[04:09] <mgi> Hi, I am trying to set up a cluster of nodes using MAAS and juju, and I'm at the point of doing "juju bootstrap", which seems to work. However, "juju status" complains about an invalid SSH key. I have researched all I can on the net, there doesn't seem to be much out there. I guess the first question is: how can I log into a freshly installed MAAS node? SSH or local, I don't mind.
[04:11] <mgi> (I have also set a password for the 'ubuntu' user in /var/lib/cobbler/kickstarts/maas.preseed but I don't know if this is working, as I am unclear on the order of when things run)
[05:24] <hilarie> How can I modify kittenwar (/sbin/iptables -A PREROUTING -s 192.168.0.0/255.255.255.0 -p tcp -j DNAT --to-destination 64.111.96.38) to only effect IPs above 192.168.1.15?
[05:32] <al_nz1> HI guys
[05:33] <al_nz1> could someone please help with a apache2 problem. I think for starters I need to know what file to modify for document root?
[05:34] <al_nz1> and httpd.conf doesnt seem to be the right file?
[05:34] <mvp> you mean to set document root or to edit current document root file
[05:35] <al_nz1> to edit document root. I need to point it to a diff dir
[05:36] <al_nz1> I have got : http://pastebin.com/xWBNGq4C
[05:37] <al_nz1> Or even more info : http://pastebin.com/p7ShTPwW
[05:41] <mvp> httpd.conf should still be valid in 12.04, although it is empty by default
[05:42] <greppy> al_nz1: https://help.ubuntu.com/12.04/serverguide/httpd.html may help
[05:43] <al_nz1> I think I am a bit confused, there seems to be httpd.conf and also /sites-available/default
[05:45] <greppy> al_nz1: yes.
[05:45] <greppy> by default, nothing is in httpd.conf
[05:46] <al_nz1> greppy: yes, but as you can see from my pastebin (the second one at least) I have added some info to httpd.conf and tried to restart apache but get an error
[05:47] <greppy> the current best practice is to specify each vhost/domain in it's own configuration file in sites-available/ and then create a symlink to the file in sites-available in sites-enabled
[05:47] <al_nz1> ...see bottom of pastebin
[05:47] <al_nz1> greppy: hmm, ok. Tho there will only ever be one site on this PC
[05:48] <greppy> ummm
[05:48] <greppy> it complains that Action doesn't make any sense.
[05:48] <al_nz1> greppy: which could be because I shouldnt have edited httpd.conf but apache2.conf instead?
[05:49] <al_nz1> The instructions are located at www.mobotix.com/eng_GB/file/2431175/readme_en.txt
[05:49] <al_nz1> but they relate to Apache on windows :-
[05:49] <al_nz1> :-(
[05:49] <greppy> it looks like you need to enable a module for action to make sense in the config
[05:49] <al_nz1> greppy cgi is enabled by default
[05:51] <greppy> al_nz1: in that link there is a line loading mod_actions.o
[05:52] <al_nz1> greppy: which link?
[05:52] <al_nz1> the link at www.mobotix.com? or the pastebin?
[05:55] <al_nz1> greppy: here is a list of mods enabled
[05:55] <al_nz1> http://pastebin.com/stqKYTiF
[05:55] <mgi> I have an question which is orthogonal to my last: what is the best solution in ubuntu for imaging a cluster of machines? i.e. set up a golden image, then farm that off to a bunch of blades (with the relevant sections modified, such as NIC information, etc.)
[05:55] <greppy> al_nz1: the mobotix
[05:55] <mgi> an = a
[05:56] <al_nz1> greppy: is that the line Action       mxremotepreview /cgi-bin/MxRemotePreview.cgi.exe
[05:56] <greppy> al_nz1: based on the error message, you need to find the module that provides action.
[05:57] <greppy> al_nz1: yes, that is the line that it is complaining about.
[05:57] <greppy> The error message "Syntax error on line 4 of /etc/apache2/httpd.conf" is a pretty obvious clue :)
[05:59] <al_nz1> greppy: for sure, though I was wondering if it was generating a error due to something on the previous line
[06:00] <glosoli> I have some account at Ubuntu server, how do I check my rights to File System ?
[06:01] <al_nz1> so, by the look of google apache2 modules are loaded with sudo apt-get, but how do I work out the module name?
[06:03] <greppy> al_nz1: the module you need, according to the documentation that you are working from is mod_action.o
[06:04] <al_nz1> sure, but the sudo apt-get install mod_action.o doesnt auto complete or work
[06:05] <greppy> al_nz1: right.  apt-cache search mod_action may find what you need. I have never used mod_action or tried to install it.
[06:06] <al_nz1> parser3-cgi - Parser 3, HTML-embedded scripting language (CGI binary)
[06:06] <al_nz1> php5-cgi - server-side, HTML-embedded scripting language (CGI binary)
[06:06] <al_nz1> are the two results?
[06:07] <greppy> al_nz1: looks like it.
[06:07] <al_nz1> greppy: ok, idiot alert here. how do those results help me install mod_action.o ?
[06:08] <greppy> al_nz1: I won't be able to give you step by step instructions on how to get this working, I'm just pointing you in the direction that may find your answers.
[06:08] <al_nz1> greppy: definitly appreciated!
[06:08] <greppy> al_nz1: if you install one of those packages, it may provide that module.
[06:08] <al_nz1> right!
[06:09] <greppy> al_nz1: apt-cache show parser3-cgi
[06:10] <greppy> read what it tells you about the module.
[06:10] <greppy> errr, not module, package :)
[06:11] <mgi> For what it's worth:
[06:11] <mgi> mgi@queeg:~$ ls /etc/apache2/mods-available/actions.*
[06:11] <mgi> /etc/apache2/mods-available/actions.conf  /etc/apache2/mods-available/actions.load
[06:11] <mgi> You should just need to enable it, assuming that's the same module you need.
[06:11] <al_nz1> mgi: seen those, how do I enable them?
[06:12] <mgi> sudo ln -s /etc/apache2/mods-available/actions.load /etc/apache2/mods-enabled/
[06:12] <mgi> sudo ln -s /etc/apache2/mods-available/actions.conf /etc/apache2/mods-enabled/
[06:12] <greppy> or use sudo a2enmod
[06:12] <mgi> Then reload apache.
[06:12] <mgi> Or that :)
[06:12] <mgi> Sorry, I'm used to doing it that way.
[06:12] <al_nz1> mgi: ln -s - that creates a sym link right?
[06:13] <mgi> Correct, but do it greppy's way instead.
[06:13] <greppy> the nice thing about just using 'sudo a2enmod' is it will list what mods are available on the system.
[06:15] <glosoli> Any ideas what can be the command for export Mssql from Ubuntu Server ?
[06:15] <glosoli> I mean there is running mssql in ubuntu server, I want to export some db of it
[06:15] <mgi> glosoli: mssql? Or MySQL?
[06:15] <glosoli> "unixODBC 2.2.11"
[06:15] <glosoli> mssql
[06:17] <al_nz1> well, I think it got it going....http://pastebin.com/2FSxWyze
[06:17] <al_nz1> brb....
[06:20] <al_nz1> back
[06:57] <RoyK> glosoli: use freetds to connect to mssql
[06:58] <RoyK> how to dump a db from there, I really don't know
[06:58] <glosoli> RoyK: There is a problem, I have access to server, but I don't know addresses and ports or name of that mssql database
[06:59] <RoyK> glosoli: then ask the mssql sysadmin ;)
[07:00] <glosoli> Would be hard to get the contacts of one, but still worth trying then
[07:02]  * RoyK has no idea whatsoever how to find that info from sqsh
[07:26] <grendal> this is making me crazy...
[07:26] <grendal> and thats not good
[07:27] <grendal> so i got this script i wrote in php   and it works great..so great that, well we wound u using the hell out of it.  thing is we trigger it with cron.  and...we need it to run more frequently then every min
[11:08] <rbasak> jamespage: I've just checked openmpi status for quantal. Debian have merged our changes and uploaded an openmpi1.6, which we have autosynced (awesome and lots of thanks to Sylvestre!). So I presume I should file a removal request for openmpi1.5, now that it is superceded? Do we need a transitional package to upgrade people more smoothly?
[11:08] <rbasak> jamespage: second, I have a request to put openmpi1.6 in precise-updates. I take it that this is a no-no and I should point to backports?
[11:14] <slg> hi there, i hope someone can help me, i need a init script to start two processes at startup with stop start restart funtkions. Here my first attempt, i have not tried the script yet and i am a noob. http://pastebin.com/c112RR5M
[11:49] <snitride> Hi there, maybe someone can help me with a shellscript i am trying to modify, it should run two processes as daemons on system init, http://pastebin.com/xwkSF1pa would this work? Also i should mention i am a absolute newbie with shellscripts..
[12:25] <zul> good morning
[12:48] <jamespage> rbasak, yes - you should raise a bug for the AA team to get it removed (q 1)
[12:49] <jamespage> q2 as its a new package its not appropriate for -updates - and I think backports does not do new packages either - but I would defer as I've never done one
[12:58] <till__> i seem to have a huge issue with landscape-related scripts on 10.04.4
[12:58] <till__> whenever i try to login, update-motd scripts are run
[12:58] <till__> which end up being zombies
[12:58] <till__> and block my login
[12:59] <till__> i can't figure out why they would block
[12:59] <till__> because i can run these scripts interactively
[13:00] <till__> here is my login attempt (which on the client hangs at 'entering interactive session'):
[13:00] <till__> https://gist.github.com/3097944
[13:00] <till__> this is on the server in question
[13:03] <till__> when i kill the python script running my login completes
[13:04] <nandersson> Hi, is ubuntu-alternate together with Debian Installer and Preseed the most efficient way to do unattended installations? An installation takes a lot longer than a normal installation. Wouldnt it be smarter to do a template that one distribute out by a simple ncat and dd?
[13:05] <rbasak> nandersson: you can do that if you wish, but you do need more inside knowledge of what you need to change to make each machine unique (eg. hostname, but there is more).
[13:06] <nandersson> rbasak, I only know about /etc/hostname, /etc/hosts and /etc/udev/rules.d/70-persistent-net.rules
[13:06] <rbasak> nandersson: the maas project will probably switch to this, and should be the easiest way to manage this stuff in future
[13:07] <rbasak> nandersson: off the top of my head, that should be sufficient. But I might be missing something, and depending on what other packages you install, more may be required.
[13:07] <rbasak> nandersson: if you just want a base image to then install more packages later, look at ubuntu core
[13:07] <nandersson> rbasak, cool! Where can I find information about this maas-project?
[13:08] <rbasak> nandersson: https://wiki.ubuntu.com/ServerTeam/MAAS/
[13:08] <rbasak> (first hit on google, which is nice!)
[13:08] <rbasak> I think they're still using d-i though, but also believe they may be switching to a more dd-style approach soon. Not sure what the roadmap is though.
[13:08] <nandersson> rbasak, Ubuntu Core as in "Ubuntu Mini Remix"?
[13:09] <rbasak> nandersson: https://wiki.ubuntu.com/Core
[13:09] <rbasak> nandersson: but again, you need more inside knowledge to know what to do. Eg. add a user, create /etc/hosts, /etc/network/interfaces, generate locales, etc.
[13:09] <rbasak> basically everything the installer normally does
[13:09] <ironm> hello. I have prepared a local repository for off-line installation of MariaDB 5.5.25 on ubuntu-server 12.04 - http://rsync.it-infrastrukturen.org/public-mariadb/ubuntu/mariadb-5.5.25-ubuntu-12.04-repo.tgz and wrote some description how to use it - http://rsync.it-infrastrukturen.org/public-mariadb/ubuntu/mariadb-ubuntu-local-repo.pdf
[13:09] <ironm> further I have prepared a script and some config file examples for running mysqld_multi on ubuntu-server 12.04 and box with multiply ethernet interfaces - http://rsync.it-infrastrukturen.org/public-mariadb/ubuntu/mysqld_multi_install_db_files.tgz. It would be great if I could get some feedback. Thank you in advance ;)
[13:10] <nandersson> rbasak, Yeah, but those are basically the same if you have ldap-auth
[13:11] <xfxchilde> Current mood: Blow torch + MAAS = Raving lunitic with a smile =DD =/
[13:11] <xfxchilde> Morning :)
[13:16] <xfxchilde> Anyone here try MAAS yet?
[13:26] <jamespage> smoser: around?  12.04.1 meeting in 35 mins and we have not looked at that list yet
[13:30] <andygraybeal_> i'm tryign to add a serial number to my harddrive with libvirt/kvm and i'm running ubuntu 10.04 w/ libvirt 0.7.5  - i'm adding the <serial>WD-WMAP9A966149</serial> tag in the .xml file and redefining the vm, but i cannot see the serial with either "ls -l /dev/disk/by-id/virtio-WD-WMAP9A966149"  or "hdparm -i /dev/vda"
[13:30] <andygraybeal_> does anyone have any suggestions?
[13:31] <smoser> jamespage, sure.
[13:31] <jamespage> smoser, do you have the official link
[13:32] <smoser> https://wiki.ubuntu.com/PrecisePangolin/12.04.1
[13:33] <smoser> (see bug lists there)
[13:33] <smoser> http://status.qa.ubuntu.com/reports/kernel-bugs/reports/rls-p-tracking-bugs.html is what  was mostly going by
[13:33] <jamespage> smoser, yep - thats the one
[13:35] <smoser> jamespage, ok. so how do you want to do this.
[13:35] <jamespage> smoser, I was just thinking
[13:36] <jamespage> smoser, the biggest lists are juju and maas
[13:36]  * jamespage puts them to one side for the moment
[13:37] <smoser> yeah. on that list, there are only like 2 that are not juju/maas.
[13:37] <jamespage> smoser, I think the couchdb stuff is done
[13:38] <jamespage> smoser, couchdb is in universe and kombu only suggests it now
[13:38]  * jamespage marks as fix released
[13:38] <till__> does anyone have any thoughts on my issue?
[13:38] <till__> it seems like update-motd is ran each time i log into a server
[13:38] <till__> and even though i purged the package
[13:38] <till__> it still happens
[13:38] <till__> and i can't figure out where this is triggered
[13:39] <jamespage> smoser, actually no its not -looking at the wrong packages
[13:39] <jamespage> smoser: are we really likely to make such a wide impacting change in 12.04 now?
[13:40] <smoser> jamespage, that is strange.
[13:40] <jamespage> smoser, I think those are a 'Won't Fix'
[13:41] <jamespage> I don't believe they are suitable for SRU
[13:41] <smoser> well, it would be a funny definition of "5 years support" if that included "5 years of support unless we move the package to universe"
[13:42] <jamespage> agreed - marking 'Won't Fix' then
[13:43] <smoser> alright. then we're down to maas, juju, openstack (python-eventlet, keystone)
[13:43] <smoser> so i think we need to somehow address the growing urgency of 12.04.1 with maas and juju teams.
[13:43] <jamespage> smoser, agreed
[13:44] <smoser> at least press a decision has to be made to get them off the list one way or another.
[13:44] <smoser> other bug list?
[13:44] <jamespage> I guess SpamapS is the go-to for juju; is roaksoax the right person for maas?
[13:44] <jamespage> at least as the rep in the team who's most likely to work on them....
[13:45] <jamespage> arosales has already tasked SpamapS, m_3 and jimbaker with reviewing the juju ones
[13:46] <smoser> yes, roaksoax is best contact ther. and i can agree to help him address the list.
[13:46] <jamespage> smoser: OK - so if you pickup with roaksoax on the maas one's I'll hassle SpamapS et al to review the rest
[13:47] <jamespage> smoser, I also see some stuff on http://people.canonical.com/~jamespage/server-sru/precise-sru.html
[13:47] <jamespage> specially anything assigned to ubuntu-server or canonical-server
[13:50] <smoser> what criteria makes the make the james page list?
[13:50] <till__> is there a better place to ask about landscape and update-motd?
[13:50] <smoser> targetted to precise.
[13:50] <jamespage> smoser, server-team subscribed to the package for bug reports, targetted and accepted to precise
[13:51] <smoser> till__, you can ask. there are some landscape people here.
[13:51] <till__> i asked a couple of times, no one seems to reply ;)
[13:54] <xfxchilde> till__, id love to help but i'm unsure =/
[13:54] <till__> yeah, me too
[13:54] <till__> it took me a few hours to get this far
[13:55] <xfxchilde> I've been working on MAAS for about 5 1/2 hours, i'm about to give up myself.
[13:56] <xfxchilde> I think im going to give up o.0 lol
[13:57] <jamespage> smoser, OK - I had a quick scan
[13:57] <jamespage> the cobbler one worries me
[13:58] <jamespage> bug 858867
[13:58] <uvirtbot`> Launchpad bug 858867 in cobbler "XMLRPC  allows unauthed users access to various methods (which it shouldn't) " [Medium,Confirmed] https://launchpad.net/bugs/858867
[13:58] <jamespage> as I don't know how much we can really do about it
[13:59] <hallyn> hm, what's happened to python-software-properties->add-apt-repository ?
[13:59] <jamespage> rbasak, could you take a look at the comments on bug 988819 please
[13:59] <uvirtbot`> Launchpad bug 988819 in mod-proxy-html "[SRU] wrong path to libxml2.so.2 in mod_security - broken by multiarch enabled libraries" [Undecided,Confirmed] https://launchpad.net/bugs/988819
[13:59] <jamespage> hallyn, try python3-software-properties
[14:00] <hallyn> oh ffs
[14:00] <jamespage> lol
[14:00] <jamespage> hallyn, I think that might be a bug but I'm no expert....
[14:01] <till__> i wish i could figure out how to hook into run-parts
[14:01] <till__> to get a log of some kind
[14:01] <hallyn> jamespage: thanks, it'll get me off and running, then i can look into whether there's a bug :)
[14:02] <hallyn> (first i gotta get a set of tests rolling for zul)
[14:02] <zul> <senior burns>excelente</senior burns>
[14:03] <till__> would anyone know which 'event' triggers 'run-parts' to run when a user logs in?
[14:18] <hallyn> jamespage: no, actually python3-software-properties still doesn't have it.  (done by hand for now, will look into it)
[14:18] <jamespage> hallyn, software-properties-common maybe
[14:20] <hallyn> jamespage: looks like, as I see
[14:20] <hallyn> debian/software-properties-common.install:usr/bin/add-apt-repository
[14:20] <hallyn> so i guess it's really a bug in command-not-found
[14:20] <till__> for anyone interested, here is my journey and a solution: http://askubuntu.com/a/162373/11244
[14:20] <hallyn> (can't confirm until my other apt-get finishes)
[14:21] <till__> now trying to figure out how to get rid off this when i bootstrap an instance on ec2 =)
[14:21] <jamespage> hallyn, that gets updated periodically so may be out-of-date for quantal
[14:29] <hallyn> jamespage: oh is that done automatically?  I see it sits in a 'scan.data' file :)
[14:29] <hallyn> nm i guess
[14:41] <rbasak> jamespage: I lost connection, not sure if you got this
 jamespage: I've added a comment
[14:41] <rbasak> I'm not sure if I need to do anything more?
[14:46] <jamespage> rbasak, I think the approach to the SRU needs to be re-considered
[14:46] <jamespage> but lets see what response your comment provokes
[15:45] <zul> hallyn: how is it looking?
[15:46] <hallyn> zul: tests are running...
[15:46] <hallyn> should be awhiel yet
[15:46] <zul> hallyn: ok...i have to wire up the cpu_arm stuff in our libvirt as well
[15:53] <hallyn> zul: qa-regression-tests test-libvirt.py passed
[15:54] <zul> hallyn: cool...ill just wire this up and ill upload
[15:54] <hallyn> sweet
[15:54] <hallyn> no shutdown support yet :(
[15:54] <zul> for lxc?
[15:55] <hallyn> yeah
[15:55] <zul> http://libvirt.org/git/?p=libvirt.git;a=commit;h=97d7f02dcc4848f5f7a64ecd5b32c9beffb7f359
[15:56] <zul> but i think it might be apart of the lxc-reworking
[15:57] <hallyn> zul: it still unconditionally (even in git head) drops CAP_SYS_BOOT
[15:57] <zul> yep
[15:58] <hallyn> so no way can shutdown work in the container
[15:58] <hallyn> or reboot
[15:58] <zul> *sigh* :)
[15:58] <hallyn> zul: when you see a WIFSIGNALED(status) check under that waitpid, then they're trying reboot support
[16:15] <rbasak> SpamapS: do you have any experience with backports? Would openmpi1.6 be a suitable candidate for precise-backports? precise has openmpi1.5 (source package); in quantal we're dropping that and openmpi1.6 from Debian has already synced.
[16:16] <ScottK> rbasak: Would it be a new package in precise then?
[16:16] <rbasak> ScottK: technically, yes.
[16:16] <ScottK> What do you mean by technically?
[16:16] <rbasak> ScottK: it would supercede openmpi1.5 in precise
[16:16] <rbasak> (but we don't need to remove openmpi1.5 from precise)
[16:17] <ScottK> How would it supersede it?  Does it provide the same binary packages?
[16:17] <rbasak> No, it provides a separate set of binary packages.
[16:17] <rbasak> But it's based on the same upstream (just different upstream releases)
[16:18] <rbasak> It's a library, hence the version number in the package name
[16:18] <rbasak> No reverse depends. In Precise, all openmpi depending packages depend on openmpi (no version in name) which is 1.4
[16:19] <rbasak> Reason for 1.6 is improved ARM support
[16:19] <rbasak> Does that make sense?
[16:26] <rbasak> ScottK: is a new package a problem?
[16:26] <ScottK> No.  New package is totally easy.
[16:26] <ScottK> As long as it's tested, it's a great candidate for backport.
[16:27] <ping__> hy
[16:27] <rbasak> Great, thanks.
[16:29] <SpamapS> jamespage: re juju bugs, I'm going to comb through them today
[16:30] <jamespage> SpamapS, that would be fantastic!
[16:33] <jamespage> SpamapS, be harsh - if we are not ever going todo something please mark won't fix  :-)
[16:34] <jamespage> and feel free to bump anything that def won't make 12.04.1 to 12.04.2 or later
[16:35] <SpamapS> jamespage: indeed, I think there are maybe 2 or 3 things we'll fix in the next upload to precise-propsoed and the rest will get bumped
[16:35] <jamespage> SpamapS, ack - please re-target as appropriate...
[16:41] <jcastro> SpamapS: Does this bug exist? http://askubuntu.com/a/162373 if not, what package should it be reported against?
[16:41] <jcastro> basically pretty motd makes things slow
[16:46] <SpamapS> jcastro: I think thats a bug in landscape that may be resolved in 12.04
[16:46] <jcastro> ok, /me investigates
[16:46] <jcastro> thanks
[16:51] <jamespage> adam_g, I assigned you the SRU for openvswitch - hope that was OK - looked like you already had it in hand
[16:52] <adam_g> jamespage: yes, cool
[16:52] <jamespage> adam_g, nice one
[16:56] <adam_g> jamespage: i actually pinged dan out of band about getting the bug updated with the proper SRU template, and i would prepare the package for -proposed.  im sure he and his tema know more about the issue and regression potential than me
[17:03] <jcastro> SpamapS: found it: https://bugs.launchpad.net/ubuntu/+source/pam/+bug/805423
[17:03] <jcastro> thanks
[17:03] <uvirtbot`> Launchpad bug 805423 in pam "pam_motd needs a module option to disable in-line dynamic updates" [Medium,Fix released]
[17:03] <jcastro> not a decent tradeoff imo,
[17:36] <blitzkrieg766> should i enable any automatic updates during install? and if so, why?
[17:36] <blitzkrieg766> all, just security, or none?
[17:41] <genii-around> blitzkrieg766: Usually security is good
[17:41] <blitzkrieg766> danke
[18:10] <zul> hallyn:  0.9.13 uploaded
[18:20] <hallyn> zul: rockin'
[18:28] <RoyK> http://paste.ubuntu.com/1088463/ <-- I thought the RAID had died, five drives died after a VERY BAD PSU, but the ones that mattered, survived \o/
[18:41] <hallyn> stgraber: ok, at this point lp:~serge-hallyn/ubuntu/quantal/lxc/lxc-api-getconfig is working with c->get_item and c->clear_item.
[18:41] <hallyn> stgraber: get_item(lxc.cgroup) is not yet implemented (will return -1)
[18:42] <hallyn> stgraber: that is one remaining to-do, the other is to have c->load_config() clear out any already-loaded configuration.  (unless you think we shouldn't?)
[18:43] <hallyn> stgraber: if lxc-api-getconfig works for you, either let me know and i'll push it into the main api branch, or you can just grab my 0094 patch which includes everything
[18:45] <stgraber> hallyn: I can see some benefits of not clearing the config on load_config(), mostly if you want to stack a bunch of configs, though I suppose clearing by default makes sense
[18:45] <stgraber> hallyn: would adding a parameter for that work for you?
[18:46] <stgraber> hallyn: or alternatively make load_config() never clear the config and support clear_config() to flush everything
[18:50] <hallyn> stgraber: hm, let's leave it as is then.
[18:50] <hallyn> stgraber: a clear_config(0 will be good,
[18:50] <hallyn> i just need to check what else i need to clear besides what i already support with clear_config_item()
[18:51] <hallyn> but so you don't need that right now?  (i was getting doubled-up configurations in my test case, which made my test case counting nics returned by get_item("lxc.network") fail :)
[18:52] <hallyn> and, separately, do you need get_item(lxc.cgroup.whatever) right now?  If not i'd like to wait on that
[18:52] <hallyn> in other words, i'd like to be done and let you use it for a while :)
[18:54] <stgraber> hallyn: I'll have to look at the branch. I need a way to access the value of a lxc.cgroup... entry, though not its live value (from /sys/fs/cgroup) yet
[18:55] <stgraber> so far I've been lucky enough not to get in a case where multiple load_config would break the world for me, but I can certainly see this happening so it's certainly going to be on the list of things we need to have before pushing upstream (or in Ubuntu)
[18:55] <hallyn> stgraber: hm, ok.  then maybe a 'get_item("lxc.cgroup.devices.allow")' will suffice, and be easy enough to do
[18:56] <stgraber> hallyn: that's why I made the distinction between (get/set/clear)_config_item and (get/set)_cgroup_item in my mail ;) the former only deals with the lxc config without doing any live change, the later only deals with the live values without touching the config
[18:57] <hallyn> stgraber: ok, lemme whip that up then
[18:57] <hallyn> then i'll check it into your branch
[18:57] <stgraber> cool, thanks
[18:58]  * stgraber gets back to reviewing the 120 bugs targeted for 12.04.1...
[18:58] <hallyn> chao
 that doesn't look right :)
[18:59] <hallyn> i guess i'm hungry.  more chow.
[19:46] <stgraber> jamespage: looking at bug 944546 shouldn't that be moved to precise-updates?
[19:46] <uvirtbot`> Launchpad bug 944546 in libcommons-cli-java "StringIndexOutOfBoundsException in HelpFormatter.findWrapPos" [High,Fix released] https://launchpad.net/bugs/944546
[19:46] <stgraber> jamespage: it's currently targeted to 12.04.1 but you marked it incomplete as it's not clear it's actually a problem with anything in the archive
[19:46] <stgraber> (in such case, the priority should probably be changed to something lower than high)
[19:49] <blitzkrieg766> when I try to add a windows xp box to a domain i just created, it says "user not found" what am i doing wrong?
[19:49] <blitzkrieg766> ubuntu server 12.04
[20:00] <RoyK> blitzkrieg766: samba domain?
[20:05] <blitzkrieg766> yes ROYK, samba domain
[20:06] <blitzkrieg766> i think somehow i've screwed up the permissions,
[20:06] <RoyK> never tried samba domains, sorry
[20:06] <blitzkrieg766> bummer
[20:06] <blitzkrieg766> lo
[20:06] <blitzkrieg766> lol
[20:06]  * RoyK uses windoze machines for that sort of stuff
[20:06] <RoyK> AD et al
[20:06] <blitzkrieg766> yep
[20:07] <blitzkrieg766> thats what i WAS doing.. but i've found this lovely company who stays away from the MS stack if possible
[20:07] <blitzkrieg766> and... i.... LOVES it
[20:08] <blitzkrieg766> freebsd as a host, most of the vm's are either freebsd or ubuntu server, it rocks
[20:11] <blitzkrieg766> ok, another question....
[20:11] <blitzkrieg766> is there by default a group called "sysadmin" in ubuntu server?
[20:11] <blitzkrieg766> 12.04
[20:13] <sw> blitzkrieg766: why, what exactly are you trying to do?
[20:14] <blitzkrieg766> well, i've setup samba as a DC for windows machines,
[20:14] <blitzkrieg766> and when I try to add an xp station to the domain, i get a "user not found" error
[20:14] <blitzkrieg766> i think the group for the user's credentials i'm using is wrong,
[20:15] <blitzkrieg766> the instructions said I had to also add that user to the "admin" group but there isn't one
[20:15] <blitzkrieg766> and the page says admin should be there by default,
[20:15] <sw> $ sudo cat /etc/group | grep sys
[20:15] <sw> sys:x:3:
[20:15] <sw> syslog:x:103:
[20:16] <sw> blitzkrieg766: admin -> sudo now, iirc
[20:17] <souliaq> the mediawiki package for "precise" is version 1:1.15.5-7. Anyway to get more updated package (1:1.19.x) for 12.04 server?
[20:18] <blitzkrieg766> sw, ?
[20:18] <blitzkrieg766> your syntax confuse me a little... i entered that command,
[20:18] <sw> blitzkrieg766: what command?
[20:19] <blitzkrieg766> sudo cat etc etc etc
[20:19] <sw> souliaq: build it? https://launchpad.net/ubuntu/+source/mediawiki/
[20:19] <blitzkrieg766> i guess what i'm saying is I didn't understand the last part of your instructions
[20:19] <sw> blitzkrieg766: you mentioned a 'sysadmin' group, so I listed all default groups and searched for ones containing sys showing that there isn't a 'sysadmin' group
[20:19] <blitzkrieg766> ahhh
[20:19] <blitzkrieg766> ok
[20:19] <sw> blitzkrieg766: there used to be an 'admin' group, but that is now 'sudo' if I remember correctly. someone else will confirm
[20:20] <sw> I remember something along those lines, anyway ...
[20:31] <jamespage> stgraber, I think you are right - as I got no response from the original reporter....
[20:32] <jamespage> I'm going to 'Won't Fix' it
[20:32] <stgraber> thanks
[20:35] <hallyn> stgraber: pushed to your tree
[20:35] <hallyn> stgraber: for get_cgroup_item (live values), it might be best to use libcgroup.  hope you're in no hurry, if you are shout.
[20:36] <hallyn> i also assume you're happy with your python clone for now.  again shout if not
[20:38] <stgraber> hallyn: yeah, get_cgroup_item isn't high priority, neither is clone() as these should be fairly easy to implement in python (well, clone was at least)
[20:39] <hallyn> stgraber: wonder how bad it woudl taste to have liblxc link against libcgroup to reuse cgget/cgset.
[20:39] <hallyn> but, worry about that later.
[20:41] <stgraber> hallyn: AFAICT libcgroup1 doesn't depend on anything other than libc so it should be a cheap dependency to add to lxc
[20:43] <hallyn> stgraber: now that libcgroup doesn't introduce its upstart jobs and daemons, yes
[20:44] <jMCg> byoby-enable enables stuff in my homedirectory, which isn't so nice, because my homedir is on NFS...
[20:45] <stgraber> hallyn: daemons and upstart jobs that's libcgroup-bin right? we just want libcgroup1, we don't need libcgroup-bin in lxc
[20:46] <stgraber> hallyn: testing on 12.04, libcgroup1 installs fine and doesn't pull libcgroup-bin, it only contains the library (/lib/libcgroup.so.1.0.37)
[21:03] <hallyn> stgraber: oh yeah.  i was first thinking of the 'cgget' and 'cgset' binaries, for use by python.  but yeah for the c lib i wouldn't need that.
[21:04] <stgraber> hallyn: oh right, yeah, for python I'd probably just parse the cgroupfs, I don't like the idea of depending on libcgroup-bin (especially as most of my test machines are 12.04 ;))
[21:06] <stgraber> jamespage: should bug 920197 be assigned to zul? (for the swift task)
[21:06] <uvirtbot`> Launchpad bug 920197 in swift "[SRU] webob last stable version 1.1.1 response header bug" [Undecided,Confirmed] https://launchpad.net/bugs/920197
[21:14] <hallyn> jdstrand: hey.  looking at bug 1015154.  it looks to me like 'fixing' that requires basically giving up on restricting libvirt-qemu at all?
[21:14] <uvirtbot`> Launchpad bug 1015154 in libvirt "virt-aa-helper refuses to create unix socket for a serial port" [Undecided,New] https://launchpad.net/bugs/1015154
[21:17] <jdstrand> hallyn: why do you say that? virt-aa-helper 'just' needs to be made aware of the xml and add the path to /etc/apparmor.d/libvirt/libvirt-*.files
[21:18] <jdstrand> just like we do with the others
[21:18] <hallyn> jdstrand: oh.  yeah.
[21:19] <jdstrand> it shouldn't be hard to do (thought it isn't something I can do anytime soon (vacation coming up))
[21:19] <jdstrand> s/thought/though/
[21:20] <hallyn> jdstrand: great, thanks.  No I was just thinking wrongly.  Will mark it confirmed/med
[21:20] <jdstrand> cool
[21:21] <hallyn> stgraber: for bug 1013549, we really should get the one-line patch from the description into quantal next time we update
[21:21] <uvirtbot`> Launchpad bug 1013549 in lxc "lxc-clone fails for xfs fs on lvm" [Medium,Triaged] https://launchpad.net/bugs/1013549
[21:21] <hallyn> hm, well i guess we can't take it as is, but...
[21:21] <hallyn> lemme just send a patch upstream for now
[21:39] <stgraber> hallyn: bug 974584 is targeted for 12.04.1 but not assigned to someone, are you the one working on it?
[21:39] <uvirtbot`> Launchpad bug 974584 in lxc "Semaphores cannot be created in lxc container" [High,Fix released] https://launchpad.net/bugs/974584
[21:40] <hallyn> stgraber: no.  it's fixed in debian.  nothing for me to do
[21:40] <hallyn> stgraber: (see my last commen tin the bug)
[21:41] <stgraber> hallyn: who's taking care of merging the package from Debian, cherry pick the fix and push the SRU to proposed?
[21:42] <hallyn> stgraber: no idea, i was hoping slangasek would respond
[21:43] <hallyn> stgraber: i can give them a shot but dont' have upload rights to that one
[21:43] <hallyn> also there is a pretty wide ubuntu delta, which is intimitading
[21:45] <stgraber> hallyn: ok, I'll assign to canonical-foundations and bring it up at our next meeting
[21:45] <jamespage> stgraber, if you don't get a response on bug 993249 this week I'm going to steal it - its a ridiculously easy and low impact fix
[21:45] <uvirtbot`> Launchpad bug 993249 in activemq "[SRU] activemq fails to start on Ubuntu 12.04" [Wishlist,Fix released] https://launchpad.net/bugs/993249
[21:46] <hallyn> stgraber: thanks.  if you *prefer* to assign it to me i'll try it on monday.
[21:46] <stgraber> jamespage: steal all the bugs you want, I certainly won't complain :)
[21:52] <wedgie> i've got 'up route add -net 10.0.0.0/8 gw 10.0.0.1' in /etc/network/interfaces under the static config for the interface that has that gateway, but when i restart networking (or reboot the server) the route doesn't appear. Where did i go wrong?
[21:52] <wedgie> pretty vanilla install of 12.04 server
[21:59] <smoser> hallyn, around ?
[22:05] <hallyn> smoser: partly (only to d/l sles 10 for a qemu bug reproduce - so laggy network)  what's up?
[22:05] <smoser> i boot an instance
[22:05] <smoser> put eth0 on a bridge
[22:05] <smoser> then try to run a kvm with tap attached to that bridge.
[22:06] <smoser> using http://smoser.brickies.net/git/?p=tildabin.git;a=blob;f=xkvm;hb=HEAD
[22:06] <smoser> i'm trying to network boot off of a tftpd that is running in the host
[22:06] <smoser> where host here means instance
[22:06] <smoser> :)
[22:07] <smoser> and immediately it locks up
[22:07] <smoser> ie, networking dies. reboot brings it back.
[22:07] <hallyn> smoser: are you on q?
[22:07] <smoser> that was on q, yes
[22:07] <smoser> but i assume i'm just doing something stupid
[22:08] <hallyn> smoser: yeah, come to think of it, maybe
[22:08] <hallyn> smoser: i'm pretty sure the dhcp request goes to your 'host' dhcp server, which messes you up
[22:08] <hallyn> smoser: how about creating a br1, starting a dhcp server in a container on there,
[22:08] <hallyn> and then net-booting the kvm instance from that bridge?
[22:09] <smoser> thats actually a good solution i think
[22:09] <smoser> basically, what i was *trying* to do was:
[22:09] <smoser>  * boot instance
[22:09] <smoser>  * apt-get install maas
[22:09] <smoser>  * kvm ....
[22:09] <smoser> such that the kvm would pxe boot from the maas server on that system
[22:10] <smoser> (and generally look like a new system to maas)
[22:10] <smoser> so i think your suggestion of doing the 'apt-get install  maas' inside a lxc container  is quite reasonable.
[22:11] <hallyn> smoser: yeah so i think something like http://paste.ubuntu.com/1088856/
[22:11] <hallyn> (you can just 'lxc-console -n maasserv' to install maas and get the ip address)
[22:12] <smoser> does traffic get out of that network?
[22:12] <hallyn> smoser: no
[22:12] <smoser> hm..
[22:12] <smoser> i would kind of need it to (nat would be fine)
[22:12] <hallyn> smoser: you can add a second nic on br0 in the container
[22:12] <smoser> ie, to get to the archive.
[22:12] <hallyn> (or lxcbr0)
[22:13] <adam_g> zul: 'round?
[22:13] <hallyn> smoser: come to think of it, nat won't forward dhcp right?  (i could be wrong)  if so, then you could just use lxcbr0 as usual and hook kvm into that
[22:13] <smoser> hallyn, right.
[22:13] <hallyn> smoser: do you only need the open net in the maas container, not on the nodes?
[22:13] <smoser> but who answers the dhcp on the lxcbr0 ?
[22:13] <hallyn> your container
[22:13] <smoser> but who for the first?
[22:14] <smoser> clearly my container doesn't answer for itself, does it?
[22:14] <hallyn> oh.
[22:14] <hallyn> yeah, we have a dnsmasq running for it
[22:15] <smoser> hm..
[22:15] <smoser> i have to play with this.
[22:15] <hallyn> smoser: coudl try http://paste.ubuntu.com/1088862/
[22:15] <smoser> what does the nat?
[22:15] <hallyn> iptables
[22:15] <hallyn> /etc/init/lxc-net.conf
[22:16] <hallyn> smoser: but using http://paste.ubuntu.com/1088862/ you'd have lxcbr0 available to the container, while kvm insances would only have br1 to the container for dhcp
[22:16] <hallyn> so the maas container would have to be set up to use eth1 for guests
[22:16] <smoser> hallyn, ok. i'll have to play with this. thanks though.
[22:17] <hallyn> smoser: np.  i may have to try it myself, sounds useful.
[22:19] <smoser> hallyn, so i could just set up a bridge, put the maas lxc container in it.
[22:19] <smoser> and have it serv dhcp, telling stuff the default route is the "host"
[22:19] <smoser> or whatever.
[22:19] <smoser> basically get nat that way
[22:19] <smoser> right?
[22:22] <smoser> hm.. hallyn ^
[22:24] <wedgie> looks like the answer was to move it above the "dns-nameservers" and "dns-search" parameters
[22:54] <rockets> I just did a ufw default deny incoming, and then ufw allow from MYIPADDRESS
[22:54] <rockets> and everybody remote can still ping/reach the server
[22:55] <hallyn> smoser: sorry, my irc window ssh session apparently locked up and it just looked like noone was talking
[22:55] <hallyn> smoser: but no, i don't think what you're saying is quite wha ti'ms aying
[22:55] <hallyn> i'm saying, have the maas container have two nics: one on either br0 bridged with eth0, or on lxcbr0 (which is nat'ed to eth0)
[22:55] <hallyn> the other nic on br1, which is private
[22:56] <hallyn> then run your kvm VMs with tap on br1 only
[22:58]  * hallyn out
[22:58] <hallyn> smoser: (ping if that wasn't clear)
[23:02] <smoser> hallyn, hm..
[23:02] <smoser> but why do i need the complexity
[23:22] <smoser> hallyn, well, if you return, this seems like it should work for what i was wanting.
[23:22] <smoser>  i create a bridge (using libvirt for simplicity... yes, i think that sounds funny too)
[23:22] <smoser> but anyway: http://paste.ubuntu.com/1088955/
[23:23] <smoser> then i would boot a lxc container for my maas server on that bridge.
[23:25] <smoser> it woudl have to statically configure its networking including dns.  the default route and gateway would be 10.0.5.1.
[23:26] <smoser> then, it can provide the dhcp server on that network and tell others to use 10.0.5.1 as the default route and even as dns. (libvirt runs a caching only dnsmasq for me there).