=== JanC is now known as Guest74845 === JanC_ is now known as JanC [00:55] tarpman: the debian package is done by upstream, afaict [00:55] but not sure, i've not looked at the changes [00:56] https://sourceforge.net/p/postfixadmin/code/HEAD/tree/branches/postfixadmin-2.3/debian/control [00:56] Belldandu: tarpman: --^ [00:56] as explicit mariadb-* alternatives [00:56] i'd expect tht's what unstable will pick up when a new release happens [00:57] Belldandu: so you might just consider pinging the debian maintainers and asking if they'd package the new release [03:42] hi anyne around to help me with a apache problem? === timmy is now known as Guest98043 [03:44] Guest98043: people come and go.. best to just ask and hope :) [03:46] hi..i just installed apache and php 7 [03:46] everything seems to work fine when serviing html files [03:46] when i am serving a json file [03:46] i realised the content repeats [03:47] this is what my content looks like as rendered by the browser [03:47] HTTP/1.0 200 OK Cache-Control: no-cache Content-Type: application/json Date: Wed, 08 Jun 2016 03:40:50 GMT [03:47] so the content to be served is displayed..followed by HTTP/1.0 200 OK Cache-Control: no-cache Content-Type: application/json Date: Wed, 08 Jun 2016 03:40:50 GMT [03:47] then the content repeats again [03:50] o_O [05:12] EmilienM: the error that you were seeing is the same that blocked the neutron gate yesterday. tempest update broke neutron-fwaas tests. fixed by https://review.openstack.org/326150 === WARDEN is now known as Sysop_WARDEN [06:13] hi there. I have some ubuntu 16.04 servers with Kernel 4.4. Can I disable "mounting" of debugfs? On Ubuntu 14.04 with kernel 4.4 the debugs entry would go away after a reboot. But not with ubuntu 16.04. Any idea? [06:13] debugfs [06:13] is it in your /etc/fstab? [06:14] noope [06:14] only disk-by-uuid and swap [06:14] example: mount [06:14] mqueue /dev/mqueue mqueue rw,relatime 0 0 [06:14] debugfs /sys/kernel/debug debugfs rw,relatime 0 0 [06:15] hugetlbfs /dev/hugepages hugetlbfs rw,relatime 0 0 [06:15] I don't have such entries with ubuntu 14.04 [06:15] (this is not the full mount of course) [06:15] well now [06:15] I've got it mounted on two of my systems too. [06:15] harupmh. [06:16] yes, and if you use icinga/nagios and check disks, you have to adjust stuff so that it won't conmplain about debugfs. SInce I won't use it actively I wonder if I can make it go away. [06:17] There was an issue request for this at icinga2 so they "fixed" it, but I just do not understand if I can just disable it and where ... [06:17] ronator: aha. I think I've got it: systemctl cat sys-kernel-debug.mount [06:17] let me see ... [06:18] wow, how did you come up with this so fast ... [06:18] probably you could symlink /etc/systemd/system/sys-kernel-debug.mount to /dev/null to forbid it from happening. maybe systemctl mask can do that for you? maybe you'd need to stop the service first to get it to umount... [06:18] ronator: just a hunch to check systemd when I didn't see it in my fstab either. I got lucky. [06:18] What about commenting out all below [Mount] ? Or is this not a true file? [06:19] it is, but the version in /lib is owned by the package -- if you want ot make changes, they should go into a file in /etc/ instead [06:19] but yeah, awesome tip sarnold, thx I was searching the net and only fin ding boring descriptions of debugfs ;-) [06:20] and indeed it feels like a mistake for it to be loaded by default.. [06:21] fun fact is: I have ubuntu 14.04 with kernel 4.4. first bott, it shows debugfs. reboot and it is gone! with ubuntu 16.04 and kernel 4.4 it won't just disappear. [06:21] so thx for your hint, this way I can check the other systems that don't have it and compare them [06:23] heh, I'm surprised debugfs shows up in the first boot of 4.4 on a 14.04 -- in this case it's the systemd unit that causes it to be mounted [06:30] yes, I saw it happening. and monitoring plugins where failing because no permissions on that "device" ... [06:30] -h [06:31] it'd be worth poking upstreams on those plugins to try to get them to blacklist the in-kernel filesystems [06:31] I mean, it's handy, it reported this issue :) but it's also probably more annoying than useful [06:32] true. and yes, icinga2 e.g. did change the defaults for check_disks especially for ubuntu 16.04 ;-) I wonder if this hapens with Debian 8 ... [06:36] and I also wonder why the heck ubuntu 16.04 ships lxcd by default ... this also leads to stuff like "Additional Info: DISK CRITICAL - /run/lxcfs/controllers/freezer is not accessible: No permission" [06:37] ah, that's an easy one :) lxd is pretty neat and we wanted a low barrier to entry for folks to try it out :) [06:38] hrrhrr === sarnold is now known as sarnold_ === wisur_ is now known as wisur [08:14] hi, wnat to setup some nodes with conjure-up. How many servers and storage do i need ? is there any kind of reference and explained where the e.g. cephs is installed or where the horizon dashboard goes? [08:14] its for openstack [08:27] Is there a recommended PPA for installing PHP7 on 14.04? === gms is now known as Guest85880 === stikky is now known as _gms [08:55] coreycb: regarding debian/patches/skip-ryu-tests.patch for neutron, there are two uses of ryu, the second one is as the reference implementation for the bgp-dragent. I'm a bit sceptical about dropping tests for that part completely, because we are actively using it in our deployments [09:41] stokachu: see conjure-up question above please. [09:41] Mc_uli: stokachu won't be in for a few hours at least, assuming he's not away. === MacroMan_ is now known as MacroMan [11:02] w/in 1 === Bruh is now known as gitgud === gitgud is now known as Bruh [11:41] frickler, I've dropped the skip-ryu-tests.patch and no longer skip those tests in the newton b1 release [11:42] frickler, build-depends no longer need to be in main so we were able to drop that patch [11:42] coreycb: ah, great, that answers my follow-up question already ;) [11:43] frickler, :) === iberezovskiy_afk is now known as iberezovskiy [12:29] frickler: can you backport https://review.openstack.org/#/c/326150/ to Mitaka? [12:36] EmilienM: I can try, it probably would help if you can add a note to https://bugs.launchpad.net/neutron/+bug/1589521 stating how this affects you for Mitaka still [12:36] Launchpad bug 1589521 in tap-as-a-service "AttributeError: type object 'BaseNetworkTest' has no attribute '_try_delete_resource'" [Undecided,In progress] [12:40] rbasak: i've got this channel highlighted on conjure-up keywords, unfortunately i just got online [12:51] frickler: well, I have the same trace so I'm not sure what can I add [12:57] EmilienM: hmm, right, looks like it is a clean cherry-pick anyway, so lets wait what the neutrinos have to say on this (https://review.openstack.org/327068) [12:57] ok [12:57] in the meantime, I unblocked our CI in disabling fwaas on ubuntu [13:01] stokachu: np. The guy asking left anyway. [13:02] rbasak: is there like an info bot in here? so we could do like !conjure-up and it'll point them to askubuntu.com? [13:03] ! [13:03] !info [13:03] I thought there was once, but evidently not. [13:04] stokachu: ah, you watch askubuntu.com for conjure-up stuff? I can just point people there then. That's probably easier even if no bot. [13:04] rbasak: yea that would be awesome [13:05] Cool, I'll do that - thanks. [13:18] !yum [13:18] Uh, don't you mean !apt ? [13:19] its still there it seems ;) [13:34] howdy all [13:46] caribou: around? I'm looking at your updated kexec-tools MP. [13:47] caribou: I'm quite happy for you to just upload without a second review, but I'm going through as if you're not an uploader in order to try out the process. [13:48] I don't have my old git repo around, and I'd like to diff from before your rebase to after your rebase, but I can't find the old branch. Usefully LP remembers, so it's apparently de45e23. [13:50] caribou: would you mind tagging de45e23 as merge.v1 and pushing that please? I'm not sure this is actually documented anywhere but I intended that to be the process when asking for a re-review (or when you want to keep an older merge branch around before rebasing for any other reason) so the reviewer can get to both branches. [13:56] hello [13:56] This is for ubuntu servers aka vpses right? [13:56] can anyone think of any valid sane reason that when someone issues a 'ping' from 14.04 server that it would send the ping out over both the MAC address on that NIC port, but also a MAC address that doesn't even exist on the system (but isn't a broadcast MAC address)? [13:57] prince: this is a support channel for Ubuntu Server, not necessarily just VPSes [13:57] teward: obviously not just vpses, but for vps support as well I assume? [13:58] prince: again, it depends - from the Ubuntu OS on the VPS, we can assist, for the actual VPS itself, go to your provider [13:58] Oh no, ubuntu server related. [13:58] and in many cases what may be an "OS" issue is not, and is actually a provider issue [13:58] prince: you're free to ask your question, but you can get pointed to elsewhere if it's not actually an Ubuntu issue [13:58] that said, next time just ask your question [13:58] God! [13:58] rather than asking the premise of the channel [13:58] Do you need me to leave or something? [13:58] uh, what? [13:58] I said it's ubuntu related [13:58] prince: so ask your REAL QUESTION [14:00] jamespage, ddellav: I just uploaded cinder b1 for real this time. Looks like just heat and nova are stuck in proposed for the known MIRs we have. [14:00] coreycb, \o/ [14:01] hey guys, im looking for some kind of product to automate sync'ing the current dev server over to a backup server, starting the backup servers required, and to update my DNS record to have the backup as the primary server [14:01] and then signal me when the backup server has replicated globally type thing [14:01] aka its ok for me to take the production server offline [14:01] and do whatever [14:01] I figure something for this must exist [14:01] though its not terrible to do by hand [14:01] coreycb awesome. I'm working on the monasclient mir now [14:02] ddellav, ok cool [14:02] jamespage, suppose we can start backporting all the clients/oslos and core packages that are missing from the newton UCA? [14:03] coreycb, please do! [14:03] jamespage, ok [14:03] coreycb, branch PPA for newton is almost deployable [14:03] jamespage, awesome [14:03] coreycb, https://review.openstack.org/#/c/326597/ that needs landing and then we are all good [14:04] no ideas folks? [14:04] ;( [14:04] Xin, you might need that question to bake for a bit more that 3 mins :-) [14:04] lol nah its what I do currently [14:05] If theres a better way im all ears! [14:05] jamespage, why not use the package config for apache? [14:05] coreycb, well its a bit of a nuance of the charm [14:05] having two files named kesytone.conf confused the crap out of the ostemplategenerator [14:05] jamespage, ok [14:15] prince: i didn't say you couldn't ask about your issue, by the way - did you still want to ask your question? [14:26] teward: It's fine besides feeling offended, I'll probably ask later. [14:27] well, sorry, it wasn't *meant* that way === Matthew is now known as Guest11143 === Guest11143 is now known as jscar [14:35] HI, ubuntu newbie question.... how can I tell if apparmor utilties is installed? [14:37] hello anyone? [14:38] jscar: do you mean apparmor itself or some specific tooling? [14:38] jscar: for apparmor itself `apt-cache policy apparmor` [14:40] I'm trying to put mysqld into complain mode but aa-complain returns 'command not found' so I think apparmor utilties might not be installed but I want to know how to check [14:40] apparmor itself is running [14:41] <\9> find out what the package name is, using apt-cache search [14:41] apt-cache search aa-complain [14:41] <\9> and then use apt-cache policy, to find out whether it's installed or not [14:41] apt-cache policy [14:41] \9: :) [14:43] thank you - so much learning! :) [14:45] apt-cache policy apparmor-utils returns Installed: (none) [14:45] Guess it's not installed :) [14:46] yep ;) [14:46] many thanks for the pointers [14:57] ddellav, I think we both have package ci duties today so just to let you know I'm working on cinder [14:57] coreycb ok. I actually put myself down for tuesday and friday but i'll do one tandem with you a couple times [14:58] ddellav, oh ok I thought jamespage had tues/thurs. fix up the table then if you can. [14:58] coreycb when i looked only monday and thursday were selected by you guys heh [14:59] ddellav, oh package ci, not bugs [14:59] coreycb ahhhh, yes, ok. I haven't looked at CI yet [15:29] Term mode is eh [15:29] but maybe I am just butthurt because I always forget to use c-c c-k and c-c c-j [15:29] I also feel like those two keybinds could be a toggle [15:29] so just one keybind [15:30] I mainly run into trouble when I do fancy rendering things like resizing terminals [15:30] but every now and then a curses application will just bug out [15:30] runs x smoothly [15:31] oops [15:31] wrong channel === Bruh is now known as HouseMD [15:38] Odd_Bloke -- following up today on https://bugs.launchpad.net/cloud-images/+bug/1569237 Solution is to get the base image to have the default user be vagrant and not ubuntu. [15:38] Launchpad bug 1569237 in vagrant "vagrant xenial box is not provided with vagrant/vagrant username and password" [Undecided,New] [15:39] I was able to connect after a while -- timing was a factor; however, that's still broken from the vagrant workflow because the 16.04 vagrant box image doesn't follow vagrant conventions. I referenced the docs in my latest comment. [15:43] jackdpeterson: o/ [15:44] 14.04, on the other hand does. [15:47] UID/GID for vagrant:vagrant would be 1000/1000, respectively. [15:47] jackdpeterson: Right; Vagrant does allow boxes to configure what username they use (which is why `vagrant ssh` works), and using the `ubuntu` user is consistent with other Ubuntu images. [15:47] jackdpeterson: So is it just that people everywhere don't know this and assume the vagrant user will exist? [15:49] That's correct -- the assumption is that the vagrant user always exists and is the 'default'. Then people stack on provisioners to sync up code between their host machine's FS and internal VM configuration. this assumption is generally safe since that is the convention. [15:50] OK, fair enough. [15:53] and that uid/gid is based on my pulling up the 14.04 image and comparing. [15:54] jackdpeterson: I've updated the bug description, does that look reasonable to you? [15:55] Yep, that's very close to what my title was on the other cross-referenced bug was -- so I'm happy with the accuracy of that :-) [15:55] jackdpeterson: Cool; now did you mention that you were interested in fixing it? ^_^ [15:55] I added some confusion with the SSH-related nonsense (my apologies) [15:55] Yeah, If I can, I'm more than happy to! [15:57] jackdpeterson: So the relevant file is http://bazaar.launchpad.net/~ubuntu-core-dev/livecd-rootfs/trunk/view/head:/live-build/ubuntu-cpc/hooks/042-vagrant.binary [15:57] jackdpeterson: It's pretty crufty, unfortunately, but the way it works is basically to set up a cloud-init metadata source which is used at box boot to do configuration. [15:59] jackdpeterson: So I would suggest pulling down the latest yakkety Vagrant box (http://cloud-images.ubuntu.com/yakkety/current/yakkety-server-cloudimg-amd64-vagrant.box), unpacking that, modifying the contents and repacking it until you get something that works. [15:59] jackdpeterson: And then work out what will need to change in that hook. [16:00] Sure, then reply with a Github Gist? [16:01] jackdpeterson: Sure. :) === Guest47242 is now known as med_ === iberezovskiy is now known as iberezovskiy|off === degorenko is now known as _degorenko|afk === deadnull_ is now known as _deadnull [17:55] ddellav, cinder should be fixed up for newton CI, looking at neutron now === _deadnull is now known as deadnull_ [18:27] Odd_Bloke -- I have a gist -- but to be honest, I'm not sure if these changes will __work__ I'm not seeing documentation on how to spin up an environment for creating images from those hooks. https://gist.github.com/jackdpeterson/64c1f8a517087e07f7e40a1ace76ada0 Lines 70-72, and 115 are the changes [18:28] and when I manually execute these scripts it complains about missing files (and stepping through this and manually fixing the references to other paths) is tedious and feels wrong [18:29] oh, and the password doesn't need to be random ... it can just be 'vagrant' === IdleOne- is now known as IdleOne [18:51] ddellav, neutron should be fixed up for CI. I'll leave designate liberty and nova newton for you. [18:52] coreycb: ok, sounds good. === alexisb is now known as alexisb-afk === alexisb-afk is now known as alexisb === lifeless_ is now known as lifeless [21:39] I just wanted to come in here and reiterate that it is a moronic idea to install lxc and lxcfs by default in the server distribution and whoever is responsible for that decision should have their hard disks formatted. May an rm -rf /* come upon their root session. [21:47] when an upstart service runs and writes output, its stdout and stderr are written to /var/log/upstart/NAMEOFSERVICE.log where does stdout and stderr go in the systemd case? [21:55] jrwren: depends on the unit file but typically, systemctl status -l $service.service shows output from the daemon [21:56] rharper: thanks. I dug some more and found that what I expecte to be on stderr wasn't there because starting the process wasn't getting that far. [21:56] ah [23:54] What does everyone here use for a combination of DCIM and Asset/Inventory Management for their hardware devices in a server room?