[00:02] <harushimo> if I setup a samba share on my desktop and I want to acces that on my virtual machine, how would I do that?
[01:07] <TransX2> soren, I've done a few things already like setup my linux server, forwarded port 80 to my servers local ip, and setup a subdomain.domain with my external ip. The problem is the domain doesn't work and I can only access my server through my external ip. Anybody have any idea what I did wrong?
[01:07] <TransX2> *So
[01:07] <TransX2> Sorry soren, didn't mean to tab you...
[01:14] <TransX2> noone around?
[03:03] <riz0n> Hi, I am having an issue with Ubuntu 12.04 and postfix. I mv'd master.cf and main.cf and then run sudo dkpg-reconfigure postfix ... got it up and running. Dovecot IMAP is letting me login fine, but postfix is saying: "None of the authentication methods supported by this client are supported by your server." Can someone help me track down the problem wih postfix and get it running properly?
[03:22] <ScottK> riz0n: The ubuntu server guide has a good discussion about that.  I'd read up on setting up SMTP auth there and see if that helps.
[09:07] <martas> hi ppl, have a question.. i need to set up rsyslog to log messages from specific host to specific dir.. it seems i made some mistake.. file and directory was created, but no messages in there.. any tip or clue how exactly set up this thing?
[10:48] <martas> i'm using this syntax from documentation: if $fromhost-ip startswith '192.0.1.' then /var/log/network1.log     & ~        but still i'm getting syntax error... any advice?
[11:07] <StevenR_> martas: which documentation?
[11:07] <StevenR_> martas: that just looks wrong
[11:08] <martas> StevenR_:  http://www.rsyslog.com/storing-messages-from-a-remote-system-into-a-specific-file/
[11:09] <StevenR_> martas: pastebin your config file
[11:09] <martas> StevenR_: ok.. sec..
[11:12] <martas> StevenR_: http://pastebin.com/nD3zqJ9K
[11:19] <rbasak> Should bug 1014044 be an SRU candidate for precise? I'm not sure.
[11:20] <StevenR_> martas: are you sure "isequal" is valid? Maybe try == ?
[11:21] <StevenR_> martas: though actually, you just might need to move it to around line 57
[11:21] <rbasak> Looks like quite a bit of work though. Quantal would have to carry a delta too. Might be best to skip it given there's a workaround.
[11:24] <martas> StevenR_: interesting... it works now... after move... hmmm thx man
[11:27] <StevenR_> martas: there were some other rules before 57 to do with users, setup, ports, etc... those needed to be read first
[11:28] <martas> StevenR_: ok.. i was wrong.. i thought according to manual that it should be after allowing TCP input
[11:35] <feisar> hi, what's the recommended way to connect a headless ubuntu server to a wireless network?
[11:41] <onekenthomas> ssh?
[11:43] <rbasak> feisar: /etc/network/interfaces? See /usr/share/doc/wpasupplicant/README.Debian.gz
[11:43] <rbasak> The recommended way would be to use a wired NIC for a server of course :)
[11:55] <feisar> rbasak: thanks, so a combination of /etc/network/interfaces and wpasupplicant?
[11:56] <rbasak> feisar: that's what I would do. Note that you just need the wpasupplicant package installed. Then your entire configuration can be in /etc/network/interfaces without calling wpa_supplicant directly or anything like that. Nice and clean.
[11:56] <feisar> rbasak: oh great stuff, cheers : )
[13:01] <hallyn> stgraber: any comment on the last comment in bugs 928524 and 231060 ?
[13:22] <hallyn> zul: /lastlog hallyn 10
[13:22] <hallyn> oops
[13:22] <hallyn> hm.  been awhile since using irssi.  trying to recall...
[13:22] <hallyn> oh well.
[13:22] <zul> use a sane client like bitchx :)
[13:23] <hallyn> zul: are you going to have time to test the qemu-kvm from my ppa this morning?
[13:23] <hallyn> zul: if i switch i'll switch back to sic :)
[13:23] <zul> hallyn: if i can test it this afternoon
[13:24] <stgraber> hallyn: commented. I believe this change will fix systems that are currently broken, so even though dnsmasq's behaviour will indeed slightly change, it'll still be a lot better than what we have now
[13:29] <hallyn> zul: thanks.  (if you won't be able to pls let me know and i'll push what i have)
[13:29] <zul> hallyn: ok
[13:29] <hallyn> stgraber: but that's the definition of a regression right?  it'll fix a few broken systems, but risk breaking currently workign systems?
[13:29] <hallyn> is it inappropriate for sru then?
[13:30] <hallyn> if my wireless nic comes up late, will dnsmasq not find it?
[13:30] <hallyn> of course right on call we get another dup (bug 1014576)
[13:31] <stgraber> hallyn: I don't think it risks breaking currently working systems as systems don't currently work when dnsmasq and lxc are installed
[13:31] <hallyn> oops, that's not a dup, nm
[13:31] <hallyn> ah, right
[13:31] <stgraber> hallyn: and for systems that don't have lxc installed won't be affected, neither will these that have lxc installed but not dnsmasq
[13:32] <hallyn> dnsmasq isn't soon going to be enabled by default everywhere?
[13:32] <hallyn> i'm sorry, i'm too fuzzy on our dns plans
[13:32] <stgraber> no, we have dnsmasq-base installed everywhere, not dnsmasq itself
[13:32] <stgraber> dnsmasq-base gives you the dnsmasq binary that you have to spawn manually, dnsmasq gives you the daemon that's breaking with lxc
[13:33] <stgraber> and when calling dnsmasq directly (like NM, lxc, libvirt, ... do), it's not parsing /etc/dnsmasq.d, so there won't be a change of behaviour for these
[13:34] <stgraber> the change of behaviour will only be for these that explicitly install the daemon
[13:36] <hallyn> stgraber: thanks
[13:40] <hallyn> stgraber: for premounting cgroups, I think that's best done with a lxc.hook.mount script.  (so i'm going to mark it blocked;  i intend to put a hooks patch into our package in a few days if there's no upstream feedback)
[13:40] <hallyn> shout if you disagree
[13:40] <hallyn> right now i'm going to make an uptreamable apparmor patch
[13:42] <hallyn> hm, but i so wanna do the userns patch
[13:44] <stgraber> hallyn: that's for premounting the right subset of the cgroup filesystem to allow for nested containers while not letting the container change its own cgroup, right?
[13:45] <hallyn> right
[13:46] <hallyn> it needs to be a script to allow for flexibility in mounted cgroups
[13:46] <hallyn> and has to be done after /sys is mounted
[13:48] <stgraber> right, I guess I'll have to write an lxc-with-nesting.conf template in /etc/lxc/ to use my custom apparmor profile and use that hook, sounds like it'll work as a good example of how to tweak your containers :)
[13:54] <zul> Daviey/jamepage: what are we tagging MIR bugs with again?
[13:55] <Daviey> zul: we don't...
[13:55] <zul> i thought we did
[13:56] <Daviey> zul: it needs to be a package which has a  ~ubuntu-server package subscription, and ~ubuntu-mir subscribed to the bug
[13:56] <rbasak> hey zul
[13:56] <rbasak> zul: https://bugs.launchpad.net/ubuntu/+source/openldap/+bug/82853/comments/26 - any reason why we can't build slapd-smbk5pwd now that heimdal is in main?
[13:56] <zul> rbasak: not that i know of
[13:57] <Daviey> rbasak: I thought that one was assigned to dendro-afk
[13:57] <Daviey> hah, he was removed.
[13:58] <zul> Daviey: ack...another one for your radar:  bug #1014657
[13:58] <zul> rbasak: in fact i dont see why our openldap package has to be so specialized either compared to debian
[13:59] <Daviey> Matthias :)
[14:00] <Daviey> zul: can you comment what work you did for that MIR so far?
[14:00] <zul> Daviey: subsscribed ubuntu-server, subscribed ubuntu-mir, check deps
[14:01] <rbasak> zul: well you did the merge :-)
[14:01] <zul> rbasak: well yes
[14:51] <zul> bdmurray: ping when you are around
[15:04] <bdmurray> zul: hi
[15:04] <zul> bdmurray: sorry to bug you about this but can you approve nova we would like users to get testing with the new snapshots
[15:05] <bdmurray> zul: its fine, I'm really not comfortable with some of these bugs that have "Test Case: none" in them
[15:06] <zul> bdmurray: so if something that we cant test what should i put in there then?
[15:07] <bdmurray> zul: could you explain how bug 967931 is not testable?
[15:09] <zul> bdmurray: hmm...we can run the testsuite for that bug
[15:11] <zul> bdmurray: updated
[15:12] <bdmurray> zul: and in bug 975043 there is a comment that this is not supported out of the box but how does that make it untestable?
[15:13] <zul> bdmurray: we dont have any documentation about it on ubuntu
[15:13] <zul> bdmurray: its a feature that we dont support really
[15:13] <bdmurray> zul: then why are you SRU'ing it?
[15:14] <zul> bdmurray: because its apart of the snapshot that got inherited by upstream
[15:17] <roasted> hello!
[15:18] <Daviey> bdmurray: ey
[15:18] <Daviey> Hey*
[15:18] <roasted> I was tinkering with a few different web services on my server... apache, nginx, etc... I think I goofed cause now I want to revert everything back to a fresh apache install.
[15:18] <roasted> When I install it, it doesn't generate anything in /var/www lke yout ypically expect.
[15:18] <bdmurray> Daviey: Hi
[15:19] <Daviey> bdmurray: So.. What we are trying to do is.. Have significant involvement in the upstream stable trees (zul and myself are both reviewers.)
[15:19] <Daviey> bdmurray: Push any Ubuntu fixes back to upstream where we can
[15:19] <Daviey> Do significant CI testing at both upstream and distro level
[15:19] <Daviey> bdmurray: Then take a whole snapshot, and SRU it.
[15:20] <Daviey> (including fixes that don't impact Ubuntu.. but crucially, do not regress either)
[15:20] <bdmurray> Daviey: I'm the new guy here but taking a whole snapshot and SRU'ing it sounds like a "micro release" to me and would fall under the micro release exceptions category
[15:20] <bdmurray> https://wiki.ubuntu.com/StableReleaseUpdates#Special_Cases
[15:20] <zul> yeah we are working towards that
[15:20] <Daviey> bdmurray: we submitted for one :)
[15:22] <Daviey> (context https://lists.ubuntu.com/archives/ubuntu-devel/2011-November/034509.html)
[15:24] <Daviey> bdmurray: and https://lists.ubuntu.com/archives/technical-board/2011-November/001142.html
[15:41] <hallyn> stgraber: d'oh, I see that src/lxc/conf.h is already #including stdbool.h - so I can clean that bit of the lxcapi code up (to use real 'bool' instead of an int).  I'll wait though until I hear more from you about the python bit, unless you say now is a good time
[15:42] <stgraber> hallyn: the python bits are mostly done but it's extremely easy to rebase on any change you do, so go ahead
[15:43] <stgraber> hallyn: also, can you fix the lxc.h/lxccontainer.h situation? IIRC you said that lxc/lxc.h should give me the lxc_container struct, but that's not currently the case, I need lxccontainer.h which isn't in the lxc-dev package by default :)
[15:44] <stgraber> hallyn: IIRC from what I did on Friday, python-lxc covers everything except for the variable arguments of start() that I still need to write magic to parse/cast and send to start() (with all the fun memory management stuff ;))
[15:57] <hallyn> stgraber: ok, lemme wrap up my apparmor bits as they are and i'll fix the api (or break it completely, depending on what sort of week this is gonna be)
[15:59] <stgraber> hallyn: ;)
[16:00] <mdeslaur> hehe
[16:02] <zul> hallyn: where is the qemu-kvm stuff?
[16:05] <stgraber> hallyn: when you're done moving things in the lxcwithapi branch, I'll work on a lxc test package (not called lxc2 this time) that bundles the API changes and the python module, so it can easily be tested by highvoltage and the others who want to test the python stuff
[16:17] <hallyn> zul: ppa:serge-hallyn/virt
[16:17] <hallyn> stgraber: awesome, thanks
[16:26] <SpamapS> can somebody who is an admin for ubuntu-server please add a bug subscription to mysql-5.5?
[16:30] <zul> SpamapS: done
[16:30] <SpamapS> zul: thanks
[16:31] <SpamapS> zul	maybe make me an admin too? :)
[16:31] <TheLordOfTime> lol?
[16:31] <TheLordOfTime> SpamapS:  any idea when/if php 5.4.x is going to be included into Ubuntu?
[16:32] <SpamapS> TheLordOfTime: I actually intend to figure that out this week
[16:32] <SpamapS> TheLordOfTime: I *think* we'll be following everybody else's lead and dropping Suhosin
[16:33] <SpamapS> TheLordOfTime: so as soon as we ACK that, it will be includedASAP
[16:33] <TheLordOfTime> wonderful
[16:33] <TheLordOfTime> lol, damn, i joined #ubuntu-bugs-announce and my ZNC autodetached me from the channel because of flood... XD
[16:34] <zul> SpamapS:  i dunno about that
[16:35] <SpamapS> zul: what's the matter Colonel Sanders, chicken?
[16:35] <zul> SpamapS:  i dont want you to crazy with power
[16:37] <rbasak> While we're talking about php, is bug 1014044 something that we need to fix and/or SRU? Since the web dev community is pretty big on ubuntu server I thought it might be useful to them. But we'd have to carry a delta for it until a fix filters through.
[16:37] <TheLordOfTime> uvirtbot:  that's not a bug
[16:37] <TheLordOfTime> erm
[16:37] <TheLordOfTime> rbasak:  that's not a bug...
[16:37] <TheLordOfTime> rbasak:  in nginx, you have to have proxy_intercept_errors
[16:37] <TheLordOfTime> or...
[16:38] <TheLordOfTime> have php show errors
[16:38] <TheLordOfTime> (turn on show errors in the conf)
[16:38] <rbasak> Apparently php has to write the errors to the right place
[16:38] <rbasak> And upstream have committed a fix
[16:38] <TheLordOfTime> rbasak:  which errors in particular are you talking about
[16:38] <TheLordOfTime> like 500s?
[16:38] <TheLordOfTime> 403s?
[16:38] <TheLordOfTime> etc.
[16:38] <TheLordOfTime> HTTP status codes need proxy_intercept_errors on in the site config
[16:39] <rbasak> I assume PHP level errors. I'm just going by the bug, which I marked as Triaged as there's an upstream bug and fix
[16:39] <rbasak> (and commit)
[16:39] <TheLordOfTime> PHP specific errors are a different issue
[16:39]  * TheLordOfTime looks at the bug
[16:40] <TheLordOfTime> that distinction needs to be made, see my latest comment
[16:40] <TheLordOfTime> if they mean HTTP status codes, there's a workaround, if they mean PHP errors, well...
[16:41] <TheLordOfTime> if there's a regression in the package in the repos that's bad, but the PPA maintained by the NGINX team correctly displays PHP errors in the errors log file
[16:41] <TheLordOfTime> so... *shrugs*
[16:41] <SpamapS> rbasak: looks like a fairly hefty, but straight forward change
[16:41] <TheLordOfTime> (PPA for nginx)
[16:42] <adam_g> anyone else having problems with dnsmasq's DHCP on quantal?
[16:43] <TheLordOfTime> SpamapS:  if you'd like i'll patch and push the debdiff for that patch
[16:43] <TheLordOfTime> i'm in need of stuff to do xD
[16:43] <rbasak> SpamapS: yeah. I was just thinking that ubuntu server in general is a very friendly web dev environment, including PHP newbies. And not all of them would find an error logging workaround. So wondering about SRU, though not sure
[16:43] <SpamapS> rbasak: right. I'm not sure I understand the impact of not having that fetaure, but I do think its just that.. a feature.
[16:44] <TheLordOfTime> SpamapS:  question for you then, would that bug be a wishlist type bug?
[16:44] <SpamapS> not necessarily
[16:44] <rbasak> So as far as I understand the impact, it is that a developer will get confused by not getting PHP error logs. There is a workaround, which is to log to a separate file.
[16:45] <TheLordOfTime> SpamapS:  i'm in agreement with rbasak, as a developer having PHP log to the error.log file is useful
[16:45] <rbasak> I'm not that clear on what sort of error logs though.
[16:45]  * TheLordOfTime has errors in PHP all the time and sees them in error.log
[16:45] <SpamapS> hrm, whats the default error logging config though? Shouldn't PHP be logging errors on the server they're happening on, not via nginx?
[16:45] <rbasak> (now that TheLordOfTime has pointed that out)
[16:45] <TheLordOfTime> rbasak:  hence the comment on the bug
[16:45] <TheLordOfTime> clarification is required
[16:45] <rbasak> Yep, understood
[16:45] <TheLordOfTime> 500 errors and 502 errors should show up in error.log with context
[16:45] <TheLordOfTime> 403s won't
[16:46] <TheLordOfTime> s/won't/shouldn't necessarily/
[16:47] <TheLordOfTime> rbasak:  was this filed against precise?
[16:47]  * TheLordOfTime sees a blatant lack of information about which version of Ubuntu this occurred in
[16:48] <rbasak> Yes
[16:48] <hallyn> stgraber: i *think* the bzr tree should be all set
[16:48] <rbasak> "Since updating Ubuntu" to 12.04 and "Using php5-fpm version: 5.3.10-1ubuntu3.1"
[16:51] <TheLordOfTime> i wish these things were filed against the actual version of the OS, rather than just the source package in [devRelease]
[16:52] <TheLordOfTime> s/version of the/version of the package in the correct/
[16:52] <rbasak> That would be really awkward for the development version, which is where most bugs are worked on
[16:52] <TheLordOfTime> *shrugs*
[16:52]  * TheLordOfTime prefers the security bugs
[16:52] <TheLordOfTime> where *all* the versions of a package get tagged (CVEs)
[16:53] <TheLordOfTime> helps me figure out where to pull from for source when making debdiffs :P
[16:53] <zul> php is genrally evil for those who do not know
[16:53] <rbasak> :-)
[16:54] <rbasak> I avoid it too, but I also understand that a ton of web developers use ubuntu server so it would make sense to keep it in good shape :)
[16:54] <TheLordOfTime> zul:  +1
[16:54] <TheLordOfTime> well, that and tagging the bugs with the actual release...
[16:55]  * TheLordOfTime tends to look at tags when seeing bugs, since 'precise' or 'lucid' seem to be autotagged on bugs when using ubuntu-bug, which most bugs in nginx get filed with
[16:55] <zul> hallyn: seems to launch instances ok
[16:55] <rbasak> There was a good article on this: http://me.veekun.com/blog/2012/04/09/php-a-fractal-of-bad-design/
[16:56] <stgraber> hallyn: cool. Will have a look after lunch.
[16:56] <zul> well lamp stacks without the p is lam and that is one letter away from lame
[16:57] <SpamapS> rbasak: that article is just whining about the realities of evolution  IMO
[16:58] <SpamapS> There is no "design" in PHP. It evolves in the direction its users want it to
[16:58] <SpamapS> frustrating to those who want things all neat and tidy
[16:59] <hallyn> zul: great, thanks
[16:59] <TheLordOfTime> SpamapS:  should i get started on debdiffs for php with the upstream patch?  or do you want me to wait on getting a debdiff ready
[17:00] <rbasak> EOD
[17:00] <TheLordOfTime> :P
[17:01] <SpamapS> TheLordOfTime: I don't think its that critical. It will arrive via upstream fixes..
[17:01] <TheLordOfTime> wonderful, i can take something off of my to do list :P
[17:01] <SpamapS> By all means do it if you want to
[17:01] <TheLordOfTime> SpamapS:  considering its either make the debdiff or sit at the office waiting for a Microsoft tech to respond to an email, i'd rather make the debdiff xD
[17:01] <SpamapS> But perhaps hold off anyway, since I want to see if we can get 5.4 merged soon
[17:02] <TheLordOfTime> indeed
[17:02] <TheLordOfTime> i'd *love* to see 5.4 merged :P
[17:02] <TheLordOfTime> (considering i have at least one patch in there)
[17:11] <zul> SpamapS: 5.4 on precise?
[17:12] <SpamapS> no
[17:15] <zul> ok you got me scared for a sec
[17:24] <GeminiDomino> (12.04 64-bit) Okay, I've got a RAID1 swap partition that seems to have me locked into a mutually exclusive charlie-foxtrot here.  The device was "removed" but can't be re-added even though it's already added but... *gets dizzy and just throws up the pastebin* http://pastebin.com/raw.php?i=Kv4jNF6g
[17:26] <RoyK> GeminiDomino: did you try to --zero-superblock it?
[17:30] <GeminiDomino> I didn't want to do that until I got someone who knew what they were talking about to suggest that it might not be a Very Bad Idea (TM) :)  Can I zero it and add it as a spare like it says, then make it active?
[17:31] <RoyK> just zero the first few megabytes or so and try to add it
[17:32] <GeminiDomino> okay
[17:32] <RoyK> better add it as a full member than a spare, since a spare won't have much of a function without redundancy
[17:32] <RoyK> as in 'hey, that drive crashed, let me take over, huh, from where can I copy the data? ....'
[17:33] <GeminiDomino> It said that zeroing the superblock would add it as a spare, but when I did so, it worked, then immediately rebuilt.
[17:33] <GeminiDomino> So it seems I'm good to go. Molto Grazie
[17:33] <RoyK> :)
[17:37] <axisys> configure: error: no acceptable C compiler found in $PATH
[17:37] <axisys> what pkg to install to get gcc?
[17:37] <axisys> build-essential?
[17:37] <RoyK> yep
[17:37] <axisys> RoyK: thanks
[18:44] <adam_g> zul: fyi im working on getting the folsom-on-precise stuff built and deployed on there atm. ive hit some dnsmasq/nova-network bug /w folsom-on-quantal and really want to be able to check it on precise first
[18:44] <zul> which dnsmasq bug? the vlan one?
[18:46] <adam_g> zul: might be related but AFAICS something different. instances get no OFFERs
[18:48] <zul> adam_g: try the dnsmasq from proposed
[18:49] <adam_g> zul: quantal-proposed?
[18:50] <adam_g> ive only hit it only quantal atm
[18:50] <adam_g> as soon as i get these folsom/precise deploys going i can test on P
[18:50] <zul> adam_g: k ill leave you alont then ;)
[19:40] <reisi> what is the runlevel recovery -boot menu items go, and what is the default runlevel on normal boot entries?
[19:48] <adam_g> jamespage: still around?
[19:53] <fraterm> I'd like to set up bridged networking with a virutal machine running on ubuntu(through kvm).
[19:54] <fraterm> I've got 2 IP addresses, but only one nic.  I have gotten to the point where I have a bridge device set up manually.
[20:00] <RoyK> fraterm: just start virt-manager and assign the nic to br0 or similar
[20:01] <Womkes> If I have a new disk with the following info in parted ( http://paste.ubuntu.com/1047971/ ) and I just want to have one large parition but optimally aligned, do I first need to create a primary partition with certain alignment? If so, how do I do this I keep getting message that the partition is not optimally aligned.
[20:03] <fraterm> RoyK, at that point, if I select the right bridge device, whatever the 'guest' os has set up for a network configuratino (eth0 static blahblahblah) should be pingable browsable and all that?
[20:03] <fraterm> I think I just haven't selected the right bridge device.
[20:04] <RoyK> fraterm: the guest isn't relevant if the bridging is done correctly
[20:07] <fraterm> RoyK, the guest (the virtual machine) would still have to have a /etc/resolv.conf and static network configuration information applied to it though would it not?
[20:09] <RoyK> fraterm: that's just networking
[20:11]  * fraterm nods Alright then.
[20:55] <hallyn> stgraber: hm, i wonder if i should wait until someone responds about the hooks, before sending the apparmor patch upstream.  flooding the list probably just works against us
[20:55] <hallyn> OTOH, it'd be nice to drop 4 patches from our queue
[20:57] <stgraber> hallyn: would be nice to have Daniel look at the list history and merge some of our stuff, the list is getting long enough that I can't remember what was submitted and what wasn't ;)
[21:04] <stgraber> hallyn: what's the status of destroy() and create() in the API? I'm pushing their python implementation for now until we have something in the API. I also added get_ips() as a python function.
[21:07] <hallyn> stgraber: uh, i don't know.  i haven't been thinking about them.  i thought we were going to postpone those until we were sure we needed them
[21:08] <hallyn> bc it's not clear to me how they woudl interact with the scripts (especially lxc-create)
[21:08] <hallyn> woudl they be replacments, or wrappers?
[21:09] <hallyn> stgraber: hm, then maybe i'll send the patch i have, and then tomorrow go through the mailing list archives and make a new git tree with unmerged stuff
[21:11] <stgraber> hallyn: git tree with unmerged stuff sounds good
[21:11] <stgraber> hallyn: http://paste.ubuntu.com/1048114/
[21:16] <hallyn> stgraber: hm, so notice i had 'get_config_item' and 'set_config_item' and 'load_config/save_config()'.  I just have this hunch that the configs are meant to be somewhat independent from rootfs.  maybe that's wrong
[21:16] <hallyn> the reaosn i mention it is that it affects what create means
[21:16] <hallyn> maybe i'm being silly
[21:17] <hallyn> but, from the c api, what i think i'd expect to do is:  (hold on lemme pb it)
[21:19] <hallyn> stgraber: http://paste.ubuntu.com/1048131/
[21:21] <stgraber> hallyn: yeah, my current implementation essentially requires you to flush your container including its config, then call create(), then load the config that create generated, but I'd definitely be interested in being able to create() without flushing the config
[21:22] <stgraber> I guess to get a clean implementation we'd have to essentially create some kind of template API, defining how they get their arguments and what they output
[21:22] <stgraber> with their output usually being a rootfs and a list of configurations keys that need to be set for the template (but not duplicating all the usual ones)
[21:24] <hallyn> stgraber: right.  so what i'm waffling on is, do we wait until we figure that out, or do we risk polluting the api with cheap (but useful!) wrappers around the scripts?
[21:26] <stgraber> hallyn: I guess we can wait for create() at least. Do you see a similar problem with destroy() ?
[21:27] <fraterm> RoyK, well, of all things, virt-manager seems to have a bug in which the Virtual Network Interface selector persists in setting NAT settings rather than changing them to any of the other available Source Device types.
[21:27] <stgraber> my create() implementation will be good enough for the initial proof of concept. Which reminds me I should ask some python warning on import of the lxc module that the API isn't stable yet ;)
[21:29] <hallyn> stgraber: hm, no, i guess no problem with destroy (except symmetry :)
[21:29] <hallyn> stgraber: i'll be leaving soon.  let me think on the create thing and decide tomorrow?
[21:29] <stgraber> sounds good
[21:34]  * fraterm grins as his virtual machine configuration was working just fine.  Tools just don't exactly work the way one would suspect.