved to a file?
[20:22] <mason> DammitJim: If you have a chance to vary the FusionIO cards that'd be useful too. Also, have you tried dropping mdadm and just having lvm as the base on *one*?
[20:22] <DammitJim> no, we just came across this
[20:22] <mason> DammitJim: The other thing that comes to mind is, try what you're doing as a once-off with only one underlying device in the mdadm.
[20:23] <mason> Basically, twiddle all the available knobs to map where it fails and where it doesn't.
[20:23] <DammitJim> yeah
[20:23] <mason> DammitJim: http://manpages.ubuntu.com/manpages/xenial/man8/kdump-config.8.html
[20:23] <DammitJim> I wish there was something else to look at that would tell me where to go
[20:23] <mason> DammitJim: If you can get a vmcore that'll be hugely useful.
[20:23] <blackflow> DammitJim: well that dump will tell you _where_ the crash occurs
[20:24] <mason> Not a bad start really. :P
[20:25] <sarnold> DammitJim: I think the price seems fair, especially compared to the cost of your time https://www.ubuntu.com/support/plans-and-pricing
[20:36] <_KaszpiR_> hm this reminds me some bug with multipath under some system.....
[20:41] <DammitJim> what?
[20:43] <_KaszpiR_> someone had vm with multiple disks like you with multipath enabled and it was causing some weird shit&dumps
[20:43] <_KaszpiR_> remembering it through a haze, so ignore it.
=== elsheepo_ is now known as elsheepo