[01:30] <maveas> Hi guys.
[01:31] <maveas> Anyone with experience with MAAS?
[01:31] <maveas> I am trying to test MAAS in VirtualBox. The MAAS server is installed but nothing happens when I try to boot my node (PXE)
[01:32] <maveas> Installed maas-dhcp etc.
[01:33] <maveas> Debugged with wireshark. The node sends dhcpdiscover but there is no response from the MAAS. Have no idea what the problem is. The MAAS is totally open (no rules in iptables)
[01:38] <sarnold> maveas: is the networking on your system configured for the virtualbox clients to be able to talk with each other?
[01:38] <sarnold> s/clients/guests/
[01:39] <maveas> It might not (don't know) :)
[01:39] <maveas> If each VM is bridged on the same interface (eth0) should it really be a problem?
[01:40] <maveas> I am able to set a static IP on each VM and access the internet..
[01:41] <sarnold> maveas: hrm, there's the chance that the packets are leaving your local machine's nic...
[01:42] <maveas> local as in ?
[01:42] <sarnold> maveas: try setting up the MAAS master on another physical machine on the physical network? :)
[01:42] <maveas> guest or host?
[01:42] <sarnold> sorry, host.
[01:42] <sarnold> oof. :)
[01:42] <maveas> Ah
[01:42] <sarnold> Or it might be easier to fiddle with the networking to unbridge them
[01:43] <maveas> So even if I tried to setup maas on my host and then boot a guest it would not work?
[01:43] <sarnold> (and no, I don't know the lingo well enough to use the actual technical term then... :)
[01:43] <maveas> That's ok :)
[01:43] <sarnold> maveas: it may not, if the 'bridged ethernet' mode means packets leave your host nic without being subjected to routing 'internal' on the host.
[01:43] <maveas> I'll try with my desktop as a node
[01:43] <sarnold> maveas: woo. nice half-way testing point. :)
[01:44] <maveas> Thank you for your time sarnold. Appreciated :)
[01:44] <sarnold> please do repotr back, I've been meaning to fiddle with maas in kvm for ages now...
[01:44] <maveas> Always online @ irc?
[01:44] <sarnold> maveas: yes
[01:45] <maveas> Ok. I'll try again tomorrow (2:45am here) and report back with my discoveries :)
[01:45] <maveas> Caio
[02:35] <pndemc> is there something like htop where I can view live dl/ul stats?
[02:38] <sarnold> pndemc: I know I've seen something nicer than this: http://iptraf.seul.org/shots/iptraf-iptm1.gif
[02:38] <sarnold> pndemc: but man, I cn't remember the name :(
[08:59] <soren_> hallyn: Not that it matters to the problem at hand, but I'd avoid versioning the dependency from qemu-kvm to qemu-system.
[09:00] <soren_> hallyn: Also, qemu-kvm used to provide both system and user qemu's, didn't it? If so, I'd make qemu-kvm (the transitional package) depend on both, otherwise people using qemu-user will lose functionality in the upgrade.
[09:45] <eagles0513875> hey guys I am setting up a mixed setup environment. client computers are windows based and the server back end is going to be linux based. in terms of user authentication with the network from the front end is RADIUS sufficient for that purpose or is its purpose only for wifi authentication
[10:19] <vezq> it can be used but depending on amount of users, also think what database RADIUS will use for users, e.g. LDAP if there are a lot of users
[10:23] <vezq> also security is one aspect, see this: http://lists.freeradius.org/pipermail/freeradius-users/2012-April/059947.html
[10:34] <peterrus> when trying to upgrade from 11.10 to 12.04 I get:
[10:34] <peterrus> Exception during pm.DoInstall():  E:Could not perform immediate configuration on 'mysql-client-5.5'. Please see man 5 apt.conf under APT::Immediate-Configure for details. (2)
[10:34] <peterrus> that package is not installed
[10:41] <tizz> hi everybody! where is an IP alias (eth0:0) defined if it is not in /etc/network/interfaces? I deleted everything in /etc/network/interfaces but still eth0:0 is up after each reboot, with a static IP assigned. where the heck could this be defined in? thanks a lot!
[10:42] <vezq> tizz: check this directory /etc/network/if-up.d
[10:42] <vezq> also can be some initscript
[10:44] <tizz> vezq, thanks a lot, saved my a lot of time! there was actually a bash script in /etc/network/if-up.d creating the alias.
[11:29] <jamespage> jdstrand, hey - I'm reviewing a MP for a new version of iptables for raring; its the same version as is Debian experimental (1.4.16.3)
[11:30] <jamespage> and thoughts on whether thats a good idea or not?
[12:00] <samba35> how do i get X windows on windows with xming ,what i have to change in putty
[12:03] <greppy> samba35: you need to use port forwarding, specifcally x11.  connection, ssh, x11, check enable forwarding.
[12:03] <samba35> ok
[12:13] <soren_> jamespage: Do you know why it's in experimental rather than in unstable?
[12:14] <soren_> jamespage: 1.4.16.3 is in unstable.
[12:15] <jamespage> soren_, you are quite correct  - I misread the output of rmadison
[12:18] <soren_> jamespage: Back when I maintained iptables, I don't recall ever having any reservations about pulling in new versions.
[12:18] <soren_> fwiw
[12:18] <jamespage> soren_, my reservation was really around it being in experimental still :-)
[12:23] <scalability-junk> hey I'm working on a kickstart file and I have issues with encrypted raid.
[12:24] <scalability-junk> I wanted to use --fstype and --encrypted and --passphrase within raid of the kickstart, but it says sees options are not supported.. any hints?
[12:27] <scalability-junk> am I right to assume that kickstart files are sort of not supported with ubuntu?
[12:41] <jdstrand> jamespage: right, so 1.4.16.3-4 is in unstable. you'll be able to drop a patch or two iirc, so yeah, I think that would be fine
[12:41] <jamespage> jdstrand, ack
[12:42] <Daviey> jamespage: We should probably look at a rsyslog merge aswell soonly?
[12:43] <jdstrand> jamespage: before you upload it, can you at least 'sudo ufw enable' then reboot, then 'sudo ufw status' to make sure it still comes up?
[12:43] <jamespage> jdstrand, OK
[12:43] <jamespage> Daviey, most probably
[12:44] <jdstrand> I know there are a couple upstream bugs for things that iptables has deprecated. it might be noisy, but still work. I'll fix the noisy bits. if it doesn't work, pass it to me and I can fix ufw and upload both at the same time
[12:44] <jdstrand> jamespage: ^
[12:44] <jamespage> jdstrand, gotcha
[12:44] <jdstrand> thanks
[13:29] <hallyn> soren_: it was before oneiric that qemu-kvm provided qemu-user, though
[13:30] <hallyn> soren_: what would be the reason to avoid versioning the qemu-kvm -> qemu-system depend?
[13:31] <soren_> hallyn: You should generally only do versioned relationship if the version really does matter.
[13:31] <soren_> hallyn: ...and I don't see how it does here?
[13:31] <soren_> hallyn: The looser the relationship is, the easier things will be for apt.
[13:32] <hallyn> soren_: well an older qemu-system which doesn't have qemu-kvm wouldn't suffice, is all...
[13:32] <hallyn> but i can drop it
[13:32] <soren_> Oh. I thought qemu-system was a new binary.
[13:33] <hallyn> no qemu-linaro provides it now
[13:33] <soren_> But still:
[13:33] <hallyn> for qemu without kvm
[13:33] <soren_> Erm... /me thinks some more
[13:34] <hallyn> soren_: meanwhile last night i reproduced what i thought was the situation with some dummy pkgs ina container with private mirror...  and all worked fine
[13:34] <hallyn> so i'm going to go pour over the control file and look for another typo
[13:34] <hallyn> :(
[13:34] <soren_> Heh .
[13:34] <soren_> Enjoy.
[13:34]  * soren_ is of little use today due to some sort of flu
[13:35] <hallyn> thanks :)  I really should try to enjoy - it would be fun if i didn't feel like i was on a deadline
[14:20] <caribou> jamespage: howdy, did you see the not from scott Kitterman about the walinuxagent SRU ?
[14:21] <jamespage> caribou, yes - Daviey sorted things out with the SRU team I believe
[14:22] <caribou> jamespage: ah, ok. What was the private bug he was reffering to ? I don't remember seeing one
[14:22] <jamespage> caribou, there is one referenced in the change log
[14:22] <caribou> jamespage: I thought it was the public bug I was working on...
[14:43] <rbasak> smoser: busy? I've got a couple of tools that could go into cloud-utils that I'd like you to look at before I integrate it.
[14:44] <smoser> rbasak, fire away
[14:44] <caribou> jamespage: so looks like they removed the SRU for Quantal for walinuxagnet
[14:44] <rbasak> So the first one is the libvirt cloud-localds wrapper. It takes one of three subcommands, create, destroy and import.
[14:45] <rbasak> First you import <image_name> <filename>. image_name might be "raring", and filename would be your download of the raring cloud image. That'll stick it into libvirt in /var/lib/libvirt/images/raring
[14:46] <rbasak> Then "create <hostname> <base_image>" which will create a cow image /var/lib/libvirt/images/<hostname> and start it. It injects your ssh key, a hostname and avahi-daemon install into the localds
[14:46] <caribou> :12
[14:47] <smoser> rbasak, ok. i hope that we will have wonderful tools for doing the downloading by the end of this cycle, and also tools that will keep your local image  list up to date.
[14:47] <rbasak> After that "ssh ubuntu@<hostname>.local" should work
[14:47] <smoser> so hopefully we can use that.
[14:47] <rbasak> We can add that. I haven't done it yet for a first iteration
[14:47] <smoser> right.
[14:47] <rbasak> destroy gets rid of it completely
[14:47] <rbasak> I have no idea what to call this tool!
[14:47] <smoser> why 'a' a hostname, and 'b' a avahi-daemon ?
[14:48] <smoser> rbasak, well, i started something very similar with lxc
[14:48] <rbasak> I figure that zeroconf is the easiest way to find machines in a test/dev environment. No messing with IPs then. ssh just works too.
[14:48] <smoser> and called it "lxc-cloud"
[14:48] <rbasak> So libvirt-cloud then?
[14:49] <smoser> well, i was thinking the other way.
[14:49] <smoser> localcloud <hypervisor>
[14:49] <smoser> lcloud
[14:49] <smoser> lcloud <kvm> import
[14:49] <smoser> lcloud lxc import
[14:49] <med_> caribou, sounds like the walinuxagent thing is being worked (various ways)
[14:49] <smoser> ?
[14:50] <rbasak> As a single tool or as a wrapper to call the other tools or don't care?
[14:51] <smoser> rbasak, what do yo uthink ?
[14:51] <rbasak> cloud-fingerprint has subcommands grep, import and fix. fix just wraps grep and import. grep filters console output for the cloud-init fingerprint printout. import takes that on stdin, calls ssh-keyscan, ssh-keygen, verifies fingerprint etc.
[14:51] <smoser> i think single python would be nicer.
[14:51] <caribou> med_: k
[14:52] <rbasak> So I can do "euca-get-console-output i-123|cloud-fingerprint fix 1.2.3.4" and it just works
[14:52] <rbasak> I thought it was time to split that out as this is useful for both local and remote clouds
[14:52] <smoser> rbasak, right.
[14:52] <smoser> how are you doing networking with libvirt ?
[14:52] <smoser> is this on the system libvirt ig uess ? (qemu:///system)
[14:53] <rbasak> The tool currently hardcodes qemu:///system. That could be configurable.
[14:53] <smoser> and how do you create and delete images ? it requires sudo ?
[14:53] <rbasak> The domain xml is picked up from a template. I intend that to go in /etc and ~/.something so that the user could override.
[14:53] <smoser> hm.
[14:53] <rbasak> Using lxml to manipulate the template, changing disks and hostname only. Networking stays as-is.
[14:53] <smoser> k.
[14:54] <smoser> i just come to the realiszation that we poitentially have
[14:54] <rbasak> Using libvirt python binding to create and delete images. No sudo needed.
[14:54] <smoser> lcloud libvirt-kvm
[14:54] <smoser> lcloud kvm
[14:54] <smoser> lcloud libvirt-lxc
[14:54] <smoser> lcloud lxc
[14:54] <rbasak> I'm not sure about non-libvirt
[14:54] <smoser> (ie, with or without lxc in the middle).  apparently there are people that do not think libvirt-lxc is "the real lxc"
[14:54] <rbasak> We would then have to manage things about where the persistent information about images is kept, at which point we'll have reinvented libvirt
[14:55] <rbasak> for kvm
[14:55] <smoser> rbasak, i agree, but there are many annoyances on libvirt and kvm.
[14:55] <rbasak> For lxc, the lxc package already does this
[14:55] <rbasak> So that's OK
[14:55] <smoser> well, networking is different is the big thing.
[14:56] <rbasak> What networking can you not do with libvirt?
[14:56] <zul> jamespage:  ping http://paste.ubuntu.com/1416957/
[14:56] <smoser> rbasak, the biggest issue i have with libvirt is its insistence on root owning my images.
[14:56] <smoser> rbasak, just that the networking is different.
[14:56] <smoser> not insufficient, just idfferent.
[14:56] <jamespage> zul, in a bit - have my head in swift charms from adam_g ATM
[14:57] <rbasak> If you go through the API, then that is no longer an issue, since you can still create/delete etc without root. With a wrapper tool that understands the API, is it really needed for someone who just wants to use images?
[14:57] <zul> jamespage: okies its just more cloud-archive tooling
[14:57] <smoser> rbasak, how do you delete images through api ?
[14:58] <smoser> i guess i didn't realize you coudl do that.
[14:58] <rbasak> you can virsh vol-delete --pool default <image>
[14:58] <rbasak> There are definitely things you can't do via the API
[14:58] <smoser> if your'e telling me i dont have to be root, but just have qemu://system access, then that rocks.
[14:58] <rbasak> And the API is really annoying and tedious to use
[14:58] <rbasak> But with suitable well-designed wrappers that pain should go away
[14:58] <smoser> but what is '<image>' there?
[14:58] <smoser> a path to an image ?
[14:59] <rbasak> I'm not sure what will work. I use just the filename as it appears in /var/lib/libvirt/images
[14:59] <rbasak> The abstraction is that directory is a storage pool
[14:59] <smoser> right. but then how did you put the image there.
[14:59] <rbasak> virsh vol-upload
[14:59] <smoser> oh good night.
[14:59] <rbasak> (yes, it's a pain!)
[15:00] <smoser> i didnt realize that was there. wow.
[15:00] <smoser> i guess our end game would be that those full images would be synced/populated as root
[15:00] <smoser> and your instance images would be qcow deltas backed by those
[15:00] <rbasak> I do belong to libvirtd btw
[15:01] <smoser> and then you'd hvae to upload "full localds ISOs"
[15:01] <smoser> rbasak, well, generally, i really like what you're doing.
[15:01] <smoser> and i really want to have nice tools like this for kvm and lxc
[15:01] <smoser> as i want to use them for testing cloud-init :)
[15:01] <rbasak> I'd like to write juju providers for this too :)
[15:02] <smoser> rbasak, 'lc2' (local compute cloud)
[15:02] <smoser> ?
[15:03] <rbasak> I like it. Makes it sound like a product though!
[15:03] <smoser> at the moment, we're only repllicating *compute*
[15:03] <smoser> in the future, we'll have analogs for object store, block storage, networking
[15:03] <smoser> ...
[15:03] <smoser> :)
[15:03] <smoser> and then profit
[15:04] <rbasak> that sounds like quite an interesting stack of tools. I wonder what we could call the entire collection?
[15:05]  * smoser wonders how much VC money we can get based only on the IRC conversation above.
[15:06] <smoser> rbasak, one thing that i want to at least attempt to support is non-ubuntu
[15:06] <smoser> images
[15:06] <rbasak> So far nothing is specific to ubuntu, only cloud-init, I think
[15:06] <smoser> with, of course, cirros being the first set of non-ubuntu images.
[15:06] <smoser> so what all do you do in the guest ?
[15:06] <smoser> oh, and do you take user-data ?
[15:06] <smoser> it has to
[15:07] <rbasak> Oh, I am assuming that avahi-daemon exists and can be installed.
[15:07] <smoser> what does that get you?
[15:07] <rbasak> I'm creating user-data. Right now I don't have the ability to take anything supplied by the user
[15:07] <rbasak> avahi-daemon makes zeroconf work. I can ssh <hostname>.local and it finds it
[15:07] <rbasak> No messing with IPs then
[15:08] <rbasak> I've come to the conclusion that it's the cleanest way of finding ephemeral machines in local use
[15:08] <smoser> i think i'd rather mess with IPs if i can find a way to deal with it.
[15:08] <smoser> i really dont want to make assumptions about the guest.
[15:09] <smoser> other than possibly that it brings up eth0
[15:09] <rbasak> It's a standard, doesn't really depend on anything apart from IP networking, will work with IPv6 well, doesn't depend on the dnsmasq trick which always felt horrible to me
[15:09] <smoser> well, if you're doing libvirt, you can specify
[15:09] <smoser> right?
[15:09] <smoser> you can set the ip that it will get on dhcp based on its mac that you also specify
[15:09] <rbasak> Not sure of the details. I struggled the last time I tried. libvirt was a bit buggy with the network xml stuff then (oneiric maybe)
[15:10] <rbasak> Suddenly we need to have an IP allocation mechanism
[15:10] <rbasak> And know what it is and how to query it
[15:10] <rbasak> With zeroconf, it doesn't matter what the user's using. He could be bridged to his LAN and using his own dhcp server and I don't need to care - it'll still work
[15:11] <rbasak> I'm not saying that we shouldn't support an alternative, just that this is a very easy case that takes no effort to get started
[15:11] <smoser> all this in exchange for making grave assumptions on the guest
[15:11] <rbasak> And we could add options for other mechanisms in the future
[15:11] <smoser> and not supporting user-data
[15:11] <smoser> :)
[15:11] <rbasak> I'm doing three things in user-data currently
[15:11] <rbasak> I'm automatically grabbing your ~/.ssh/id_rsa.pub and sticking that in
[15:11] <rbasak> I'm adding avahi-daemon
[15:11] <rbasak> And I'm setting your hostname based on what you asked for it to be called
[15:12] <smoser> id_rsa.pub can be meta-data not user-data
[15:12] <smoser> same with hostname
[15:12] <rbasak> Oh, OK
[15:12] <smoser> so avahi-daemon is the sticking point.
[15:12] <smoser> but libvirt probably makes this work, honestly.
[15:12] <rbasak> Well user-data is YAML, so it's easy enough to have options that manipulate user-supplied user-data
[15:12] <smoser> rbasak, well, its not guaranteed to be yaml
[15:12] <rbasak> Oh
[15:12] <smoser> and it sucks to merge.
[15:13] <smoser> well, anyway.
[15:13] <smoser> tahat can be solvve.d
[15:13] <rbasak> Well how about making it yaml if you want zeroconf, or if you supply non-yaml then you don't get that or have to do it yourself
[15:14] <smoser> rbasak, well, the thing you really need to do is use multi-part input.
[15:14] <smoser> one part comes from you, one from the user.
[15:14] <smoser> if you're doing that.
[15:14] <smoser> the issue then is that multipart cloud-config does not actually merge well, but you could use boothooks for your portions.
[15:14] <rbasak> I'm not familiar with the details of that. How would two yaml parts be merged? What if I want avahi-daemon installed and the user wants hello installed?
[15:14] <rbasak> OK
[15:14] <smoser> but then you're makign assuptions on the guest again (ie, you'd have to know to use 'apt-get' rather than 'yum install')
[15:15] <rbasak> Yes
[15:15] <smoser> hm..
[15:15] <rbasak> At some point, assumptions on the guest allow us to make the tool more useful and do more automatically, which for a command line tool like this I think is important.
[15:15] <smoser> lets table this for now. i think with libvirt and a network of our own we can make this work.
[15:16] <smoser> but, i'm ok with your solutio nfor now.
[15:16] <rbasak> I favour making it simple by default, but possible to turn off some of the automation in return for removing assumptions on the guest, so then it'll still work with non-ubuntu guests
[15:56] <eagles0513875> hey guys what is the group name for samba is it samba or something else as I created a share directory in /srv/samba/SHARE NAME it has right permissions and everything but I cannot add it to the samba group
[16:16] <eagles0513875> hey guys what is the group name for samba is it samba or something else as I created a share directory in /srv/samba/SHARE NAME it has right permissions and everything but I cannot add it to the samba group i get chgrp: invalid group: `samba when i try to change the group
[16:22] <SpamapS> eagles0513875: samba runs as root
[16:22] <SpamapS> eagles0513875: because it needs to map uids <-> samba users
[16:22] <eagles0513875> :-/ ok what i was planning on doing is restricting users to shares for their particular department
[16:22] <eagles0513875> does that mean i need to add each user to the root group?
[16:26] <eagles0513875> SpamapS: im a bit lost as to now go about setting up my permissions in terms of users :-/
[16:27] <SpamapS> eagles0513875: well you need to either tell samba what to set the user to in the share config, or tell samba how to map users.
[16:28] <eagles0513875> how do i go about doing that as there is no mention https://help.ubuntu.com/12.04/serverguide/samba-fileserver.html about having to do that
[16:29] <eagles0513875> SpamapS: are you talking about what is mentioned here https://help.ubuntu.com/12.04/serverguide/samba-fileprint-security.html in terms of the groups?
[16:30] <SpamapS> eagles0513875: there is, its just not obvious.    security = user
[16:30] <eagles0513875> I have that set
[16:30] <SpamapS> sudo mkdir -p /srv/samba/share
[16:30] <SpamapS> sudo chown nobody.nogroup /srv/samba/share/
[16:30] <SpamapS> pretty clear there
[16:31] <alex88> hi guys, I'm starting an fcgi daemon this way
[16:31] <alex88> exec su -c "spawn-fcgi -s /tmp/imageserver.sock -u www-data -g www-data -n -- /usr/bin/multiwatch -f 10 /bin/imagefogimageserver" www-data
[16:31] <alex88> inside an upstart script
[16:31] <alex88> but sometimes the socket doesn't get created… what can be the issue?
[16:31] <eagles0513875> SpamapS: if i have /srv/samba/product and another share at /srv/samba/product/123 can i change the permissions of that directory to have the group 123 and users samba
[16:34] <SpamapS> alex88: that looks good to me. no errors reported
[16:36] <alex88> SpamapS, in fact usually it works fine, now the service was giving 502, i've checked and the socket wasn't there, but service service name status gave the proc id
[16:37] <alex88> I've restarted and the socket still wasn't there, restarting service instead make it working
[16:39] <eagles0513875> Tm_T: can i ask you a question really quickly if i have a share /srv/samba/A and i create another share in its subdirectory /srv/samba/A/B can i on subdirectory B modify the permissions and group and ownership. ownership to samba and then the group to a particular department group?
[16:40] <Tm_T> eagles0513875: why not
[16:40] <eagles0513875> right now I cannot change the parent directory its set to nobody nogroup
[16:41] <eagles0513875> should the parent directory A in my example be changeable as well?
[16:41] <Tm_T> eagles0513875: I assume you get an error message when you try change ownership
[16:41] <Tm_T> and yes why not
[16:41] <eagles0513875> yes i get an error on the parent directory
[16:42] <eagles0513875> be it the user or the group i get chown: invalid user: `samba'
[16:42] <eagles0513875> i checked the group file and the group samba exists
[16:43] <Tm_T> "getent group | grep samba" would tell you something
[16:43] <eagles0513875> sambashare:x:112:jaquilina oops its samba shre
[16:43] <eagles0513875> my bad
[16:44] <eagles0513875> Tm_T: epic fail on my part
[16:47] <Tm_T> eagles0513875: this is exactly what is worrying me
[16:47] <eagles0513875> i just figured it out
[16:48] <Tm_T> you get clear error message, and yet...
[16:50] <eagles0513875> i got it
[16:54] <jamespage> adam_g, see MP's for comments on swift charms.
[17:17] <jamespage> adam_g, the only thing I did not manage to do was actually verify that swift was working :-)
[17:26] <hallyn> jdstrand: still no comments from upstream on the patches from lp:~serge-hallyn/ubuntu/raring/libvirt/libvirt-hugepages .  Do you think it would help if you commented on those threads instead of doing the bzr merge review?
[17:27] <jdstrand> hallyn: I plan on commenting
[17:28] <jdstrand> hopefully today, otherwise tuesday (it's at the top of my todo list after some MIR reviews)
[17:28] <jdstrand> s/tuesday/by tuesday/
[17:32] <hallyn> jdstrand: on the libvir-list ?
[17:32] <jdstrand> yes
[17:33] <hallyn> jdstrand: awesome, thanks.  have a good weekend
[17:33] <jdstrand> thanks, you too
[17:34] <jdstrand> hallyn: do you have packages with those patches somewhere?
[17:35] <hallyn> jdstrand: no, but i can push them to a ppa if it helps
[17:35] <hallyn> (i mean yes, but compiled locally on an instance)
[17:35] <jdstrand> hallyn: a source package on chinstrap would be fine
[17:35] <hallyn> k
[17:36] <jdstrand> hallyn: I noticed that '[PATCH 1/1] add vnc unix sockets to apparmor policy' does not include a test case in tests/virt-aa-helper-test. could you add one?
[17:37] <hallyn> jdstrand: http://people.canonical.com/~serge/libvirt_1.0.0-0ubuntu3ppa1.dsc  and http://people.canonical.com/~serge/libvirt_1.0.0-0ubuntu3ppa1.debian.tar.gz
[17:38]  * hallyn goes to look at how those testcases are done
[17:38] <jdstrand> hallyn: can you also add one for '[libvirt] add security hook for permitting hugetlbfs access (v2)'? that one would just be doing the equivalent call to virt-aa-helper that ApparmorSetHugepages() is doing
[17:39] <hallyn> jdstrand: what is that meant to test exactly?
[17:39] <jdstrand> hallyn: that virt-aa-helper is doing what you meant for it to do
[17:39] <hallyn> ok
[17:42] <jdstrand> it exited with the correct exit code. also, if you pass '-d' to virt-aa-helper-test, then you will get debugging output to visually inspect
[17:42] <jdstrand> hallyn: ^
[17:45] <hallyn> jdstrand: ok that test program looks grok-able, thanks for that :)
[17:45] <jdstrand> np
[18:08] <phunyguy_work> Hey folks, I am not on Ubuntu-server, but I need to write a script for some proxy settings, and I need some advice.  I have a script that currently checks to see if my LAN is pingable, and if not, try to connect via VPN, and if I am at home, it will disconnect the VPN.  What I want to accomplish is have it change system-wide proxy settings when I am away from home and connected VPN to proxy my web traffic through
[18:08] <phunyguy_work>  home, (including aptitude), and when I am home, stop using the proxy.   Any ideas hwo to accomplish the proxy changes dynamically?
[18:09] <phunyguy_work> (the reason I ask here is due to command-line stuff)
[19:05] <adam_g> jamespage: re swift review, i'd rather spend time doing a proper rewrite of that charm than working on the current bugs. i think the underlying design of how swift-proxy + storage are deployed with the current charms is broken, and we can do better
[19:18] <halvors> I ran chmod 6755 /usr/lib/dovecot/deliver wich conflicts with the setuid. How to revert it?
[19:19] <sarnold> halvors: "conflicts with the setuid"? do you want that file to be setuid or setgid or neither when you're done?
[19:19] <halvors> I want it back to default...
[19:38] <jamespage> adam_g, ack
[20:05] <ankushsachdeva> my wifi does not connect to adhocs :(
[20:50] <raub> Trivial question: how do I partition a hard drive during the server installation using command line?
[20:51] <sarnold> raub: look for commands cfdisk or gparted first, then fdisk.
[20:54] <raub> sarnold: cfdisk, gparted, and fdisk are not in the built-in shell offeredn during install. And I forgot how to add a package at that point
[20:54] <raub> shame on me
[20:55] <sarnold> raub: try find / -name '*fdisk*' ... it may be there, just under a different name.
[20:55] <sarnold> err
[20:56] <sarnold> raub: under a different _path_ than you may expect
[20:57] <raub> sarnold: hate to say only fdisk thingie it reports is a package (.udeb).
[20:58] <sarnold> raub: darn.
[20:59] <raub> I am going to dl a new iso just in case
[21:08] <raub> Still no nice. Even with a single ext4 partition
[22:08] <raub> sarnold: it seems I am not the only one https://bugs.launchpad.net/ubuntu/+source/qemu-kvm/+bug/802081
[22:10] <sarnold> raub: aha. Perhaps you can revive that bug... it never got the information hallyn asked for..
[22:10] <xnox> raub: sarnold: I thought this was fixed in lvm2 and SRUed.
[22:11] <raub> sarnold: thing is that it only happens during install
[22:11]  * xnox just had a dejavu moment.
[22:11] <raub> xnox: had that happen to you before?
[22:12] <raub> sarnold: If I run the livecd, I can partition it to my heart's content
[22:12] <xnox> raub: well. I never had the qemu problem myself, but I remember uploading fixes to resolve an approx similar problem you are discussing here.
[22:12] <raub> But if I try to install it it will go boink
[22:12] <raub> xnox: sounds like they were not committed?
[22:15] <jamespage> adam_g, still around?
[22:19] <adam_g> jamespage: yes
[22:20] <jamespage> adam_g, with regards to the rewrite of the swift charms; how long do you think that might take?  I'm concerned about the quanlity of whats in the charm store
[22:21] <jamespage> seriously considering whether we should unpromulgate it
[22:23] <adam_g> jamespage: im working on it now, hope to have it up for review before EOD mon? i'd +1 that idea. i dont know how those canonical-specific changes got in there to begin with
[22:24] <jamespage> adam_g, me neither
[23:52] <Daviey> adam_g: updates published
[23:52] <Daviey> https://lists.ubuntu.com/archives/cloud-archive-changes/2012-December/thread.html
[23:52] <adam_g> Daviey: danke
[23:53] <Daviey> adam_g / zul: Just a reminder to upload with debuild -v, to have a fuller log