[00:22] <zul> hallyn:  disable the test :)
[02:23] <lexbanks1985> hey guys... lookin for some help here
[03:27] <Macer> in case someone was wondering about ssds... heh...
[03:27] <Macer> http://techreport.com/review/26523/the-ssd-endurance-experiment-casualties-on-the-way-to-a-petabyte
[03:28] <Macer> makes me wonder if SSDs can easily outlast platters in servers nowadays
[09:25] <`ajven> hi im looking for command pear what package i need to install to use it ?
[09:25] <`ajven> its php-pear maybe ?
[09:26] <cfhowlett> !info pear
[10:04] <Ajven_> !info zsh
[10:52] <lkthomas> does anyone know how does preseed work with partman ?
[11:19] <pmatulis> morning
[13:24] <hallyn_> zul: ok so i'm skipping the *(&$%* test and pkg built at https://launchpad.net/~serge-hallyn/+archive/libvirt-host-upstart-fix . running qa regression tests now, then i'll push to utopic
[13:25] <zul> hallyn_:  coolio :)
[13:41] <hallyn_> zul: so.. uh.  i should call it 1.2.5-0ubuntu1  right? :)  or dod something happen with the archive that i don't know about?
[13:41] <zul> yep
[13:41] <zul> thats right
[13:41] <hallyn_> (will take a bit longer to get the qrt running, not ready to push yet)
[13:42] <hallyn_> ok thanks
[13:42] <zul> ack
[14:21] <hallyn_> zul: feh    2014-06-17 14:20:28.044+0000: 14370: error : AppArmorGenSecurityLabel:457 : internal error: cannot load AppArmor profile 'libvirt-19f903d3-56b
[14:21] <hallyn_> 8-493e-b255-8dce9c953da6'
[14:23] <zul> 1.2.5?
[14:38] <hallyn_> zul: yeah.  something there's broken
[14:38] <hallyn_> heh, tons of old patches in there, had to do a for f in debian/patches/*; do f=`basename $f`; if [ $f != "series" ]; then grep -q $f debian/patches/series || rm debian/patches/$f; fi; done
[14:38] <hallyn_> but i assume this is broken upstream
[15:19] <jdstrand> zul: do you have any idea how to boot from a ceph volume on essex? I've tried what feels like a hundred things
[15:19] <jdstrand> (dd, rsync/grub-install, qemu-img convert, variations on all of it...)
[15:20] <jdstrand> zul: nm
[15:32] <hallyn_> zul: jdstrand: feh, commit 43c030f046932b728fcfd2c3dd5fe5f01b707df4 LXC driver: generate apparmor profiles for guests looks like it broke kvm apparmor
[15:33] <zul> hallyn_:  qualtity testing that upstream has
[15:35] <jdstrand> hallyn_: that has:
[15:35] <jdstrand>  profile LIBVIRT_TEMPLATE {
[15:35] <jdstrand> -  #include <abstractions/libvirt-qemu>
[15:35] <jdstrand> +  #include <abstractions/libvirt-driver>
[15:35] <jdstrand>  }
[15:35] <jdstrand> could it be as simple as updating what we ship?
[15:36] <hallyn_> it could be, still looking.  why on earth call it 'libvirt-driver' though?  when the lxc one is libvirt-lxc?
[15:36] <hallyn_> that looks liek somethign was broken in the author's setup and he did a quick fix
[15:36] <jdstrand> the 'libvirt-driver' is the abstraction that all profiles will include
[15:37] <hallyn_> jdstrand: thanks for the pointer, that's simpler than i'd hoped :)  testing
[15:37] <jdstrand> tbh, it should probably be separate for VMs and containers
[15:39] <hallyn_> i owe you one :)
[15:49] <doge_wow> I would like to know: is it possible (and wise) to upgrade Ubuntu Server to a major version with a remote SSH connection? For many operating systems it is recommended to do a clean install.
[15:59] <genii> doge_wow: In worst-case that it fails, do you have some way back into the machine remotely, like a KVM over IP switch, etc?
[16:01] <doge_wow> genii: yes, I can drive to the server center, about an hour's drive from here.
[16:02] <genii> doge_wow: In that case, it is probably fairly safe to attempt, with do-release-upgrade
[16:02] <genii> ordering lunch, afk
[16:07] <hallyn_> zul: jdstrand: with that qrt passes, so pushing to utopic, we can proceed from there
[16:09] <zul> hallyn_:  yay!
[16:11] <jdstrand> hallyn_: ah, good. though, I'm thinking that the libvirt-driver abstraction is probably the wrong way to go about it. qemu VMs will certainly need different accesses than containers. Seems to me would be the way to go is continue to have libvirt-qemu and then have a separate libvirt-lxc. the adjust virt-aa-helper.c to use one or the other based on the driver. then we ship both abstractions
[16:12] <jdstrand> hallyn_: I don't think that needs to be fixed in this upload per se, but worth thinking about
[16:13] <jdstrand> hallyn_: if you look in /etc/apparmor.d/abstractions/libvirt-qemu, you'll see what I mean (why give the container profile ix on /usr/bin/qemu-system-*, for example)
[16:13] <hallyn_> jdstrand: yeah, I agree, that's why the commonization seemed strange to me.  I imagine that in the next few months we'll be seriously wanting apparmor in libvirt containers and that'll be a good time to address that
[16:14] <hallyn_> (right now there are bigger fish to fry :)
[16:14] <hallyn_> s/bigger/other/
[16:19] <jcastro> jamespage, where can I wishlist a feature for the openstack charms?
[16:19] <jcastro> http://openstack.prov12n.com/openstack-lumberjack-part-3-logstash-and-kibana/
[16:19] <jcastro> we have all of these components charmed, I think having an interface for this would be <3
[16:20] <zul> jcastro:  everytime i see java i cringe
[16:20] <jamespage> jcastro, awesome
[16:20] <jcastro> you won't care, I have a charm (handwave)
[16:20] <jamespage> just raise a bug against the nova-cc charm
[16:20] <jcastro> ta
[16:21] <lordievader> Good evening.
[16:46] <jdstrand> hallyn_: fyi, I file bug #1331081
[16:46] <jdstrand> filed*
[16:47] <jdstrand> IMO, this should be fixed for 14.10
[16:48]  * hallyn_ makes no comment
[16:49] <shrikant> is any one faceing problem network card ,i am getting ip address from dhcp server but not able to ping to any network
[16:52] <RoyK> shrikant: can you ping machines on the same network, such as the dhcp server?
[16:52] <RoyK> (which usually is on the same network :P)
[16:52] <shrikant> hey , how are you
[16:53] <shrikant> i am not able to ping even gateway
[17:07] <shrikant> brb
[17:48] <stickystyle> Does anyone have information about the execution order of cloud-init? Like what is the order that it evalulates the commands in the user-data file?
[18:18] <robsparker> Trying to install ubuntu server 14.04 and it has been stuck at 'Configuring systemd-services' for several hours and there is no longer any apparent disk activity.  Any ideas?
[18:20] <sarnold> robsparker: can you get another virtual console with control+alt+f1 or f2 etc?
[18:20] <sarnold> robsparker: if so, is 'ubuntu-bug' available? :) It'd be nice to get a bug report about that
[19:08] <doge_wow> genii: so, you're saying it's not a 100% water proof to upgrade to a major version by SSH? What makes it not entirely 100% safe? BTW, u finished lunch already?
[19:10] <genii> doge_wow: It's just that because you never know what custom things someone has done to their server may break an upgrade. Like a PPA that was added, or a piece of software was recompiled, or permissions on a system folder were altered, etc. But for the most part on a stock system with stock repositories and apps it's safe.
[19:27] <doge_wow> genii: this system has not yet been build. I am planning to install an operating system on this server which can be upgraded on-the-fly, remote with an SSH connection. Is this feature to upgrade major versions remotely officially supported by Ubuntu Server or is it safe "because there are rarely any complaints due to major upgrades"?
[19:33] <genii> doge_wow: If you are running a stock Server then LTS to LTS upgrades are fully and officially supported.
[19:33] <doge_wow> genii: what do you mean by stock Server?
[19:34] <dasjoe> doge_wow: see https://help.ubuntu.com/14.04/serverguide/installing-upgrading.html
[19:35] <genii> doge_wow: By stock, i mean that you have not added any PPAs to the system, not running any software that you compiled yourself manually, not altered any of the default system permissions, and so on.
[19:36] <genii> Basically a vanilla install with whatever apps you needed got installed from the standard Ubuntu repositories
[19:37] <sarnold> self-compiled stuff will usually not affect much, but ppas can be finicky.
[19:37] <doge_wow> genii: what are PPA's? I never heard about that before.
[19:37] <genii> !ppa
[19:38] <dasjoe> sarnold: as long as nobody does stuff like "sudo gcc -o /sbin/init hello-world.c" ;)
[19:38] <genii> sarnold: If they are put somewhere like /opt or /usr/local not normally an issue, I agree. But some people replace the system binaries with their own stuff on occasion
[19:41] <doge_wow> genii: oic... please correct me if I'm wrong. An OS like Ubuntu Server has many dependencies because it has many automated features, meaning if you tamper with it too much (too many custom changes, compile your own software etc) then a major upgrade cannot be guaranteed 100%. But an OS like Slackware which has little automated features (bare) can be manually tweaked much more and still be able to guarantee major upgrades remotely. Is this th
[19:43] <sarnold> dasjoe,genii, hehe, yeah. caveat compiler? :)
[19:44] <doge_wow> dasjoe: yeah, thnx for the link, I read it.
[19:46] <genii> doge_wow: I can't speak for Slackware.
[19:47] <doge_wow> genii: oki, thnx for your help though. I learned something :)
[23:06] <xibalba> i need a rsyslog example of taking syslog from a specific host and dumping it to a single file