[00:35] <rocio> alguien habla español
[00:36] <twb> !es
[00:36] <twb> Solo ingles in esta canal, por favor
[00:58] <smoser> Daviey, it seems that i cannot.
[00:58] <smoser> # euca-modify-image-attribute --launch-permission --add foof ami-0000003; echo $?
[00:58] <smoser> ApiError: user or group not specified
[00:58] <smoser> 1
[00:58] <smoser> not a good error, because clearly a user or group *was* specified, but at least it exited non-zero
[01:06] <rocio> alguien????
[01:11] <twb> rocio: no.
[01:11] <rocio> que tal
[02:33] <twb> GRARGH
[02:34] <twb> http://paste.debian.net/131692/
[02:34] <twb> I am *REALLY* sick of ubuntu doing that
[02:35] <twb> I think it's because of that crap PATH setting moved from /etc/profile to pam_environment.so or something
[02:35] <twb> *because of that crap where
[02:36] <twb> Basically what happens is, if you ssh to root@<lucid host>, PATH is not set appropriately.
[03:05] <twb> OK, I have an airgapped network.  I need NTP to work on it.  I cannot afford a cesium hardware clock, and I am not allowed to deploy a hardware clock that uses GPS.
[03:05] <twb> Currently, the NTP server reports a stratum of 16, which I *think* is why the desktops are ignoring its time information.
[03:06] <twb> Can I simply tell NTP on the server to force its stratum to 3 ?
[03:22] <CluelessPerson> how do I setup my server to allow for VMware?
[03:23] <twb> CluelessPerson: define "allow vmware"
[03:23] <CluelessPerson> twb, I'm learning about stuff, I don't know much about it, and I'm googling it and stuff, wondering if someone could tell me what they think
[03:25] <twb> That is not a sufficient answer.
[03:25] <CluelessPerson> twb, What do I need to use it?
[03:25] <CluelessPerson> minimum specs?, etc
[03:25] <CluelessPerson> twb,  I have a headless ubuntu server.
[03:26] <twb> To use what?
[03:26] <CluelessPerson> twb,  How do I handle virtualization?
[03:26] <CluelessPerson> I know OF it, but I odn't know it
[03:26] <twb> Is your actual goal something like "I want to learn how to make VMs"
[03:28] <CluelessPerson> twb, I think so. :p
[03:29] <twb> In that case, don't use VMware
[03:30] <twb> The technology you should be using is KVM, which is built into Linux.  The userspace component is also called "KVM" (based on "QEMU"), which is a CLI program.  There is an optional abstraction layer called "libvirt" to help manage VMs, and a GUI for that called "virt-manager".
[03:32] <CluelessPerson> twb, So if I wanted to run ubuntu -netbook on my netbook, and remote ui to a virtual os running in kvm on the server, what do you think of that?
[03:32] <twb> The key hardware requirement for x86 and x86-64 systems is hardware VT
[03:32] <twb> 13:32 <dpkg> To confirm if your processor supports x86 virtualization extensions. run: egrep -q -e '^flags.*vmx' -e '^flags.*svm' /proc/cpuinfo && echo 'virtualization extensions available'.  Note that Intel VT (IVT) support may be switched off in firmware for some implementations.  See also <virtualization>.
[03:34] <CluelessPerson> twb,   It tells me "virtualization extensiosn available"
[03:35] <twb> Then you are fine
[03:35] <CluelessPerson> twb,  so you suggest installing kvm
[03:36] <twb> Yes.
[03:36] <twb> You should read the Ubuntu Server Guide corresponding to your release.
[03:37] <CluelessPerson> on it
[04:29] <Jasonn> anyone know how to make user-specific disk space limits without making partitions in a user friendly way?
[04:38] <qman__> Jasonn, quotas
[04:38] <Jasonn> qman__: And can I use that to create them too
[04:38] <Jasonn> and if so, what is the command to do that
[04:40] <AlecTaylor> hi
[04:40] <AlecTaylor> How do I stop the default VNC server of Ubuntu via CLI?
[04:40] <qman__> it requires extra packages, a quick google found this guide: http://www.debianadmin.com/implement-and-manage-disk-quotas-in-linux.html
[04:41] <qman__> AlecTaylor, ubuntu server does not have a VNC server by default
[04:42] <AlecTaylor> Mine does!
[04:42] <AlecTaylor> :\
[04:42] <Jasonn> AlecTaylor: killall *vnc*
[04:42] <Jasonn> qman__: thanks :)
[04:42] <qman__> ubuntu server does not have a GUI by default, either
[04:42] <qman__> a prerequisite to running VNC
[04:44] <AlecTaylor> Jasonn: No process found
[04:44] <qman__> the point of this being, either you installed it and should know which one you installed, or you are not running ubuntu server
[04:44] <Jasonn> AlecTaylor: sudo apt-get install htop && sudo htop
[04:44] <Jasonn> find it there
[04:44] <qman__> you may be able to determine it with something like 'dpkg -l | grep vnc'
[04:45] <AlecTaylor> Jasonn: htop? - Hmm, what does the h stand for?
[04:45]  * AlecTaylor has ntop and top installed
[04:45] <Jasonn> its another prog
[04:45] <AlecTaylor> Jasonn: htop? - Hmm, what does the h stand for?
[04:45] <Jasonn> imo, its the best one
[04:45] <Jasonn> its another prog
[04:45] <Jasonn> imo, its the best one
[04:45] <AlecTaylor> Hmm, what does the h stand for?
[04:45] <Jasonn> I dont know
[04:45] <AlecTaylor> k
[04:46] <Jasonn> qman__: And what would the command be to set a limit of 1GB on user a ?
[04:46] <Jasonn> user_a *
[04:48] <qman__> I don't know, see the man pages for the various quota programs
[04:48] <qman__> I'd start with quotaon
[04:48] <qman__> also, please disregard the webmin stuff, as webmin is not supported on ubuntu
[04:49] <qman__> I didn't notice that bit when I first linked
[04:49] <qman__> the rest applies
[04:49] <Jasonn> oh
[04:49] <Jasonn> alright
[04:49] <Jasonn> thanks :)
[04:49] <AlecTaylor> I'm controlling a VNC session through RDP (xrdp) remotely. How do I force a change in resolution to one not listed?
[04:50] <Jasonn> take out the host comp's monitorr
[04:50] <AlecTaylor> :P
[04:50] <qman__> you'd have to hack X up to include your desired video mode
[04:51] <qman__> but that's not really a question suited to this channel, since ubuntu server does not have a GUI by default
[04:51] <twb> AlecTaylor: iirc the h in htop is the author's first name initial
[04:51] <AlecTaylor> qman__: Using that logic, anyone with a non clean install of Ubuntu Server can't ask questions here?
[04:52] <AlecTaylor> twb: Ahh
[04:52] <twb> AlecTaylor: like, his name is Hagbard or whatever
[04:54] <Jasonn> qman__: ok, if you install this for me and make it work, I will let you use my dedicated server and give you like 50GB HDD space
[04:54] <qman__> AlecTaylor, not exactly, but X and GUIs in general are a very wide subject and are outside the scope of ubuntu server's typical, supported configurations
[04:54] <qman__> I didn't say it wasn't allowed, just not well suited
[04:55] <qman__> you'll have better luck elsewhere, in more X specific places
[04:59] <qman__> Jasonn, I appreciate the offer but I don't really need such a thing, and I unfortunately don't have time to help you in further detail right now
[04:59] <qman__> if you still need help saturday or sunday I may be able to
[04:59] <Jasonn> alright
[04:59] <Jasonn> thanks man :)
[05:07] <AlecTaylor> How do I upgrade from 11.04 to 11.10?
[05:08] <jmarsden> AlecTaylor: You wait for 11.10 to be released.
[05:08] <twb> jmarsden: say, do you run unbound?
[05:08] <jmarsden> twb: No, sorry.
[05:08] <twb> Oh well
[05:09] <AlecTaylor> jmarsden: So I can't upgrade to beta?
[05:09] <twb> AlecTaylor: if you have to ask, you shouldn't be running pre-releases
[05:09] <AlecTaylor> But I want this package, only available in oneric: http://packages.ubuntu.com/oneiric/request-tracker4
[05:09] <jmarsden> So backport it.
[05:09] <jmarsden> Migrating to a a beta OS for one package is totally outrageously overkill...
[05:10] <AlecTaylor> How to backport?
[05:10]  * AlecTaylor thought it might be easier to upgrade rather than backport
[05:10] <AlecTaylor> !upgrade
[05:10] <jmarsden> AlecTaylor: Download the package sources, rebuild the package under your current production OS, and install the resulting package.
[05:11] <AlecTaylor> Easier to upgrade methinks
[05:12] <jmarsden> You will note the advice in the first of your two links re betas is "skilled and confident users may wish to upgrade" .  if you were a skileld and confident user of Ubuntu Server, you would not be asking these questions.  Therefore, upgrading to a beta is, for you, a bad idea.
[05:13] <Jasonn> AlecTaylor: yes, for a basic user to upgrade to a highly-unstable beta OS, sounds like a good idea to me :D
[05:13] <jmarsden> There is another option: patience.  11.10 will be released in October.
[05:19] <AlecTaylor> Jasonn: It'll be released next month, so why not make upgrading easier for the WinAdmin who takes over?
[05:19] <Jasonn> why not wait a month and let him do it?
[05:24] <AlecTaylor> Jasonn: Then he'd call me back in. I'm just doing 1 project, I don't want callbacks!
[05:24] <Jasonn> lol
[05:25] <Jasonn> youd get paid more!
[05:25] <AlecTaylor> Reputation would go down
[05:40] <CluelessPerson> So I'm thinking
[05:40] <CluelessPerson> from inside this network, I could create my own private network with my router, server, and netbook, while my family resides on the outer shell.
[05:41] <CluelessPerson> my server could act as an anonymolizing proxy to tor for all connections on whatever port
[05:42] <CluelessPerson> and my server could run vms, for various functions, including dummy vms that upon access could lock down the private network.
[07:12] <CluelessPerson> for some reason I can't install java 6 on my ubuntu server
[07:13] <ikonia> define "can't"
[07:15] <ersi> CluelessPerson: Please detail what your troubles are. Not finding a suitable package? Getting complains from apt installing a specific package?
[07:15] <CluelessPerson> can't find the package to install
[07:15] <CluelessPerson> sudo apt-get install sun-java6-bin
[07:15] <CluelessPerson> fails, can't find it
[07:16] <ersi> "sun-java6-jre - Sun Java(TM) Runtime Environment (JRE) 6 (architecture independent files)"?
[07:16] <ersi> Oh. sun-java6-bin.
[07:16] <ersi> What specific error are you getting? That apt isn't finding it on the mirror?
[07:17] <CluelessPerson> ersi, sudo add-apt-repository ppa:ferramroberto/java
[07:17] <CluelessPerson> "command not found"
[07:17] <CluelessPerson> like, wtf.
[07:17] <ersi> Uh.
[07:18] <ersi> That package is available in the regular mirrors though
[07:18] <ersi> Atleast in 11.04.
[07:23] <AlecTaylor> This upgrade is taking forever. How long is it meant to take to upgrade from 11.04 to 11.10?
[07:24] <ikonia> depends on a lot of factors
[07:24] <AlecTaylor> Hmm, what were the specs...
[07:24] <ikonia> (have a look on another terminal what processes are up to)
[07:24] <ikonia> what specs ?
[07:24] <AlecTaylor> 2GB of RAM
[07:25] <AlecTaylor> 2 processors
[07:25] <AlecTaylor> So about how long? (estimate)
[07:25] <ikonia> you can't estimate
[07:25] <AlecTaylor> Currently installing ubufox
[07:25] <AlecTaylor> Is it in alphabetical install order?
[07:25] <ikonia> have a look in antoher terminal if any processes are hung
[07:25] <ikonia> or if anything's going on
[07:25] <AlecTaylor> Nothings hung, it's progressing
[07:26] <AlecTaylor> Just want to know how much longer I need to wait
[07:26] <ikonia> there we go then, so sit tight
[07:26] <AlecTaylor> Yay, up to "Setting up" stage, after 1 or 2 hours
[07:26] <AlecTaylor> :D
[07:27] <ersi> AlecTaylor: Packing up packages and putting them in their places can take a while.
[07:27] <AlecTaylor> What are they, union workers?
[07:28] <ersi> Without the breaks, and it works for free.
[07:28] <ersi> And doesn't mind worse conditions
[07:28] <ersi> So I guess.. No :P
[07:32] <AlecTaylor> hehe
[07:32]  * AlecTaylor can't wait any longer, time to watch some more anime!
[07:33] <ersi> Sounds like a good distraction :]
[07:42] <AlecTaylor> Hmm, PostgreSQL upgrading to 9.1 from 8 but 8 db server still running
[07:42] <AlecTaylor> strange
[08:02] <soren> AlecTaylor: It depends on how much stuff is installed on the system being upgraded.
[08:02] <soren> AlecTaylor: ...and how fast the machine is.
[08:02] <AlecTaylor> It's currently at the remove redundant apps stage
[08:02] <CluelessPerson> does anyone here play minecraft?
[08:03] <CluelessPerson> I need help with an init script
[08:03] <AlecTaylor> WOOT - Restart stage
[08:03] <CluelessPerson> I'm looking for an init script that uses ramdisk and backups
[08:14] <Daviey> adam_g_: are you alive?
[08:15] <SpamapS> Daviey: he may not be, but I am.. :-P
[08:15] <SpamapS> 3 nights in a row of severe onset insomnia brought on by late coffee ingestion .. :-P
[08:18] <Daviey> SpamapS: heh, good to see your dedication :)
[08:18] <SpamapS> Mostly just tweaking my demo toy so I'll look smart with Jane and Jorge on stage at ODS :)
[08:19] <Daviey> SpamapS: non-shared storage instance migration is working \o/
[08:19] <Daviey> which is perfect for rebooting compute nodes \o/
[08:19] <Daviey> SpamapS: Ah, so you'll get the blame when i goes wrong? :)
[08:20] <Daviey> You keep signing up to stuff that'll backfire >:)
[08:20] <SpamapS> No, I'm preparing to throw rbasak under the bus. New guys are always at fault.
[08:20] <Daviey> True.
[08:20] <rbasak> :-(
[08:20] <SpamapS> Thankfully JamesPage stepped right in the pile of euca after I stepped in it.. :)
[08:21] <SpamapS> rbasak: buck up.. its only a 5 minute demo.. I'm sure they'll just give you a warning for screwing it up. ;)
[08:21] <AlecTaylor> I've got no screens. I've tried apt-get remove ubuntu-desktop, apt-get remove xauth. I then rebooted and reinstalled them. I still can't startx due to no screens being found. What's next?
[08:21] <rbasak> Hmm, maybe it's a good thing I'm not going to ODS :)
[08:22] <SpamapS> "If only rbasak would have showed up, we could have had a good demo."
[08:22] <jamespage> SpamapS: I noticed your footprint just as I did :-)
[08:22] <rbasak> lol
[08:24] <rbasak> Daviey: qemu-nbd: Could not access '/dev/nbd15': No such file or directory
[08:24] <rbasak> Daviey: what was the fix for that?
[08:29] <jamespage> SpamapS, other than whats in PPA whats the current state of Cassandra packaging?
[08:29] <jamespage> or is that it?
[08:31] <Daviey> rbasak: NFI
[08:31]  * rbasak reboots
[08:31] <AlecTaylor> I can't ping google via its hostname. When I try via its IP Address I get "Packet filtered" errors
[08:39] <ersi> AlecTaylor: Well, what have you done? Just apt-get remove and apt-get upgrade a lot of stuff? Did you have Networkmanager taking care of your connections before? (You're converting from a desktop install to a server install?) Have you configured any interfaces in /etc/network/interfaces?
[08:40] <rbasak> Daviey: you were right, libvirt is hanging because it's trying to open the fifo
[08:45] <Daviey> Gah
[08:46] <Daviey> rbasak: does echo > fifo unblock it ?
[08:46] <rbasak> Daviey: yep. Well "cat > fifo" does.
[08:46] <rbasak> (so just the open, no write required)
[08:46] <rbasak> I'm just trying a workaround by having my logger hold it open for write :)
[08:47] <Daviey> rbasak: Sounds like as part of the fifo creation hunk, touching the fifo might be the solution?
[08:47] <Daviey> ah!
[08:47] <Daviey> :)
[08:49] <jamespage> Daviey: asterisk merge? hows that working for you?
[08:55] <Daviey> jamespage: Hmm
[08:56] <nigelb> Hi, I have a very curious problem.
[08:56] <nigelb> When I run a bash script, the $PATH is empty inside that script.
[08:56] <nigelb> Its normal wen I echo $PATH in my shell.
[08:57] <ersi> Is there anyone familiar with freeRadius here? Im trying to locate why my radiusd is checking the system/local passwd file, even though I've commented out 'unix' and have 'file' uncommented in sites-available/default
[08:58] <SpamapS> jamespage: Cassandra puts out their own packages
[08:58]  * SpamapS should sleep soon :p
[08:58] <SpamapS> jamespage: that is the primary reason I stopped maintaining the PPA.. nobody cared about my packages when upstream was always more up to date. :)
[09:00] <Daviey> SpamapS: sleep is for the weak.
[09:01] <nigelb> Daviey++
[09:01] <jamespage> SpamapS, I thought that was the case
[09:02] <jamespage> I was reviewing negronjl's cassandra charm
[09:02] <jamespage> which ATM pulls a .deb from elsewhere
[09:02] <jamespage> which felt a bit odd - I'll ping him later
[09:03] <SpamapS> We need to start thinking *very* hard about cryptographic guarantees on files that charms pull.
[09:03] <jamespage> yes
[09:03] <Daviey> We discussed using archive, PPA and upstream source in charms - but i don't think external debs was considered
[09:03] <jamespage> I agree
[09:03] <SpamapS> And on configurability so that they are repeatable.
[09:04] <jamespage> I did a branch of the jenkins charm yesterday which defaults to the distro, but has configuration to use upstream stable or trunk archives as well
[09:04] <SpamapS> I'd actually move that we have an apt plugin for charms to make use of which also builds a repository of packages on the bootstrap node so that updates happen when you ask for them, not just when you add-unit and oops new package, yay.
[09:05] <jamespage> SpamapS, that would be nice
[09:05] <SpamapS> its needed for juju as well
[09:05] <jamespage> true
[09:05] <SpamapS> this business ofpulling from ppa or distro or branch or wtf ... its not going to fly
[09:06] <Daviey> SpamapS: Well it's plausible that upstream doesn't allow rebuilding of debs, ie, hacked packages... So something like debmarshal would be a good win for that.
[09:06] <SpamapS> Should be relatively easy actually. Just stick an apt-get and add-apt-repository in the path which use this plugin...
[09:07] <SpamapS> Daviey: yeah, the cassandra upstream debs are nearly impossible to rebuild
[09:07] <SpamapS> I *did* solve that.
[09:07] <SpamapS> but then I started trying to get rid of the embedded jars
[09:07] <Daviey> The key needs to be exposed aswell
[09:07] <SpamapS> and thats where I started wanting to throw cassandra off a moving high speed train
[09:07] <Daviey> heh
[09:07] <rbasak> OK now I'm hitting an apparmor problem: [Errno 6] No such device or address: '/home/ubuntu/nova/nova/instances/instance-0000000a/console.fifo'
[09:07] <rbasak> I've tried aa-complain on libvirtd
[09:08] <SpamapS> Daviey: right, we can think it through at implementation time. The idea is to be able to know that add-unit will actually add the same unit... not one with updates.
[09:08] <rbasak> That didn't work but I didn't really expect it to - how is it that apparmor is being applied to nova? Is libvirt doing it?
[09:08] <SpamapS> Anyway
[09:08] <twb> rbasak: apparmor has per-VM profiles instantiated from a libvirt template
[09:08] <Daviey> rbasak: You can reproduce it with apparmour teardown?
[09:08] <twb> rbasak: maybe you need to set the specific VM profile to complain?
[09:08] <SpamapS> chatting about work has made me want to sleep
[09:08] <Daviey> apparmor*
[09:08] <SpamapS> \o/
[09:09] <Daviey> SpamapS: hah, nn
[09:09]  * rbasak doesn't know that much about apparmor
[09:09] <SpamapS> goodnight, especially to my long lost english cousins ... :)
[09:09] <rbasak> Can I just disable the whole thing for testing?
[09:09] <rbasak> twb: how do I do that?
[09:10] <rbasak> Daviey: I don't follow, what's apparmor teardown?
[09:10] <Daviey> rbasak: as in, you get the same error with apparmor not part of it.
[09:10] <twb> rbasak: nfi
[09:10] <Daviey> rbasak: sudo /etc/init.d/apparmor teardown
[09:11] <twb> teardown will basically turn off apparmor AIUI
[09:11] <rbasak> Ah that is what I was after, didn't know how to achieve it
[09:11] <rbasak> No, same error :-/
[09:12] <rbasak> But doesn't libvirt dynamically create the profiles - will apparmor teardown work with that?
[09:12] <Daviey> It *should* do.
[09:12] <twb> Perhaps a better question is how to ask apparmor what profiles are currently enforced
[09:12] <twb> Anyway what makes you think that's an apparmor problem?
[09:13] <Daviey> if you stop + start libvirt you know for usre.
[09:13] <Daviey> (restart doesn't have the same behaviour as stop/start btw)
[09:13] <twb> They usually take the form of a dmesg/syslog audit() complaint
[09:13] <twb> Eh, I rewrote my libvirtd job so that containers will not be hard-killed if the host OS is shut down
[09:14] <Daviey> twb: Interesting, care to share the diff?
[09:14] <twb> http://paste.debian.net/131730/
[09:15] <rbasak> I thought it was an apparmor problem as occasionally I saw apparmor complaints in the nova-compute trace
[09:15] <twb> TBH I don't understand how ubuntu/libvirt could even THINK of shipping a system that just hard-kills VMs
[09:15] <twb> Daviey: note that this means you can't reconfigure libvirtd without rebooting all VMs
[09:16] <twb> kirkland: ^^ FYI, paste might interest you
[09:16] <rbasak> OK it's not apparmor at all
[09:16]  * rbasak looks sheepish
[09:19] <koolhead17> hi a
[09:19] <koolhead17> *all
[09:21] <spajderix> hi
[09:33] <Daviey> twb: nice!
[09:34] <Daviey> twb: you don't fancy raising a bug with that attached do you?
[09:34] <twb> ICBF, sorry
[09:37] <Daviey> :(
[09:45] <soren> twb: I don't see how that is functionally different from what we have in Oneiric now? Both first attempt a shutdown, wait for a while, then kill them hard if they're not yet stopped.
[09:46] <twb> soren: I only look at LTS, maybe you already have it in non-LTS versions
[09:46] <soren> twb: Ah.
[09:46] <twb> If so, sorry for the noise
[09:46] <soren> twb: Yeah, in Lucid it doesn't do anything at all, right?
[09:46] <soren> twb: So they just die.
[09:47] <soren> bug 350936
[09:47] <twb> Right, they just run until the old sysvinit shutdown code does a killall -9
[09:48] <soren> twb: Looks like it was pushed to Lucid as well.
[10:15] <HellSpawn1> добрый день
[10:15] <HellSpawn1> с ubuntu server на hyper-v работал кто-то ?
[10:16] <Myrtti> !ru | HellSpawn1
[10:16] <Myrtti> HellSpawn1: alternatively you could try English
[10:17] <HellSpawn1> Hi! I'm try use ubuntu server 10.04 x86_64 with Hyper-V. Have a problem with a large HDD
[10:20] <HellSpawn1> after instalation, i'm use hv_ modules for emulated hardware. And in my system appear second storage. And i have 2 HDD: first - sda and size 137Gb and second hda - full size.
[10:23] <HellSpawn1> after mounting system via hda drive, i'm receiving errors about write buffer at disk sda
[10:24] <Retrik> anybody know how stop this error: segmentation fault, i use ubuntu server 64 bit 10.04
[10:26] <rbasak> Daviey: was it lp:~hudson-openstack/nova/milestone-proposed you pulled from last time?
[10:31] <Daviey> rbasak: yes
[10:32] <twb> soren: in backports, or updates?
[10:32] <twb> soren: I haven't seen any dpkg merge offer with the latter enabled...
[10:32] <Daviey> rbasak: BTW, nova trunk has now moved to git.. so once you have a patch, we'll try and get it lnded in trunk before cherrypicking it back to our packages
[10:32] <rbasak> OK
[10:34] <HellSpawn1> some one using Ubuntu Server on Hyper-V ?
[10:35] <rbasak> HellSpawn1: I think AlecTaylor was yesterday
[10:35] <soren> twb: /me checks
[10:36] <soren> twb: It's in proposed.
[10:36] <soren> twb: Not sure why it hasn't move on from there.
[10:37] <twb> inertia
[10:38] <soren> twb: I have a guess, but I'm verifying.
[10:39] <twb> Don't trouble yourself on my account
[10:39] <soren> Too late.
[10:39]  * twb recoils in horror
[11:12] <Guest43860> Q). i just did a tasksel install lamp-sever but am unsure why i cant get .htaccess files working ? any clues?
[11:13] <Guest43860> im working on the default site at the mo but will need other vhost to include there own .htaccess file as well as directory level access control based on ip.
[11:21] <jamespage> \o/
[11:27] <Dulcin> Hmm, for some reason my system fails to fetch packages with apt-get update
[11:27] <Dulcin> I get a lot of 404 errors, I tried changing my sources list
[11:28] <Dulcin> but the same problem, any idea waht it could be?
[12:02] <rbasak> OK now I really am getting an apparmor problem: (nova.exception): TRACE: libvirtError: internal error cannot load AppArmor profile 'libvirt-984032ec-630f-c8e1-36ce-8d7b6efd1266'
[12:02] <rbasak> I can't see a file with that name in /etc/apparmor.d/libvirt, but similar ones from successful previous instances refer to a console.log file
[12:03] <rbasak> which I've now changed to console.fifo and it's now a fifo rather than a plain file
[12:04] <rbasak> Next I'm going to have to pull nova out of the equation and try getting libvirt to start instances with fifos manually unless anyone has any ideas?
[12:05] <rbasak> soren: ^^, re https://bugs.launchpad.net/nova/+bug/832507
[12:06] <soren> Oh.
[12:06] <soren> jdstrand: One for you ^
[12:07] <soren> jdstrand: The filename referencing a fifo in the libvirt xml is really a basename for the actual fifos.
[12:07] <soren> jdstrand: So, if you specify a fifo named "foo", the actual fifos will be named "foo.in" and "foo.out".
[12:07] <soren> jdstrand: I guess the apparmor magic will need to take that into account.
[12:07] <soren> jdstrand: (the logic that defines that lives in kvm, not libvirt)
[12:08] <soren> rbasak: jdstrand is your man when it comes to apparmour and libvirt.
[12:08] <rbasak> soren: I've not actually told libvirt that it's a fifo
[12:09] <soren> rbasak: ?!?
[12:09] <soren> rbasak: So how is kvm supposed to know?
[12:09] <rbasak> well, this is what I wondered when we started this conversation a couple of weeks ago!
[12:09] <rbasak> that's why I said it sounded dodgy!
[12:10] <ersi> Hm. I'm trying to bring up a VLAN Interface on my Ubuntu 11.04 with a Realtek RTL8111/8168B NIC and I'm getting "SIOCSIFADDR: No such device" when I try to bring up eth.100 - what gives? Here's my /etc/network/interfaces - http://dpaste.org/dx8MH/
[12:10] <rbasak> I suppose I could change it to do that, but I will have to sort out a writer for the other direction that's not used
[12:11] <Daviey> erm, why does kvm care it's a fifo?
[12:11] <Daviey> soren / rbasak ^
[12:12] <rbasak> imho, fifo semantics are different (eg. blocking, EPIPE, SIGPIPE) so it stops random breakage like the libvirtd hanging problem we had earlier
[12:13] <patdk-wk> ersi, did you install the vlan package?
[12:13] <ersi> patdk-wk: No. Lol.
[12:13] <soren> Daviey: So that it can do the right thing. I suppose it's possible that it'll work without telling it that it's a fifo (the special care you need to take with fifo's doesn't hurt if applied to other files), but I think it's probably safer to let kvm know what's going on.
[12:13] <ersi> patdk-wk: Doing that now, I'll also sit in the shame corner for a min :]
[12:14] <patdk-wk> don't worry, I use vlans all the time, and last install had the same issue
[12:14] <soren> rbasak: Just ignore the .in fifo.
[12:14] <patdk-wk> took a few min to figure that out myself :)
[12:14] <ersi> patdk-wk: Dunno why I assumed I had all the prerequisites :D
[12:14] <ersi> Might be cause I had fiddled with networking on that machine earlier, setting up bridging and shiz
[12:15] <ersi> Uhh, weird. Why is it still routing traffic for 10.100.0.0/16 through my default gateway on another device :E
[12:15] <Daviey> soren: it worked fine with my mknod from my legacy branch, i can't see why kvm would care what is under it.
[12:16] <soren> Daviey: Shall I repeat myself or can you read it again on your own? :)
[12:16] <Daviey> soren: either way.
[12:17] <soren> Daviey: Ok. Are you aware of the special care one needs to take to work with fifo's?
[12:17] <rbasak> soren: so http://libvirt.org/formatdomain.html#elementsCharPipe ?
[12:17] <soren> Daviey: Specifically fifo's where "the other end" might go missing and you don't want to let that take you down with it?
[12:17] <Daviey> soren: you'll notice i didn't ask a question, but made a statement agreeing with your supposition.
[12:17] <Daviey> soren: I always put on eye protection.
[12:17] <jdstrand> rbasak, soren: re .in and .out> this relates to 832507? if so, could someone connect the dots with xml to reproduce and I'll write a fix
[12:18] <soren> rbasak: Yes.
[12:18] <Daviey> soren: Do i need to provide logs of this same conversation from 2 weeks ago?
[12:18] <soren> Daviey: Sorry, I don't see your agreeing with me. I do see you, once more, question its necessity? What am I missing?
[12:19] <koolhead17> hey soren
[12:19] <soren> koolhead17: o/
[12:19] <Daviey> soren: I didn't ask a question.
[12:19] <rbasak> soren: just to make sure I understand, if I use type="pipe", I still need to append .in and .out? And by ignore are you saying that I need to open for write (so it won't block the open for read at the qemu end), or just do nothing?
[12:20] <soren> Daviey: It's not my fault you failed to punctuate correctly.
[12:20] <koolhead17> soren: got keystone working finally. :D
[12:20] <soren> Daviey: "i can't see why kvm would care what is under it" is a question in my book regardless of the lack of question mark at the end.
[12:22] <rbasak> jdstrand: yes, it's for 832507, but atm I'm not telling libvirt that it's a fifo, and that's what's not working. I'll try the type="pipe" as soren says and see if that works.
[12:22] <jdstrand> it'll be interesting to see if it does
[12:22] <jdstrand> (from an apparmor driver perspective)
[12:23] <soren> Daviey: Anyways.. You say you can't see why it would care. It would care, because you need to do someting special in order to properly handle it when the other end of a pipe dies (or closes the fd).
[12:23] <jdstrand> I was not aware of the .in and .out parts. incidentally, if you are seeing apparmor denials, they will be in kern.log (of audit.log if using auditd)
[12:23] <soren> Daviey: We can hope that kvm applies this same care to every fd, regardless of whether it *knows* that this special care is needed..
[12:24] <soren> Daviey: Or we can tell it that it needs to do it and be sure.
[12:24] <soren> Daviey: Seems like an easy decision in my book. I don't understand why you're so opposed to telling it that it's talking to a pipe? (question mark)
[12:24] <rbasak> jdstrand: right now I'm just seeing libvirtError: internal error cannot load AppArmor profile 'libvirt-984032ec-630f-c8e1-36ce-8d7b6efd1266' which doesn't make sense to me
[12:25] <rbasak> soren: this is exactly what I thought I was raising a week ago, and I went away from the conversation thinking that you didn't think it was an issue
[12:25] <Daviey> soren: ISTR it was me that raised reservations with the fifo concerns? :)
[12:25] <jdstrand> that can happen if things die prematrely. the logging in the security drivers leaves something to be desired
[12:26] <ersi> patdk-wk: Would you mind looking at this route table ( http://dpaste.org/UerNe/ ) and helping me understand why I can't reach 10.100.0.0/16 over eth0? Do I need a default gateway?
[12:26] <rbasak> jdstrand: aiui, all I'm doing different that is triggering that is that I've renamed console.log -> console.fifo (so the xml has changed accordingly) and that the file is now a fifo instead of a plain file
[12:26] <soren> rbasak: What I remember was you wanting to take care of the fifo specificities outside of kvm, because you didn't seem to trust it.
[12:26] <ersi> patdk-wk: I get routed via my default gateway over eth1 for some reason >_>
[12:26] <soren> Daviey: So why are you going the opposite way now?
[12:26] <soren> Is this candid camera?
[12:26] <jdstrand> rbasak: reading backscroll-- you are doing this currently with apparmor disabled, for testing purposes?
[12:27] <rbasak> jdstrand: no I rebuilt the instance since then. do you want me to try that?
[12:27] <Daviey> soren: Expect a cream pie in the face, shortly.
[12:27] <jdstrand> rbasak: no, not particularly. I just wanted to verify that if you tried to disable the driver, you did it correctly
[12:28] <soren> ersi: Which IP are you trying to reach?
[12:28] <rbasak> soren: I didn't trust it because I thought you were telling me to not tell it about the fifo, and I was worried that this would break something!
[12:28] <soren> ersi: ...because you're not routing 10.100.0.0/16 over eth0.
[12:29] <soren> ersi: only 10.100.0.0/18 (if my binary arithmetic is right)
[12:29] <ersi> soren: Uhm, I'm trying to reach 10.100.66.12, which is my VLAN network, which should get sent on eth0
[12:29] <rbasak> jdstrand: I don't understand why it would fall over at this point, but I might as well make the change to type="pipe" and worry about it if it persists then. I might have to get back to you :)
[12:29] <ersi> yeah, I meant /18.. Crazy ass Set top boxes want them netmasks
[12:30] <patdk-wk> that is like a /26 netmask, not /16 or /18
[12:30] <soren> ersi: ersi 10.100.66.12 is in 10.100.64.0/18
[12:30] <patdk-wk> oh, I can't read :)
[12:30] <patdk-wk> just got out of bed
[12:30] <soren> ersi: Not 10.100.0.0/18
[12:31] <ersi> Mmmh, d'uh. How silly of me
[12:33] <ersi> Wonder if I made it more silly now... :D Dosh darn this crazy network. > http://dpaste.org/X6PKE/
[12:33] <someuser> i'm trying to install glpi, but the install procedure says: can't create or write to <various names> folder. this is weird, since i'm the owner of these folders. any suggestions?
[12:33] <soren> rbasak, Daviey: I think it's fair to say that someone misunderstood someone. Now, can we please just fix this?
[12:33] <rbasak> soren: np
[12:34] <rbasak> soren: just to make sure I'm clear now though
 soren: just to make sure I understand, if I use type="pipe", I still need to append .in and .out? And by ignore are you saying that I need to open for write (so it won't block the open for read at the qemu end), or just do nothing?
[12:34] <ersi> soren or patdk-wk: mind having another look? http://dpaste.org/X6PKE/
[12:36] <soren> rbasak: Just ignore it. Forget it exists. <jedi mind trick>there's no .in pipe</jedi mind trick>
[12:37] <rbasak> soren: so I shouldn't even create the .in fifo?
[12:38] <rbasak> ah: "QEMU will not create these fifos, and requires them to be present."
[12:38] <rbasak> But if I create it and don't write to it, qemu will block on opening it for read
[12:39] <rbasak> But if I create and and don't _open_ it for writing, qemu will block on opening it for read
[12:39] <soren> What makes you say that?
[12:40] <rbasak> which part? manpage says that they must be present. fifo semantics are that opens block when opening for read if no writers, unless O_NONBLOCK is specified
[12:40] <soren> This part: 12:39 < rbasak> But if I create and and don't _open_ it for writing, qemu will block on opening it for read
[12:40] <soren> We established this weeks ago.
[12:40] <soren> QEmu *knows* this is a pipe, becuase we tell it's a pipe.
[12:40] <soren> So it doesn't do stupid things like hang when we don't write to it or die when we stop listening.
[12:41] <rbasak> OK, now I'm clear on what you're saying - thanks
[12:41] <soren> As I said back then: 10:35 < soren> Daviey: If everyone knows they're taling to a fifo, the semantics are well understood and pretty easy to deal with.
[12:41] <soren> Complete with typos and everything.
[12:43] <rbasak> Yeah a big "If" there. As I thought you were proposing that we *didn't* tell qemu that it was talking to a fifo, the sense in that sentence is completely reversed
[12:43]  * Daviey screams
[12:44] <soren> Ah.
[12:45] <soren> No, qemu lets you specify the type of file its chardevs are connected to. There are a bunch of them.
[12:45] <soren> There are plain files, pipe, serial, parallel, stdio, braille, udp, tcp, socket, tty.
[12:45] <soren> And more.
[12:46] <soren> Each with their own idiosyncracies.
[12:46] <twb> You're supposed to call them "features" until the users are at the gates with pitchforks and torches
[12:47] <Daviey> twb: Shouldn't you be asleep?
[12:47] <zul> morning
[12:47] <soren> It's amazing anyone lets me talk to other devs, let alone users.
[12:47] <twb> Daviey: 10PM
[12:47] <twb> Er, 11
[12:48] <Daviey> ah, ok.
[12:48] <zul> soren:  i agree :)
[12:48] <twb> I'm trying to replace crappy ol' Ubuntu armel with nice shiny Debian armhf, on my new netbook
[12:48] <Daviey> soren: It's amazing you are still breathing.
[12:48] <Daviey> twb: happy days
[12:49] <twb> I fucked up the bootloader so now I'm recompiling u-boot, and since the host is hosed, I'm doing so in a qemu-user-static chroot, from my old Atom netbook
[12:50] <spajderix> I have apache2 with mod_php enabled, hosting around 40 vhosts. I noticed something which I can't explain. I have entry in access.log with ip sending POST request to a page, and in this exact second, a file with random_numbers.php name get's created. There is no form or POST data processing what so ever on a given page. I have no idea how this happens, and more importantly how to prevent this
[12:50] <spajderix> from happening. Could you help me with figureing this out?
[12:50] <soren> Daviey: I may very well not be.
[12:51] <twb> Daviey: if I'm lucky qemu won't cause gcc-4.6 gnueabihf to have an internal error, like it does when compiling yacc output
[12:52] <soren> spajderix: Someone's found a bug in the page they're posting to that lets them hack you. Probably.
[12:52] <twb> \begin{bigot}The bug is: he enabled PHP.\end{bigot}
[12:52] <zul> php is still evil
[12:53] <rbasak> libvir: QEMU error : internal error Process exited while reading console log out
[12:53] <rbasak> put: chardev: opening backend "pipe" failed
[12:53] <rbasak> Am I missing something obvious?
[12:53] <rbasak> My xml is http://paste.ubuntu.com/695611/
[12:54] <spajderix> soren: That was my first thought, but i've checked the code. There is only a few lines of code getting items from db, and checking for pagination stuff. That's all. Is there a way to get POST data logged for requests in apache?
[12:54] <rbasak> Incidentally if I write to console.fifo.out then it goes into the ringbuffer OK :)
[12:54] <soren> rbasak: 2 pipes will be created called path.in and path.out. Data written to path.in will be received by the guest. Data written by the guest can be read from
[12:54] <soren>            path.out. QEMU will not create these fifos, and requires them to be present.
[12:55] <soren> rbasak: Maybe they both need to exist. Do they?
[12:55] <rbasak> soren: yes I've created them both
[12:55] <rbasak> prw-r----- 1 root root        0 2011-09-23 12:50 console.fifo.in
[12:55] <rbasak> prw-r----- 1 root root        0 2011-09-23 12:52 console.fifo.out
[12:55] <twb> rbasak: what's the kvm invocation look like?  (it's logged into /var/log/libvirtd-bin somwhere)
[12:55] <soren> rbasak: Oh.
[12:55] <soren> rbasak: Then need to not be owned by root.
[12:55] <soren> rbasak: ..but libvirt should be handling that :(
[12:55] <rbasak> soren: I'm running everything as root at the moment
[12:56] <soren> rbasak: Are you absolutely sure?
[12:56] <soren> rbasak: How did you do that?
[12:56] <twb> You can probably cheat in the short term by making those fifos 666
[12:56] <soren> twb: I don't love you anymore.
[12:56] <twb> *in the short term*
[12:56] <soren> mmmm
[12:56] <twb> I'm assuming this host is not connected to another
[12:56] <rbasak> I can try that
[12:57] <soren> I kind of doubt that you're actually running kvm as root.
[12:57] <twb> *I* do, because I'm a bastard :-)
[12:57] <rbasak> twb: qemu invocation: http://paste.ubuntu.com/695616/
[12:57]  * koolhead17 bows to Daviey 
[12:57] <koolhead17> hey zul
[12:57] <twb> goddamn ubuntu
[12:57] <twb> I don't want to log into openid just to get "download as text"
[12:58] <soren> twb: Yeah, that's pretty obnoxious.
[12:58] <twb> It probably won't work anyway in w3m
[12:58] <soren> rbasak: Well, unless you've changed it, kvm runs as some other user.
[12:58] <soren> libvirt-qemu or something.
[12:58] <rbasak> soren: OK, I'll try 666
[12:58] <twb> XSLT to the rescue
[12:59] <ersi> soren, patdk-wk: nevermind. I was using a IP-address that was in the wrong subnet/netmask. That's why "nothing" worked :)
[12:59] <twb> rbasak: why --no-kvm there?
[12:59] <rbasak> twb: no idea. I'm just using nova.sh
[13:00] <twb> rbasak: do you have hardware VT?
[13:00] <rbasak> twb: I don't know, but it does work without my patch
[13:00] <twb> 23:00 <dpkg> To confirm if your processor supports x86 virtualization extensions. run: egrep -q -e '^flags.*vmx' -e '^flags.*svm' /proc/cpuinfo && echo 'virtualization extensions available'.  Note that Intel VT (IVT) support may be switched off in firmware for some implementations.  See also <virtualization>.
[13:01] <rbasak> evidently I don't
[13:01] <rbasak> (it is virtualised once already)
[13:01] <twb> Oh
[13:02] <twb> I was about to say "who cares about the actual problem, first go buy some appropriate hardware"
[13:02] <rbasak> OK same error: libvir: QEMU error : internal error Process exited while reading console log out
[13:02] <rbasak> put: chardev: opening backend "pipe" failed
[13:02] <rbasak> FIFO is now 666
[13:02] <rbasak> Is there a component of qemu I'm missing or something?
[13:03] <twb> rbasak: can you just run kvm by hand, and if so does it get at least as far as that error?
[13:04] <rbasak> I tried running the same command line but it didn't work - I think because lots of setup like maybe pts etc that nova provides is missing
[13:04] <twb> yeah i had that problem with libvirt too
[13:04] <rbasak> I will need to cut the line down carefully to a version that will work to test just the fifo behaviour
[13:05] <twb> Well, you can test taht bit just using the fifo part plus a d-i kernel and initrd
[13:05] <rbasak> OK
[13:05] <rbasak> I just checked the docs carefully
[13:05] <rbasak> The manpage contradicts itself
[13:06] <rbasak> "On other hosts, 2 pipes will be created called path.in...QEMU will not create these fifos..."
[13:07] <soren> Check dmesg
[13:07] <soren> See if apparmour blocks it.
[13:07] <rbasak> Aha, yes it is
[13:07] <rbasak> Thanks soren, that just saved me the next half hour
[13:08] <twb> Good catch
[13:08] <rbasak> OK so this is where jdstrand comes in I think? The generated profiles have console.fifo only, not console.fifo.in and console.fifo.out
[13:10] <rbasak> jdstrand: example xml is here: http://paste.ubuntu.com/695611/
[13:10] <jdstrand> rbasak: and do you have apparmor denials?
[13:11] <rbasak> jdstrand: example generated profile: http://paste.ubuntu.com/695621/
[13:11] <rbasak> jdstrand: dmesg: http://paste.ubuntu.com/695622/
[13:11] <rbasak> jdstrand: they may not match exactly, if you need them to let me know and I'll do another test run
[13:11] <jdstrand> rbasak: can you attach all that to the bug?
[13:12] <rbasak> jdstrand: will do
[13:12] <rbasak> jdstrand: I'll tidy up and make everything consistent and attach my patch to the bug as well then. It's a good point to have a checkpoint anyway
[13:32]  * Daviey wonders if smoser will ever cut a new upstream version of cloud-init
[14:04] <hallyn> Daviey: he's on a 'linux 3.0' schedule
[14:05] <twb> kool-aid boy, huh
[14:12] <zul> if i create a nova group it shouldnt do anything bad should it?
[14:24] <laos> hey there guys - i was wondering if any of you would know if apache2 would start acting weird and ddns also after an internal ip change from 192.168.1.7 to 192.168.1.10
[14:24] <raubvogel> Anyone using k5start?
[14:24] <laos> i had a similar issue in the past and it went away after the router gave the original ip back to the server, but now my router is refusing to release the lease on the device that took the server's old IP
[14:47] <lynxman> morning everyone o/
[14:50] <Daviey> hey lynxman
[14:50] <Daviey> zul: Are you working on s/socat/netcat ?
[14:50] <lynxman> Daviey: hello sir
[14:50] <zul> Daviey: yeah
[14:50] <Daviey> zul: right now?
[14:51] <zul> Daviey: 10 minutes ago i moved on to the iscsi stuff
[14:51] <zul> why?
[14:53] <Daviey> zul: Just prepairing for the release meeting, wondered ETA
[14:54] <zul> Daviey: this afternoon
[14:54] <Daviey> cool!
[14:56] <zul> ill do the dh_python2 stuff while we are at it
[15:03] <smoser> new upstream version of cloud-init coming, dont worry daviey.
[15:04] <smoser> i was kind of hoping to have it not *terribly* broken for gholms' fedora work.
[15:10] <bcessa> hi, I've a problem, how can I adjust the SMPT port from 25 to something else to use it with postfix?
[15:14] <rbasak> Am I supposed to be able to submit multiple attachments in a bug comment? Or is a tar acceptable?
[15:19] <rbasak> jdstrand: I've updated bug 832507. Can I leave the apparmor fix to you please?
[15:19] <jdstrand> rbasak: yes, thanks
[15:20] <bcessa> got it
[15:33] <robbiew> Daviey: adam_g:  so it just occurred to me when we rollout official HowTo documentation for Orchestra, we should include any ARM specific gotchas (in terms of pandaboards)
[15:33] <robbiew> rbasak: zul: ^
[15:34] <zul> robbiew: just the cobbler bits afaik
[15:34] <robbiew> zul: that's what I thought
[15:34] <Daviey> Talking of which, i'm not sure the PXE fix has landed in uboot.
[15:35] <rbasak> I have followed http://pad.ubuntu.com/arm-server-netboot through for cobbler (directly, no Orchestra) and it worked without any issues
[15:35] <rbasak> Although I did some of the parts by hand so didn't test the scripts
[15:35] <Daviey> rocking.
[15:36] <Daviey> robbiew: Next week probably involves doc consolidation.. pushing together all the snippets into the main docs.
[15:40] <MiaFarrows> Hello there
[15:41] <MiaFarrows> I've just installed 11.04 and I don't know how to change the keyboard layout so it uses the Spanish(Spain) keyboard layout
[15:41] <RoAkSoAx> robbiew: I should have the orchestra/juju docs finished by the end of the day today then adam_g_ could add the openstack stuff
[15:41] <MiaFarrows> I've tried dpkg-reconfigure console-setup, but just lets me select the encoding, charset and font settings
[15:42] <RoAkSoAx> robbiew: i should aslo include ARM if desired
[15:42] <robbiew> Daviey: RoAkSoAx: ack
[15:43] <Daviey> zul: Is "List of missing bits pieces of libvirt for lxc (libvirt 0.9.1-1 merge is pending)" still required?
[15:43] <zul> nah i think we are good
[15:44] <Daviey> cool
[15:44] <Daviey> RoAkSoAx: Did you have a chance to "Validate desktop deployment through cobbler" ? (alt cd?)
[15:45] <MiaFarrows> Anyone can help me please?
[15:46] <RoAkSoAx> Daviey: I did once long ago, I;ll do it again today
[15:46] <Daviey> RoAkSoAx: awesome
[15:46] <MiaFarrows> As I've already said, how can I change the keyboard layout from Ubuntu Server 11.04's console?
[15:46] <Daviey> RoAkSoAx: If it works, can you update the blueprint?
[15:46] <RoAkSoAx> Daviey: sure
[15:47] <MiaFarrows> AAAAAAAAAAGGGGGGHHH Fuck you all! and Thanks for nothing idiots!
[15:47] <robbiew> MiaFarrows: are you using ubuntu server?
[15:48] <robbiew> or desktop?
[15:48] <MiaFarrows> Server, as I said like 3 times
[15:48] <oCean> MiaFarrows: better drop that attitude
[15:48] <MiaFarrows> im so fuckin desperate+
[15:48] <robbiew> MiaFarrows: calm down
[15:48] <oCean> MiaFarrows: control your language here, please
[15:48] <MiaFarrows> alright sorry, I've been in front of the console for 2 hours
[15:49] <robbiew> MiaFarrows: understand the frustration...give me a sec
[15:49] <MiaFarrows> I'm kind of irritated right now.
[15:49] <MiaFarrows> Sorry for the attitude, once again.
[15:49] <koolhead17> MiaFarrows: your not alone hitting your head against it
[15:50] <robbiew> koolhead17: is this a known bug?
[15:50]  * robbiew asks before diving into it
[15:50] <koolhead17> robbiew: no. i dont think so
[15:51] <robbiew> ok, thx
[15:51] <zul> Daviey: can you have a quick look at lp:~zulcss/nova/ubuntu-changes
[15:52] <koolhead17> MiaFarrows: so when you installed the server did it not asked you for keyboard layout?
[15:53] <MiaFarrows> nope
[15:53] <MiaFarrows> after installing the server I got the loving prompt
[15:53] <Daviey> zul: you know nova has moved to git now?
[15:53] <koolhead17> MiaFarrows: care to explain once more please
[15:53] <MiaFarrows> damn autocorrection lol, logging*
[15:53] <zul> Daviey: yeah i know
[15:53] <zul> this is packaging changes for ubuntu
[15:54] <Daviey> oh
[15:54] <Daviey> zul: stacked against nova?
[15:54] <zul> yes the diablo packaging branch
[15:54] <robbiew> MiaFarrows: sudo dpkg-reconfigure console-setup
[15:55] <just-a-visitor> (I'd suggest to check /etc/default/keyboard and set XKBLAYOUT="your_2_letters_language_code")
[15:56] <koolhead17> echo "setxkbmap us" >> ~/.bashrc
[15:56] <koolhead17> :)
[15:56] <zul> Daviey: like we talked about
[15:56] <Daviey> zul: rev 197 to head?
[15:56] <MiaFarrows> I downloaded the 11.04 64 bits .iso from ubuntu.com, installed it on vmware and as soon as it finished installing the server, vmware-tools were installed, and after that a login prompt appeared on the server's console, I can log in, but the keyboard layout is set to english I guess, so I did dpkg-reconfigure console-setup but there is no option to change the keyboard language, just let's you change the encoding that the
[15:56] <MiaFarrows> console uses, the font size/face and the charset.
[15:57] <zul> Daviey: the latest and greatest
[15:57] <MiaFarrows> So whenever I want to type '-' I get '/' instead
[15:57] <MiaFarrows> the ñ is a ; and so on
[15:57] <Daviey> MiaFarrows: vmware does some magic to make the install less (more) painful.
[15:58]  * koolhead17 have never used VMware
[15:58] <MiaFarrows> Davley, I've installed it manually, no easy install
[15:58] <Monotoko> hey guys...I'm trying to create a global LAN...in that I want people to be able to connect to my server (authenticated, there will be about 100 users) and then go where they need to. I want to add another TLD inside the network, for example www.main.dev takes a user to the main page...could OpenVPN do this?
[15:58] <Daviey> MiaFarrows: you probably want to change, /etc/default/locale
[16:02] <hallyn> Daviey: should bug 842845 be assigned to someone (zul :)?  If not, I might feel like I have to go spend time on it
[16:02] <zul> i wasnt able to reproduce it actually
[16:02] <Daviey> hallyn: sounds wonderful, zul you are pretty loaded at the moment! :)
[16:03] <zul> but there is some funkyness when you are shutting down the instances
[16:03] <Daviey> zul: Have you tested the deprecated auth stuff?
[16:03] <zul> Daviey: no...i was waiting to get the package out of the pending state :)
[16:04] <Daviey> zul: it has been for a few hours.
[16:04] <zul> eh? even with iscsitarget and socat?
[16:04] <Daviey> zul: those are not build deps.. :)
[16:06] <zul> Daviey: gah...interesting well it wont be using them in ubuntu2
[16:07] <Daviey> zul: Hmm d/changelog has duplication?
[16:07] <koolhead17> zul: hope your not doing anything to the keystone.
[16:07] <zul> Daviey: hold on lemme check
[16:07] <hallyn> Daviey: zul: I should take the bug?
[16:07] <zul> hallyn: please
[16:07] <Daviey> hallyn: +1
[16:08] <hallyn> kewl, thx
[16:08] <Daviey> zul: the netcat replacement - is it tested?
[16:08] <hallyn> I've got some kernel patches to fix up first though.  Kick me if you want me to get to it today.
[16:08] <zul> no i dont think anyone tested socat either
[16:09] <zul> Daviey: pushed
[16:09] <Daviey> zul: regarding the patch, is it worth linking to the gerrit review page?
[16:09] <Daviey> for Forwarded:
[16:09] <zul> Daviey: good idea
[16:10] <Daviey> zul: Yeah, i think you are right.. if we can just get socat off the list, it helps.. we should make a point of testing it this week
[16:10] <zul> agreeed
[16:11] <Daviey> zul: Do you want to add one more patch?  Seems silly for me to do it seperately
[16:11] <zul> Daviey: sure
[16:12] <Daviey> zul: https://code.launchpad.net/~nttdata/nova/850602/+merge/75480
[16:12] <Daviey> the test didn't cleanly apply.. it might do now with a new snapshot
[16:14] <zul> Daviey: ok ill pick it up after I eat
[16:14] <zul> but other than that you are ok with it?
[16:16] <Daviey> zul: i think so
[16:17] <Daviey> zul: changelog exceeds 80 chars width
[16:18] <Daviey> zul: In retospect, as it seems to build on Lucid - perhaps we should keep that change?
[16:21] <zul> the dh_python2 change?
[16:21] <Daviey> yeah
[16:22] <zul> nah
[16:37] <Daviey> adam_g_: are you alive?
[16:38]  * SpamapS just had MAJOR de ja vu
[16:39] <lynxman> Daviey: he was pretty tired yesterday after the talk
[16:39] <Daviey> SpamapS: you've only been gone 3 mins!
[16:39] <Daviey> you didn't sleep
[16:39] <Daviey> lynxman: thanks
[16:40] <SpamapS> Daviey: son of a!
[16:41] <SpamapS> jdstrand: some time last week you pinged me about PHP 5.3.8 .. http://news.php.net/php.internals/55598 ... I raised the issue that we had been dragging our feet on shipping it and it seems Rasmus agrees that they need to revert the change, so 5.3.9 should be good.. but too late for 11.10. :-/
[16:41] <SpamapS> Or that might have been 2 weeks ago
[16:41] <SpamapS> time is a blur :p
[16:43] <jdstrand> SpamapS: thanks
[16:43] <jdstrand> sbeattie: ^ for your updates
[16:43] <jdstrand> SpamapS: it may have even been 3 :P
[16:43] <koolhead17> soren:
[16:43] <jdstrand> SpamapS: thanks for following up though
[16:53] <adam_g> Daviey: here
[16:53] <lynxman> hey koolhead17 o/
[16:54] <robbiew> adam_g: how'd the juju talk go?
[16:54] <robbiew> you survive?
[16:54] <adam_g> robbiew: eh
[16:54] <koolhead17> hey lynxman
[16:55] <adam_g> Adri2000: :)
[16:55] <adam_g> er
[16:55] <koolhead17> am much relaxed today
[16:55] <Daviey> adam_g: Cool!  I got live migration working..
[16:55] <Daviey> the charm could do with some changes to enable it by default?
[16:55] <adam_g> Daviey: yeah, it would be easy. what was needed?
[16:56] <Daviey> adam_g: generating and sharing a psk.. and a /etc/default/libvirt* change
[16:56] <Daviey> I'm currently using it with noauth, rather than a psk
[16:57] <adam_g> Daviey: /etc/default/libvirt* changes on the compute nodes i assume, but psk shared between who?
[16:59] <Daviey> adam_g: the compute nodes need to talk directly with eachother.. so the psk between them
[16:59] <Daviey> /etc/default/libvirt to add -l (listen mode)
[17:07] <adam_g> Daviey: ah yeah, should be doable in the charm. what about migration flags in nova.conf? did you figure out exactly which ones were required?
[17:10] <zul> Daviey: just doing a test build
[17:15] <Daviey> adam_g: Yeah... don't touch the flags :)
[17:15] <Daviey> red herring :P
[17:15] <Daviey> leave it as default
[17:16] <SpamapS> is the nova in the archive the final diablo release?
[17:16] <zul> yes
[17:16] <SpamapS> zul: <borat>HIGHFIVE</borat>
[17:16] <zul> hehe
[17:17] <adam_g> Daviey: dont touch as in dont set them at all?
[17:21] <Daviey> adam_g: yeah, vanilla settings
[17:21] <Daviey> smoser: hmm, why do you hate dep3 so much?
[17:21] <smoser> i didn't realize i did
[17:21] <smoser> oh. in cloud-init ?
[17:22] <smoser> i think its cleaner and more obvious when you're just cherry picking
[17:22] <smoser> they're straight forward
[17:22] <Daviey> smoser: It does feel pretty damn ugly to include a LICENCE as a pach to 'upstream'
[17:22] <Daviey> yeah
[17:23] <smoser> i use some text in debian/README.source to easily cherry-pick
[17:23] <smoser> and turning that into a DEP-3 formated diff would be a PITA and useless.
[17:23] <Daviey> smoser: why not just cut a new release?
[17:23] <smoser> because i got close enough to what i had hoped was going to be in oneiric that i just said "i'll cut a release with what is in oneiric"
[17:23] <smoser> :)
[17:24] <Daviey> smoser: Considering you are upstream, and you are including all patches to upstream (right?), this aproach seems to be doing it wrong :)
[17:25] <smoser> cutting binaries every 3rd day is wrong
[17:25] <smoser> uploading binary snapshots to ubuntu as new releases is wrong
[17:26] <smoser> both of those make it harder to see what is changing than this.
[17:26] <Daviey> smoser: well i disagree, but i won't block on it.
[17:27] <smoser> really?
[17:27] <smoser> how would you review the changes in a new snapshot?
[17:27] <smoser> here there is a single new patch
[17:28] <smoser> and launchpad is less filled with meaningless binary original source tarballs
[17:29] <Daviey> smoser: Launchpad generates diffs
[17:30] <Daviey> as in debdiff's
[17:30] <Daviey> http://launchpadlibrarian.net/80735555/cloud-init_0.6.1-0ubuntu19_0.6.1-0ubuntu20.diff.gz is what i reviewed.
[17:30] <smoser> ok, so in this rare case that you're reviewing something post upload
[17:31] <Daviey> smoser: if it was a bzr merge proposal it would STILL be easier to review if it wasn't flat patches.
[17:32] <smoser> completely disagree
[17:32] <smoser> if it was a bzr merge proposal, and i was making changes to the files (as opposed to patching them) and using quitl 3.0, you'd get double diffs
[17:32] <Daviey> smoser: how?
[17:33] <Daviey> smoser: exactly.. by using flat patches i'd get double diffs.
[17:33] <Daviey> if you cut a new release, i'd get single diffs
[17:33] <smoser> you dont have doulbe diffs now. you have a diff that adds a patch
[17:33] <smoser> if i were using quilt 3.0, you'd get a diff that added a patch, and diffs to the files
[17:33] <smoser> and you'd have to use 'filterdiff' or some other sillyness to tell bzr to do what it should do
[17:34] <smoser> anyway
[17:34] <Daviey> smoser: no.. i think you are failing to listen to what i am saying :)
[17:34] <Daviey> if you didn't patch, but cut a new release - what job does quilt have?
[17:37] <smoser> why would i cut a new release?
[17:37] <smoser> for 3 lines of changes?
[17:37] <smoser> what upstream woudl do that?
[17:37] <zul> alrighty...nova ubuntu2 coming your way
[17:41] <robbiew> RoAkSoAx: hey, so awhile back I created https://wiki.ubuntu.com/ServerTeam/Orchestra
[17:41] <robbiew> could you take ownership of that?
[17:42] <RoAkSoAx> robbiew: sure
[17:42] <RoAkSoAx> robbiew: I will jupdate it accordingly
[17:42] <robbiew> thanks!
[17:44] <koolhead17> hey robbiew RoAkSoAx
[17:44] <robbiew> koolhead17: what's up?
[17:45] <koolhead17> robbiew: notthing much. much relaxed today !! :D
[17:46] <koolhead17> RoAkSoAx: and am still stuck at same bloody custom partitioning in cobbler :(
[17:47] <RoAkSoAx> koolhead17: hehe how so
[17:48] <koolhead17> RoAkSoAx: http://www.tylerlesmann.com/2008/jul/06/fun-preseed/  This is what some wise man suggested me to go through, but was too occupied with dear keystone. :D
[17:52] <koolhead17> zul: just request. don`t spend time on this https://bugs.launchpad.net/keystone/+bug/855467 i think it has much to do with path.
[17:54] <koolhead17> i have keystone working now from github i will see if i can check this issue myself
[17:55]  * koolhead17 rushes home
[18:18] <adam_g> Daviey: details about the psk stuff around anywhere?
[18:22] <Daviey> adam_g: libvirt wiki
[18:22] <Daviey> one mo
[18:23] <Daviey> adam_g: http://libvirt.org/remote.html#Remote_certificates
[18:24] <adam_g> Daviey: thanks
[18:38] <Jasonn> hi, I have a problem when uploading files via FTP. IT seems to give me: 550 Permission Denied. When I try to upload to my home dir.,
[19:02] <SpamapS> Jasonn: your ftp service probably doesn't run with enough privileges to write to your home directory
[19:43] <soren> koolhead17: hm?
[19:44] <koolhead17> soren: was trying dashboard with keystone, was stuck somewhere then got https://github.com/cloudbuilders/devstack/blob/master/files/keystone_data.sh
[19:45] <koolhead17> so pinged you that time. am home so will try it 2morrow
[19:45] <koolhead17> :)
[19:47] <soren> alright.
[22:28] <Dulcin> is setting the /var/www mode to 755 secure enough?
[22:35] <stiv2k> ok wtffffff
[22:35] <stiv2k> i dont understand why this keeps happening
[22:36] <stiv2k> every morning at ~5:45am, my server gets a dhcp address from my router... even though it's configured static in /etc/network/interfaces and also configured static by wicd
[22:36] <stiv2k> and i have to unplug/replug the network cable to make it get a static address again
[22:36] <stiv2k> suggestions?
[22:47] <lynxman> smoser: ping
[22:48] <adam_g> lynxman: lol
[22:48] <lynxman> adam_g: hm? :)
[22:49] <adam_g> lynxman: cloud-init?
[22:49] <lynxman> adam_g: lol, yes
[22:49] <lynxman> adam_g: where are you
[22:49] <qman__> stiv2k, make sure it's configured in /etc/network/interfaces, then sudo ifdown, find and kill any DHCP processess still running, then sudo ifup
[22:49] <adam_g> lynxman: back row.
[22:49] <qman__> also, if you have networkmanager, remove it
[22:49] <lynxman> adam_g: I'm 2nd row
[22:51] <stiv2k> qman__: i dont have networkmanager but i use wicd
[22:51] <qman__> I don't know anything about wicd
[22:51] <stiv2k> holy fuck
[22:51] <stiv2k> i think the apocalypse is happening
[22:51] <qman__> but the above process will resolve any other rogue DHCP configurations
[22:52] <stiv2k> qman__: ill show you my interfaces file
[22:52] <stiv2k> qman__: http://fpaste.org/MFu1/
[22:52] <stiv2k> no ip info there, i let wicd get the static ip
[22:52] <qman__> well, you have to configure an address
[22:52] <stiv2k> wicd does it
[22:53] <qman__> like I said, I don't know anything about wicd
[22:53] <stiv2k> ok
[22:53] <stiv2k> well, i pretty much did everything you said EXCEPT i never issued ifdown and ifup commands
[22:53] <stiv2k> and
[22:53] <stiv2k> i wouldn't know which processes are dhcp processes
[22:53] <qman__> dhclient mostly
[22:53] <stiv2k> aha
[22:53] <qman__> it's very persistent that once it's run, it'll stick around and keep renewing it
[22:54] <stiv2k> thats probably where my problem is
[22:54] <qman__> unless it is killed off, it'll override your static configurations
[22:54] <stiv2k> root     32240  0.0  0.0   2548   792 ?        Ss   Sep20   0:02 dhclient3 -e IF_METRIC=100 -pf /var/run/dhclient.eth0.pid -lf /var/lib/dhcp3/dhclient.eth0.leases eth0
[22:54] <stiv2k> ^
[22:54] <qman__> the ifdown/up process does that
[22:54] <stiv2k> okay
[22:54] <qman__> so does a reboot
[22:54] <stiv2k> yeah never rebooted since installing
[22:54] <qman__> provided it's configured properly, at least
[22:54] <stiv2k> 38 days, 12h ago
[22:54] <stiv2k> ok this should resolve it then
[22:54] <stiv2k> im going to ifdown ifup
[22:54] <stiv2k> brb
[22:57] <qman__> it's always a good idea to reboot a new server at least once after it's set up, to make sure it comes back online
[22:57] <qman__> because it will inevitably reboot when you're not around to fix it
[23:00] <stiv2k> ok
[23:00] <stiv2k> wtf
[23:01] <stiv2k> qman__: after ifdown/ifup, the process is still there
[23:02] <qman__> that's what the other bit was for, sometimes they're stubborn
[23:02] <qman__> ifdown, kill manually, ifup
[23:02] <stiv2k> ah
[23:02] <stiv2k> ok
[23:02] <qman__> if it comes back, something is misconfigured
[23:03] <qman__> if not, you're good
[23:04] <stiv2k> i think i goti t now
[23:04] <stiv2k> got it*
[23:46] <smoser> lynxman, what is up ?