[00:06] <kirkland> RoAkSoAx: you da man!
[00:10] <macros> I'm having trouble getting console input/ouptut for an exec in a pre-start stanza in upstart when using console owner
[00:27] <josePhoenix> Hello all
[00:28] <josePhoenix> I'm trying to find a secure way to set a password non-interactively
[00:28] <patdk-lap> edit the shadow file
[00:28] <josePhoenix> That seems a bit.. dangerous, no?
[00:29] <patdk-lap> use usermod -p
[00:30] <josePhoenix> But that shows the password in the process list (at least for a second or so)
[00:31] <patdk-lap> use ldap or sql pam backend and just update ldap/sql
[00:32] <josePhoenix> eh perhaps
[00:33] <josePhoenix> Seems like there should be a way to read a password from a file, or non-interactively from stdin
[00:35] <josePhoenix> I guess I can just put it in a file twice and use the < operator
[02:23] <Ganymede> On an Ubuntu VPS, mysql never restarts when the server is rebooted but lighttpd, postfix, openssh server, and denyhosts start up fine as do cron and syslog. As far as I can tell with sysv-rc-conf and chkconfig, MySQL should be starting on runlevels 2345...but it just doesn't. I have to go in and type sudo service mysql restart manually after each reboot and it works then.
[02:25] <stgraber> Ganymede: do you know what technology is used by your vps provider?
[02:26] <stgraber> Ganymede: does /proc/vz exist in your vps?
[02:26] <Ganymede> Yes, it's OpenVZ and the directory /proc/vz exists.
[02:26] <Ganymede> stgraber, ^
[02:26] <stgraber> ok, so it's normal. I had similar issues with openvz and upstart
[02:27] <Ganymede> Okay...good to know it's not me being stupid.
[02:27] <Ganymede> I guess I'll just add it into rc.local.
[02:27] <stgraber> http://paste.ubuntu.com/672849/
[02:27] <stgraber> put that in /etc/init/openvz.conf
[02:27] <stgraber> that should simulate a standard boot (well, kind of)
[02:29] <Ganymede> VPS is rebooting ATM...
[02:29] <Ganymede> Surprised that a CONTAINER takes so long to reboot.
[02:31] <stgraber> depends how old their openvz server is
[02:31] <stgraber> newer openvz has a service that handles the restarts, old one does it with a cronjob
[02:31] <stgraber> so it can take up to 5 minutes
[02:32] <Ganymede> I currently have something very similar in my /etc/init/openvz.conf: http://paste.ubuntu.com/672852/
[02:32] <Ganymede> It only doesn't have the emit net-device-up lines.
[02:33] <Ganymede> At the moment, sticking in "start mysql" in /etc/rc.local appears to have done the trick although it's a sloppy hack.
[02:36] <Ganymede> Okay, whatever, I guess I'll just leave it at that. K, thx, bye.
[04:22] <ujid> NEW WINDOW
[04:58] <mark33> hiya
[05:00] <mark33> I'm trying to not use a clear text password in my svn command that is run via php command line - otherwise it ends up as clear text in the sudolog
[05:00] <mark33> sudo -u ABC /usr/bin/svn --password 'PASS123' update /www/website/ 2>&1
[05:01] <mark33> is there a way to not use "PASS123" and use a folder etc instead? (kinda new to this stuff)
[05:04] <jmarsden> mark33: --password "$(cat /path/to/file-with-pw-in-it)"
[05:05] <mark33> I'll try that, thanks
[05:05] <jmarsden> You're welcome.  No guarantees exactly what will be put in the log, but it will get the pw from the file :)
[09:48] <philipballew> I need help setting up noip2 on my server and getting it to run on start up
[10:13] <ersi> Daviey: Sounds like a good idea. :-)
[10:14] <ersi> hggdh: Ping! I'm a newcomer who'd like to get on track with some bugcontrol, and possibly some packaging
[10:22] <Daviey> ersi: good day sir!
[10:22] <ersi> Daviey: Heya :)
[10:23] <ersi> Grabbed another New bug by the way
[10:25] <Daviey> ersi: top banana!]
[10:30] <lynxman> Daviey: good day top honcho :)
[10:34] <Daviey> lynxman: hiya
[10:51] <beric> hi there. I'm having hard time with adding my debs to the installCD. The installer stops and in the logs I see a message saying my package was not found. when I look in /cdrom/dists/extra it is there. and I also have a proper Packages.gz file.
[10:52] <zyga> hi
[10:53] <zyga> who should I talk to with regards to server QA?
[10:54] <soren> Daviey: That'd still be hggdh, right? ^
[10:55] <Daviey> soren / zyga: Depends on the nature of the question TBH :)
[10:56]  * soren -> lunch
[10:56] <Daviey> hggdh is still favourable to us, but is not tasked specifically with server qa work AIUI.
[10:56] <zyga> Daviey: I'm interested in deploying a small ubuntu server image (automatically) to run tests (LAVA)
[10:56] <Daviey> soren: have fun munches.
[10:56] <Daviey> zyga: sounds good.
[10:56] <zyga> in general stuff related to automating the installation enough
[10:57] <Daviey> zyga: https://launchpad.net/lava-test ?
[10:57] <zyga> Daviey: lava-* (dispatcher most likely)
[10:57] <zyga> Daviey: lava-test works very well on an existing installation
[10:58] <zyga> Daviey: I just want to deploy it automatically in a VM (initally) and on some hardware (eventually)
[10:58] <Daviey> zyga: I have no personal experience with lava, sadly.
[10:58] <zyga> Daviey: that's okay, I'm working on lava :0
[10:58] <Daviey> zyga: For the issue you are describing, we use jenkins
[10:58] <Daviey> I suspect you want to re-use the code that handles that.
[10:58] <zyga> Daviey: I'm interested in deploying ubuntu-server/minumal somehow
[10:58] <zyga> Daviey: any links that I could use?
[10:59] <Daviey> zyga: hold fire
[11:01] <Daviey> zyga: You probably want to fork https://code.launchpad.net/~ubuntu-server-iso-testing-dev/ubuntu-server-iso-testing/trunk , and add a lava-test case.
[11:01] <Daviey> essentially, preseeding ubuntu-server and running code as defined.
[11:01] <zyga> Daviey: excellent, that's what I was looking for
[11:01] <zyga> Daviey: thanks!
[11:02] <Daviey> zyga: Handles libvirt + generating preseeds + spawning libvirt + running code.
[11:02] <Daviey> zyga: I would be interesting in seeing your changes and the end result :)
[11:03] <Daviey> zyga: That branch is largely based around recieve instructions from jenkins to fire a test.
[11:03] <zyga> Daviey: I'll stay in touch, it could be easier than I though
[11:03] <Daviey> zyga: super
[11:03] <zyga> Daviey: most likely we'll migrate to lava-dispatcher for invoking the actual tests, not jenkins
[11:03] <zyga> Daviey: as that fits closer to the rest of lava
[11:05] <Daviey> zyga: oh, totally understand.. I'm hoping to see if i can steal your fork for another project. :)
[11:27] <lynxman> any libvirt experts awake? :)
[11:35] <soren> lynxman: Wazzup?
[11:40] <Daviey> lynxman: You will usually get a better result if you just ask a question... or were you just doing a poll? :)
[11:45] <zul> morning
[11:45] <Daviey> zul: hey zul
[11:45] <Daviey> zul: Have you sniffed the libvirt in experimental?
[11:46] <zul> so libvirt 0.9.4 is in experimental it looks like it has better support for non-intel arches and better lxc support although its missing two features that is in the git tree
[11:48] <Daviey> zul: I think it needs to be sniffed to work with openstack, if it looks good - raise a FFe.
[11:48] <Daviey> I am concerned about the fact that we are considering a new upstream version of libvirt and qemu this late in the cycle.
[11:49] <Daviey> zul: Probably a good idea to test it against hallyn's qemu 0.15 build.. that would help add confidence to that aswell.
[11:52] <zul> okies i can do it this afternoon
[11:53] <zul> besides no one uses libvirt/qemu ;)
[11:57] <Daviey> zul: hah
[11:58] <Daviey> zul: https://launchpad.net/~serge-hallyn/+archive/virt/+sourcepub/1883706/+listing-archive-extra
[12:07] <ersi> FFe = Feature Freeze exception?
[12:07] <mdeslaur> ersi: yes
[12:09] <Daviey> ersi: Although, to help with confusion it also stands for Final Freeze Exception. So you have to work out what it likely means based on where we are in the cycle.
[12:09] <Daviey> *awesome*
[12:09] <ersi> Ah!
[12:10] <Daviey> zul: What are the non-intel arch enrichments btw?
[12:10] <zul> disabling dmidecode basically
[12:11] <Daviey> ah!
[12:11] <Daviey> zul: I'd love for the need for a damn wrapper to go away.. stripping things which don't work from the command line
[12:11] <Daviey> (such as isa-serial)
[12:17] <zul> libvirt is one big wrapper so you are never going to get rid of the wrapper :)
[12:17] <ogra_> Daviey, poke
[12:20] <ogra_> I#m just fiddling with preinstalled rootfs tarballs and see that our image publishing code always names files with .tar.gz suffix as "UEC/EC2 tarball", would the server team massively object if i renamed that a bit more generically to i.e. "root filesystem archive" or some such ?
[12:27] <Daviey> ogra_: o/
[12:28] <Daviey> ogra_: Sounds reasonable to me... this is for where?
[12:28] <ogra_> cdimage , the script generating the html pages for the images
[12:28] <Daviey> ogra_: Ie, for cloud-images.ubuntu.com it makes sense for that string to maintain.
[12:29] <Daviey> ogra_: cloud-images (UEC/EC2 tarbals are never on cdimage unless i am mistaken)
[12:29] <ogra_> well, but i guess the publishing code is the same
[12:29] <Daviey> ogra_: in any case, the UEC string needs to go away
[12:30] <Daviey> ogra_: Yeah, i think cloud-image is an old snapshot (hence using the old theme)
[12:30] <ogra_> let me see if i can make it conditional
[12:30] <Daviey> ogra_: rocking!  If you are poking that part of code, and making it conditional, changing "UEC/EC2 tarball" -> Cloud Images tarball would be most apreciated.
[12:31] <ogra_> i'll ping you about it again, i have to find out how/if the code is shared first
[12:32] <Daviey> ogra_: I think the code we are using is forked, but it's silly if that is the case.. we should fold back :)
[12:33] <ogra_> well, it might be forked but it could as well be a shared codebase, do you have any idea who set all this up ?
[12:35] <ogra_> aha, slangasek ...
[12:36] <lynxman> soren: hey, jmp was having some problems with libvirt, if you fancy having a look at it
[12:36] <lynxman> or you zul :)
[12:36] <Daviey> ogra_: our fork was smoser i believe.
[12:36] <zul> lynxman: im swamped right now
[12:36] <lynxman> zul: aren't you always :P
[12:36] <zul> lynxman: its a given
[12:36] <lynxman> zul: indeed :)
[12:37] <smoser> i can blame slangasek
[12:38] <lynxman> smoser: hey o/
[12:38] <soren> lynxman: "jmp"?
[12:38] <smoser> we have a version of the cdimages repo on nectarine, which was originally set up by slangasek.
[12:38] <Daviey> smoser: We should look at folding back, getting the updated theme for free makes sense :)
[12:38] <smoser> i've made necessary changes.
[12:38] <smoser> free != free
[12:38] <Daviey> smoser: No such thing as a free free?
[12:39] <lynxman> soren: Jose Plans, a kernel guy
[12:39] <smoser> merging a fork isn't "free"
[12:39] <smoser> lynxman, hey. whats up?
[12:39] <Daviey> smoser: looks to me like a good lesson on why not to fork, but push back upstream? :P
[12:39] <lynxman> smoser: not much, squashing bugs :) you?
[12:40] <ogra_> smoser, well, currently i'm only intrested in knowing how my change on cdimage would affect nectarine
[12:40] <smoser> ogra_, it would not.
[12:40] <ogra_> in our tree there desnt seem to be any other cloud related code
[12:40] <smoser> at the moment, there is a bzr repo there that a 'bzr pull' doens't pull.
[12:40] <ogra_> and i wonder why the bit thats there is there :)
[12:40] <smoser> its basically stand alone.
[12:41] <ogra_> right, i dont want to break mergeability though
[12:41] <ogra_> in case we want to fold them back into one tree at some case
[12:41] <smoser> slangasek originally got that code. at the beginning, i wasn't even knowledgable enough to know where it had come from or that I should be maintaining it upstream.
[12:41] <Daviey> ogra_: Oh, smoser doesn't care about rebasing.. forks rock.
[12:41] <smoser> this is ~ 2 years now.
[12:41] <smoser> we do want to fold them back, yes.
[12:41] <smoser> what is upstream for cdimages ?
[12:42] <ogra_> antimony
[12:42] <smoser> i can look to do that merge if necessary. but really, Daviey, on your list of high priority things is a theme of a download website ?
[12:42] <ogra_> they are mirrored to launchpad somewhere too
[12:42] <smoser> ogra_, well, if you would please find that I'll see what I can do
[12:43] <Daviey> smoser: Nah, it's not worth wasting time at the moment.. i'm pulling your chain, but the update theme should happen before release.
[12:44] <smoser> so to get this fully functional, utlemming or I will need the ability to push to the real upstream and/or request merges and/or break things :)
[12:46] <Daviey> hah
[12:48] <ogra_> smoser, https://code.launchpad.net/~cjwatson/ubuntu-cdimage/mainline and https://code.launchpad.net/~ubuntu-cdimage/debian-cd/ubuntu should be the public trees i think (i never use them from LP, so i dont really know)
[12:50] <Daviey> ogra_ / smoser: Looks like just a mirror
[12:50] <ogra_> merge request should go to the ubuntu-cdimage team
[12:59] <hallyn> apologies everyone for the late meeting notes.  Should be hitting your inboxes right now.  Too late to be useful, but early by government standards.
[12:59]  * hallyn not actually in today, goes back to hiding out
[13:03] <Daviey> hallyn: Thanks for doing that!
[13:03] <Daviey> zul: That means you are chairing today, right?
[13:03] <zul> am i? I did it a while ago didnt i?
[13:04] <Daviey> zul: looks like you are next in the pop queue.
[13:04] <zul> son of a bitch
[13:06] <ersi> Daviey: Looking at bug #831907, would it sound reasonable that it is a bug of Low importance? It only affects users that would turn a specific option (ssl_enable=YES) on, which is off by default if it is not in the configuration file.
[13:07] <Daviey> ersi: I would agree that is Low.
[13:07] <ersi> Daviey: Ok! I'll add an recommendation :)
[13:08] <hallyn> zul: yeah, i don't know what happened this last month, i did it just in july (and presumably you next)
[13:08] <hallyn> i think someone pulled a fast one with the meeting page header
[13:08] <zul> i blame moop
[13:08]  * hallyn looks suspiciously over toward daviey
[13:09] <ersi> Daviey: If ya' still got it open, feel free to change it's importance :)
[13:09] <hallyn> zul: if you want to try the debian experimental libvirt i think that should be fine, *except* when i did a 2-minute experiment using yesterday's git HEAD on an already weird VM, it failed configure due to some libxml2 weirdness
[13:10] <Daviey> ersi: done.
[13:10] <zul> cool beans ill try it this afternoon, im still in the process of merging it
[13:14] <hallyn> zul: i hope you're starting based on my 0.9.3-5 version!  the patch wrangling there was a bear and shouldn't be duplicated :)
[13:17] <zul> hallyn: yeppers :)
[13:21] <Daviey> hallyn: we enjoy duplicating effort! :)
[13:23] <zul> ttx: ping
[13:23] <ttx> just a sec
[13:32] <javatexan> morning all
[13:36] <javatexan> I have server with (wan) nic and (internal) nic.  I installed ubuntu server 11.04 with kvm.  I have a VM which will act as my firewall, router, etc.  I have VM bridged with wan and internal nic respectively.  Problem is, I only have one dhcp address available from above and with bridging seems to get two, is there a way to setup the wan nic on the host to not grab the dhcp, but still let bridge work?  I tried making fake int
[13:39] <ttx> zul: yes?
[13:40] <zul> ttx: for jamespage swift change you should have merged it in lp:~openstack-ubuntu-packagers/swift/ubuntu and notlp:~openstack-ubuntu-packagers/ubuntu/oneiric/swift/ubuntu
[13:41] <ttx> rha, sigh
[13:41] <ttx> what do we keep the other one around ?
[13:41] <ttx> why*
[13:42] <ttx> zul: can you fix it ?
[13:42] <zul> yep
[13:42] <ttx> kewl
[13:43] <zul> just wanted to let you know
[13:43] <ttx> zul: thx!
[13:44] <lynxman> zul: the merging of stuff for swift is at best confusing, did the same mistake myself :/
[13:44] <zul> lynxman: yeah you need a phd for it
[13:45] <RoAkSoAx> lol
[13:45] <lynxman> RoAkSoAx: when you want to do your presentation btw? I'm ready for you
[13:47] <RoAkSoAx> lynxman: well.... was unable to actually setup the environment yesterday as I ende dup testing ensemble more time than expected due to couple new bugs found
[13:47] <lynxman> RoAkSoAx: lol, alright then, tomorrow?
[13:47] <RoAkSoAx> lynxman: yes ;)
[13:47] <lynxman> RoAkSoAx: okay!
[13:47] <RoAkSoAx> thanks
[13:48] <RoAkSoAx> alright then, I'm gonna finish changes in orchestra today to setup the environment this afternoon
[13:48] <lynxman> RoAkSoAx: godspeed
[13:49] <zul> Daviey: whats the bug number for the keystone ffe?
[13:50] <Daviey> zul: bug 828689
[13:52] <zul> Daviey: thanks..
[13:53] <FlashDeluxe> hi! does anybody know a good tool for administrating clients in a network (e.g. installing programs, upgrading etc)?
[13:54] <smoser> https://bugs.launchpad.net/nova/+bug/828357/comments/8
[13:54] <Daviey> zul: throw that bad boy in the queue!
[13:54] <zul> just finishing it up
[13:54] <Daviey> super!
[13:55] <Daviey> zul: Next week, we probably need to smoke the dashboard love.
[13:55] <smoser> that bit of information cost me $2.26
[13:55] <zul> Daviey: its already on my todo list
[13:57] <Daviey> smoser: Oh, you'll pay that yourself - or is work paying?
[13:57] <smoser> well i used an old credit card that i found that said 'dave walker' on it.
[13:57] <smoser> but *someone* is going to pay
[13:58] <zul> lol
[13:58] <lynxman> smoser: eventually
[14:02] <Daviey> smoser: super!
[14:21] <zul> Daviey: keystone and dtest uploaded
[14:21] <Daviey> zul: dtest?
[14:21] <zul> Daviey: yet another python testing framework
[14:23] <Daviey> zul: oh goody!
[14:24] <zul> whee
[14:29] <lynxman> zul: another one? *sigh*
[14:29] <zul> lynxman: yeah they should standardize
[14:29] <lynxman> zul: for real
[14:30] <zul> ttx: fixed
[15:01] <SmokeyD> hey everyone. I have a server already running with ubuntu server edition. I want to migrate that existing install to a new disk, and in the process, also make that full disk encrypted.
[15:01] <SmokeyD> I know how to setup encrypted partitions using luks, but I am not sure about the grub2 part of this. I want to create the new disk with an unencrypted /boot partition
[15:02] <SmokeyD> but all other partitions should be encrypted, so I guess I need to change something in the grub config of the old disk, so it knows that the partition is an encrypted partition, and it should ask for a passphrase during startup
[15:02] <SmokeyD> but I can't find proper docs on what I should do in grub to enable this
[15:03] <SmokeyD> I don't want to do a full clean install on the new harddisk, but just want to rsync over the data from the old harddisk, and reinstall grub in the mbr
[15:17] <queso> How do I setup an apt source to be trusted?
[15:21] <SmokeyD> queso, is it a repository from ppa.launchpad.net?
[15:23] <SmokeyD> queso, you can lookup add-apt-repository for adding repository's from ppa.launchpad.net, which automatically imports the key for that reporistory, making it trusted
[15:24] <queso> SmokeyD: okay, thanks
[15:26] <lynxman> Daviey: ping
[15:27] <Daviey> lynxman:
[15:28] <lynxman> Daviey: Working on #822142, proposed a merge for Oneiric but how should I proceed to fix it in Natty?
[15:28] <Daviey> bug 822142
[15:29] <Daviey> lynxman: is it an issue in natty?!
[15:29] <lynxman> Daviey: it's marked as such by you ;) thought you checked, I can double check
[15:30] <lynxman> actually the bug initial description comes from natty
[15:30] <Daviey> lynxman: ah, for some reason off the top of my head i thought it was oneiric only
[15:30] <Daviey> lynxman: SRU sounds suitable if you can reproduce it.  You know the SRU procedure?
[15:30] <lynxman> Daviey: you have way too many ducks in your head :)
[15:31] <lynxman> Daviey: nope, never did it yet
[15:31] <lynxman> Daviey: If you hold my hand I promise I'll do it nicely :)
[15:31] <Daviey> lynxman: rocking.
[15:32] <Daviey> lynxman: You could probably reproduce this in the cloud or vm to confirm you can hit the issue.
[15:32] <Daviey> lynxman: useful read https://wiki.ubuntu.com/StableReleaseUpdates
[15:32] <lynxman> Daviey: cool, will read this right away
[15:35] <lynxman> Daviey: meanwhile the oneiric fix is there for your merging pleasure
[15:35] <Daviey> lynxman: the most minimal SRU justification should be something like bug 578536 (see the description)
[15:37] <lynxman> Daviey: brilliant, thanks
[15:37] <Daviey> lynxman: remember, that if you don't use debcommit, use bzr commit --fixes lp:bugnum
[15:38] <lynxman> Daviey: I'll eventually remember that sometime :/
[15:38] <lynxman> Daviey: thanks!
[15:39] <Daviey> lynxman: is writtable someone you can issue a writt against?
[15:39] <Daviey> err Weit*
[15:39] <Daviey> ahh, Writ*
[15:40] <lynxman> Daviey: you => funny
[15:40] <lynxman> Daviey: English is my third language dammit :)
[15:41] <orudie> NGEN
[15:42] <lynxman> Daviey: I'll fix that commit message then and do it the proper way
[15:42] <Daviey> lynxman: hold fire.
[15:42] <lynxman> Daviey: *holding fire*
[15:43] <Daviey> lynxman: It would make our merge delta smaller with Debian if you directly grab the patch from upstream
[15:43] <Daviey> lynxman: rather than the extra  comment.
[15:44] <lynxman> Daviey: k :)
[15:44] <orudie> NGEN
[15:44] <Daviey> lynxman: For the changelog, i tend to state what it does, not what the issue was.  Ie, Resolve issue of not being able to FOO, by doing BAR.  Patch courtesy of (Baz | Upstream). (LP: #98797897)
[15:44] <Daviey> but that is what i do..
[15:45] <lynxman> Daviey: I want to learn from your best practices so... :)
[15:45] <Daviey> lynxman: heh
[15:46] <GrueMaster> I have some questions for server testing.  Namely iSCSI and (separately) ensemble.
[15:47] <GrueMaster> What tests do we have that I can try on my own hw?
[15:48] <baggar11> GrueMaster: iozone3 package?
[15:48] <GrueMaster> iozone for testing which?  iSCSI?
[15:49] <GrueMaster> Do we have some documentation on setting up iSCSI host & client?
[15:49] <Daviey> GrueMaster: depends 'what' you want to test? throughput?
[15:49] <Daviey> !iscsi
[15:49] <GrueMaster> I'm testing arm server.
[15:49] <orudie> NGEN
[15:49] <lynxman> orudie: NGEN?
[15:50] <Daviey> GrueMaster: http://testcases.qa.ubuntu.com/Install/ServeriSCSIRoot/KVMExample http://testcases.qa.ubuntu.com/Install/ServeriSCSIRoot are probably good references.
[15:50] <Daviey> there may be others.
[15:51] <GrueMaster> kvm...sigh.  I'll see if I can work around that.
[15:51] <Daviey> GrueMaster: well yeah.. it's not the kvm stuff i am pointing it you to.. it's the ietd stuff.
[15:58] <Daviey> meeting starting in 2 mins in #ubuntu-meeting, all welcome to attend :)e
[16:02] <bkerensa> Daviey: What meeting is it?
[16:04] <baggar11> GrueMaster: pretty vague on what you want to test. iscsi target or initiator, throughput or cpu utilization, or...
[16:05] <GrueMaster> I just need to setup iSCSI on arm.  I would assume both sides of the equation need to be tested.
[16:05] <Daviey> utlemming: -> #ubuntu-meetng ?
[16:05] <Daviey> bkerensa: Ubuntu-server meeting
[16:06] <bkerensa> Daviey: Oh ok so not Ubuntu-server-ops meeting :P
[16:07] <Daviey> bkerensa: There is no such thing. :/
[16:07] <zul> smoser: a newer version of libvirt will make life easier for lxc
[16:08] <smoser> zul, oh?
[16:08] <smoser> i suspect by that you mean different things will break
[16:08] <Daviey> chaps this could be in the meeting ;)
[16:48] <baggar11> GrueMaster: iscsi setup should be mostly the same as on x86 hardware
[16:49] <baggar11> GrueMaster: althoug I have no experience with arm
[16:49] <GrueMaster> Yea, looking through the info now.
[16:50] <hggdh_> Daviey: ready when you are
[16:50] <GrueMaster> Need to adjust the testcase info for non-kvm install.  Also, why are we changing the terminology from other server/client nomenclature?  On most everything else, it is host/client or host/target.  On iSCSI, it is target/initiator.
[16:51] <GrueMaster> Makes it a bit confusing.
[16:52] <Daviey> hggdh_: can we do it on thurs?
[16:53] <hggdh_> Daviey: certainly
[16:54] <SmokeyD> hey everyone. If I am running a 32bit ubuntu version, is it possible to change that to a 64bit version without doing a complete fresh install?
[16:54] <ersi> GrueMaster: Isn't that "iSCSI-speak"?
[16:55] <GrueMaster> ersi: I guess.  Seems to be a bit confusing when they refer to the same thing as host/client or host/target.
[16:55] <genii-around> SmokeyD: possible but not practical. Reinstall is actually easier.
[16:56] <baggar11> GrueMaster: as far as I know ubuntu comes packaged with either iscsitarget or tgt for targets and open-iscsi for initiators
[16:56] <GrueMaster> With the world screaming for standards, you'd think we could start with nominclature.
[16:56] <SmokeyD> genii-around, well, my machine has quite a lot of config, so I would then have to copy configuration stuff, remember which packages I had installed, etc
[16:57] <ersi> GrueMaster: Heh, sure. Where are you referring to that "we" use iSCSI-speak by the way?
[16:57] <GrueMaster> baggar11: Good to know, but the definition of an iSCSI target is different than say an openssh target.
[16:57] <GrueMaster> ersi: http://testcases.qa.ubuntu.com/Install/ServeriSCSIRoot/KVMExample
[16:57] <SmokeyD> genii-around, what would it involve to switch the installation from i386 to amd64? Are there docs on it somewhere?
[16:58] <ersi> GrueMaster: Ah.
[16:58] <GrueMaster> Do we call an NFS or Samba server a host or a target?
[16:58] <baggar11> GrueMaster: I would think that has to do with iscsi dealing with block devices and openssh not
[16:58] <GrueMaster> See the confusion?
[16:59] <baggar11> GrueMaster: I didn't create the terminology :)
[16:59] <genii-around> SmokeyD: I have not found any docs on it. But i did this myself as an experiment once. You change the arch in your apt.conf, dump a list of installed packages then use that to reinstall them all in the new arch
[17:01] <GrueMaster> I'm just saying that I am having to work around the different naming conventions.  I am doing multiple arm server tests, and in this instance (especially on a lower caffeine level) it can get confusing when doing iSCSI on one side, and NFS Root on another.
[17:02] <GrueMaster> The target should be the target fs on the iSCSI Host.  the initiator is also the Client (diskless) system in my thinking.
[17:02] <GrueMaster> At any rate, I am getting around it.
[17:03] <zyga> de	ded
[17:04] <genii-around> SmokeyD: You can also just call apt-get with -o APT::Architecture=amd64 as well
[17:04] <SmokeyD> genii-around, ok cool.
[17:05] <SmokeyD> I am still considering. maybe I will actualy reinstall the machine from a fresh install
[17:05] <SmokeyD> is it easy to restore the installed packages from a previously made dump?
[17:12] <genii-around> SmokeyD: usually something like dpkg --get-selections  >> filename     and then dpkg --set-selections < filename
[17:12] <SmokeyD> genii-around, cool. Thanks a lot for all the help
[17:12] <genii-around> SmokeyD: I would still recommend to install fresh however.
[17:12] <SmokeyD> ge
[17:12] <SmokeyD> genii-around, that is what I am thinking
[17:13] <SmokeyD> do a fresh install and then restore the previously installed packages from a list of packages
[17:13] <genii-around> SmokeyD: Yes
[17:15] <SmokeyD> after that it would just involve copying the relevant stuff from /etc and data directories from the old installation
[17:16] <SmokeyD> genii-around, thanks for helping
[17:17] <genii-around> SmokeyD: You're welcome
[17:32] <smb> Daviey, zul, FYI just opened bug 832207
[17:33] <zul> smb:  cool!
[17:34] <RoAkSoAx> smoser:
[17:35] <smb> zul, Seems I at least got a usable combination of dump and vmlinux for it. Though evaluating that one takes a bit. And its not really usable as an attachment either... :-P
[17:35] <RoAkSoAx> smoser: ping
[17:36] <utlemming> qemu: hardware error: no boot device found
[17:36] <utlemming> CPU #0:
[17:36] <utlemming> R00=00000000 R01=00000000 R02=00000000 R03=00000000
[17:36] <utlemming> R04=00000000 R05=00000000 R06=00000000 R07=00000000
[17:36] <utlemming> R08=00000000 R09=00000000 R10=00000000 R11=00000000
[17:36] <utlemming> R12=00000000 R13=00000000 R14=00000000 R15=400140a4
[17:36] <utlemming> PSR=400001d3 -Z-- A svc32
[17:36] <smoser> RoAkSoAx, here
[17:37] <utlemming> qemu: hardware error: no boot device found
[17:37] <utlemming> CPU #0:
[17:37] <utlemming> R00=00000000 R01=00000000 R02=00000000 R03=00000000
[17:37] <utlemming> R04=00000000 R05=00000000 R06=00000000 R07=00000000
[17:37] <utlemming> R08=00000000 R09=00000000 R10=00000000 R11=00000000
[17:37] <utlemming> R12=00000000 R13=00000000 R14=00000000 R15=400140a4
[17:37] <utlemming> PSR=400001d3 -Z-- A svc32
[17:40] <Daviey> smb: probability of this getting fixed in oneiric?
[17:40] <RoAkSoAx> smoser: so I was looking into running python code on the cobbler snippets and I came to find out that cheetah can import modules and then stuff can be imported in the snippet. So I ended up using re-using your code for orchestra stuff
[17:40] <RoAkSoAx> smoser: by providing a module that can be imported by the snippet and return a late_command
[17:40] <Daviey> utlemming: hello!
[17:40] <Daviey> !ping
[17:42] <smoser> RoAkSoAx, oh, interesting.
[17:42] <smoser> what all are you doing in the snippet now?
[17:44] <utlemming> smoser: we've been advised to use the OMAP kernel for ARM images and use "beagle" as the machine type. The snippet is what you see when you attempt to invoke qemu.
[17:45] <smoser> did you attempt to provide it with a boot device ?
[17:46] <smoser> ie, what did you run
[17:46] <utlemming> qemu-system-arm -M beagle -m 512 -append 'root=/dev/mmcblk0 rw mem=512M console=ttyAMA0,115200 rootwait root=/dev/mmcblk0' -drive file=oneiric-server-cloudimg-armel.img,if=sd,cache=writeback --nographic -kernel oneiric-server-cloudimg-armel-vmlinuz-omap
[17:48] <RoAkSoAx> smoser: well I needed to pass keys to the orchestra-clients from the server for rsyslog as we are now using TLS
[17:49] <RoAkSoAx> smoser:  so I had to figure out a way to provide that keys so I though of doing something similar to what was done with ensemble
[17:49] <RoAkSoAx> smoser: but that required to mess up with cobbler, so I while researching found out about the cheetah importing modules on snippets
[17:49] <RoAkSoAx> so that's what I'm finishing right now
[17:49] <smoser> utlemming, you may need to say "boot=on" and/or -boot c
[17:49] <smoser> oh. wait, not -boot c
[17:50] <smoser> but probably boot=on
[17:50] <smoser> but that wont matter. your kernel is not even loading if thats all you're getting
[17:50] <utlemming> yup
[17:50] <RoAkSoAx> Daviey: ping
[17:50] <RoAkSoAx> Daviey: sorry had to go open the door
[17:51] <utlemming> I've found a whole bunch of stuff that says its looking for some sort of firmware loader
[17:51] <utlemming> http://balau82.wordpress.com/2010/03/10/u-boot-for-arm-on-qemu/
[17:56] <Daviey> RoAkSoAx: o/
[17:56] <RoAkSoAx> Daviey: what were you saying ? :)
[17:56] <Daviey> RoAkSoAx: my head is melting?
[17:57] <RoAkSoAx> Daviey: lol
[17:59] <ersi> Yippie, my first "Bug management" launchpad karma. :P
[18:01] <Daviey> RoAkSoAx: Ah, there are cobbler Fix Committed bugs.. i thought they were wedged on binary NEW queue, is this not the case?
[18:01] <RoAkSoAx> Daviey: which ones?
[18:01] <Daviey> bug #789266
[18:02] <Daviey> bug #705691 was wishlist, was that achieved?
[18:02] <RoAkSoAx> Daviey: cobbler-web already includes a working configuration
[18:03] <RoAkSoAx> it has had it for quite a long time
[18:03] <Daviey> RoAkSoAx: that is what i thought, the bug is assigned to you - can you update the status as appropriate please?
[18:03] <RoAkSoAx> sure
[18:04] <Daviey> rocking
[18:05] <RoAkSoAx> ;)
[18:05] <Daviey> If someone wants to hijack bug 810051 from zul, i don't think he'll mind!
[18:06] <Daviey> ersi: that is GREAT :)
[18:14] <ersi> Ooh, neat. A bug reporter replied to my comment
[18:14] <Daviey> \o/
[18:17] <Daviey> HOT bug of the day that someone might want to address: bug #776945
[18:22] <Daviey> smoser: bug #832123, did you see that as part of openstack - or standalone libvirt?
[18:23] <smoser> both
[18:24] <Daviey> smoser: awesome, utlemming ^^ did you see that?
[18:24] <Daviey> smoser: So that is a regression from last week?
[18:24] <utlemming> hummm...looking
[18:24] <smoser> Daviey, no.
[18:25] <smoser> just the first that i am able to actually get that far reproducibly
[18:26] <Daviey> smoser: ah, ok - stand down utlemming
[18:26] <zul> smoser: can you check to see if /dev/pts gets mounted?
[18:28] <smoser> zul, of course it did not
[18:28] <smoser> well, actually it does
[18:28] <smoser> twice
[18:28] <smoser> and  underneith /dev
[18:28] <zul> twice?
[18:28] <smoser> twice
[18:28] <zul> wha?
[18:30] <smoser> http://paste.ubuntu.com/673293/
[18:32] <zul> weird
[18:33] <smoser> yes, wierd.
[18:33] <smoser> this reproduces with ttylinux image. very basic busybox /sbin/init at this point.
[18:33] <smoser> its being done by libvirt
[18:34] <smoser> http://www.redhat.com/archives/libvir-list/2011-August/msg00213.html
[18:35] <smoser> "           * mount point in the main OS becomes readonly too which si not what"
[18:35] <smoser> i'm seeing that too
[18:35] <Daviey> smoser: so that would be resolved by using the new libvirt release?
[18:35] <smoser> shut up
[18:36] <smoser> s/would be resolved/could possibly be resolved/
[18:36] <smoser> s/could possibly be resolved/would likely be regressed to the point of not reaching this code/
[18:36] <smoser> :)
[18:36] <Daviey> smoser: heh
[18:37]  * zul feels vindicated
[18:37] <smoser> but yes, there are 2 cherry picks that by changelogs would seem to apply here.
[18:38] <zul> the mounting of lxc libvirt has been refactored btw
[18:40] <zul> er...the mounting of fileystems done by libvirt for lxc has been refactored
[18:53] <hallyn> smoser: (/me sneaks in for one quick comment) danpb did say in email last night that he would look into the libvirt symlink problem we uncovered yesterday
[18:53]  * hallyn disappears back into the shadows
[18:54] <smoser> hallyn, :) thanks. i consider that lower priority than the pts issue
[19:00] <hallyn> naturally
[19:19] <zul> bbl
[20:22] <philipballew> i cant seem to connect to my server from ssh after i attempted to set up a ddyddns cliant on it.
[20:30] <ersi> philipballew: Do you have any access to it what so ever?
[20:31] <philipballew> i have physical accsess for the next few days ersi
[20:32] <ersi> Were you able to SSH into it prior to setting up the dyndns client on it?
[20:32] <philipballew> ersi, i can type ssh philip@philip3
[20:32] <philipballew>  and log in.
[20:32] <philipballew> but not with my ddyns server name
[20:32] <ersi> Ah. Well, were you able to SSH to the external IP previously? Are you on the same LAN?
[20:33] <philipballew> i have never ssh'd from outside my networl
[20:33] <philipballew> i am on the same lan now
[20:33] <ersi> Chanses are that your firewall thinks you're being quite funny, if you're connecting on the LAN - but with the dyndns hostname which is your external IP
[20:33] <genii-around> Probably thinks you're spoofing
[20:33] <ersi> Do you have a machine on the outside, that you can SSH into, and SSH into your dyndns hostname?
[20:33] <ersi> yeh
[20:34] <philipballew> no i dont. i can always run to starbucks. heres a image bin of my router pf i did
[20:34] <philipballew> http://imagebin.org/169358
[20:34] <ersi> If you don't, you could give me (in a pm) (or here) the hostname and I could tell you if you're connectable from the outside
[20:35] <ersi> philipballew: It works :)
[20:35] <ersi> philipballew: If you check your logs, you have an login try for user 'ersi'
[20:35] <philipballew> what did you type that i didnt. send that in a pm
[20:35] <ersi> so if you'd run to starbucks, work, a friends house - you will be able to ssh into your machine as long as you have user/pass
[20:36] <ersi> philipballew: you can't connect while on the same LAN, then you'd have to use the internal "LAN IP"
[20:36] <philipballew> what do i type when i connect then?
[20:36] <ersi> your dyndns hostname, or the IP (69..)
[20:37] <ersi> both will work :) But when you're on the LAN, you need to use the 192.. one
[20:37] <ersi> or "philip3"
[20:37] <RoyK> http://jmckinley.posterous.com/dc-earthquake-devastation
[20:38] <philipballew> so ssh philip@philipserver1.no-ip.org  ersi ?
[20:38] <ersi> philipballew: Yeah, that will work when you're not at home :)
[20:39] <philipballew> alright. ill go to somewhere and see what i can do!
[20:40] <ersi> philipballew: Enjoy! :)
[20:40] <genii-around> As a side note, you may want to use a different port than 22, it is well known
[20:40] <ersi> And on another side note, make sure you got root logins disabled, and use a ssh-key instead of a password :)
[20:40] <adam_g> /win/win 15
[20:40] <adam_g> jeez
[20:59] <RoAkSoAx> adam_g: i have the same issue every single day :P
[21:03] <adam_g> RoAkSoAx: suppose its better than /lose/lose
[21:07] <RoAkSoAx> adam_g: heh
[21:07] <RoAkSoAx> adam_g: alright orchestra fixes finally uploaded
[21:07] <RoAkSoAx> adam_g: will let it build for a bit and i'll hit the lab
[21:10] <kirkland> RoAkSoAx: awesome \o/
[21:12] <RoAkSoAx> kirkland: ;)
[21:14] <adam_g> RoAkSoAx: cool.
[21:14] <adam_g> RoAkSoAx: do you know if orchestra is making use of cobbler-enroll, or is that for the ISO?
[21:16] <kirkland> adam_g: is cobbler-enroll Daviey's thing?
[21:17] <RoAkSoAx> kirkland: it is
[21:17] <RoAkSoAx> adam_g: haven't test cobbler-enroll though I think he was working on getting it working on the installer
[21:17] <kirkland> yeah, i haven't seen it land in oneiric yet
[21:18] <kirkland> we *need* something like that for sure
[21:18] <kirkland> but i doubt it's making 11.10
[21:18] <adam_g> yah, i knew it was going there but didn't know if orchestra was going to preseed it or what
[21:18] <kirkland> adam_g: check with daviey, but I'm betting that gets deferred for 12.04
[21:18] <kirkland> adam_g: and if so we should have a uds session around it
[21:18] <RoAkSoAx> indeed
[21:19] <kirkland> adam_g: as I have a few ideas how it could work, and I'm sure others do too
[21:20] <RoAkSoAx> ahhh crap
[21:20]  * RoAkSoAx forgot to re-enable the import of ISO's
[21:20] <adam_g> its pretty straight forward, but (AFAIK), its required to be coded in C (or shell?) if its to e included on the CD
[21:29] <RoAkSoAx> adam_g: C
[21:40] <hggdh_> kirkland, adam_g: if I undertood correctly, Daviey still hopes to get it in 11.10. See bug 831496
[21:41] <kirkland> hggdh_: ah, cool, that's great
[21:41] <kirkland> adam_g: ^ I'm outta the loop :-)
[21:41] <hggdh_> the only part I could not get done was the postinst integration with d-i... docs sucks there :-(
[21:42] <soren> adam_g: Is this something for the installer? If so, shell should be fine.
[21:43] <hggdh_> soren: it is both for the d-i and for normal usage, both packages get built
[22:24] <soren> smoser: Glancing at bug 828211, I'm kind of surprised. In my reading of the code, these calls happen before the clone() call.
[22:25] <soren> smoser: Do you have a specific example of where it happens inside the guest?
[22:31] <soren> smoser: Oh, I see it now. NEver mind.
[22:34] <soren> smoser: Did you consider using ifaceCtrl in src/util/interface.c?
[22:35] <soren> smoser: ...and also sticking the interface-renaming stuff in there and calling it from src/lxc/veth.c?
[22:35] <soren> Anyways, bedtime.
[23:07] <kirkland> RoAkSoAx: man, you went on an orchestra bug squashing safari :-)
[23:17] <SpamapS> man, I really like that oneiric has quiet turned off.. and plymouth tells you what upstart is up to.
[23:24] <RoAkSoAx> kirkland heheh i juet filed the bugs i fixed that i could remember to keep.track of them
[23:40] <Daviey> SpamapS: Are you able to look at the SRU's attached to bug 578536 please?
[23:41] <SpamapS> Daviey: looking
[23:41] <Daviey> Note that there are two Maverick's in unapproved.. the former really wants nacking as it reverts a patch, rather than carries the cherry picked patch from usptream to resolve it
[23:42] <SpamapS> Agreed and its been there for > 3 months .. oi
[23:42] <Daviey> SpamapS: Hey! we can't rush these things.
[23:43] <SpamapS> is it not present in natty, or are we just ignoring it for natty?
[23:43] <Daviey> SpamapS: unknown status regarding natty.
[23:47] <SpamapS> That introduces a regression when upgrading from maverick -> natty
[23:47] <Daviey> SpamapS: That isn't declared in SRU policy, is it?
[23:48] <SpamapS> No, its just something to consider.
[23:49] <Daviey> SpamapS: ack
[23:49] <Daviey> patches welcome.
[23:49] <Daviey> :)
[23:50] <SpamapS> Daviey: if nothing else, it needs to be documented in the bug that it most likely does affect natty, at least a New task. I'll add that.
[23:50] <Daviey> SpamapS: So, i did just add a task.
[23:50] <SpamapS> ahh well then. :)
[23:50] <Daviey> There is some debate going on at the moment if raising a task if it's unlikely to be addressed should be done or not
[23:50] <Daviey> Personally, i think it should
[23:51] <SpamapS> unlikely to be addressed should be Won't Fix then.
[23:51] <RoAkSoAx> lifeless: ping
[23:51] <SpamapS> Daviey: whats the point of a bug tracking system if we can't use it to keep users quiet^H^H^H^H^Hinformed? ;)
[23:52] <Daviey> hah
[23:52] <lifeless> RoAkSoAx: hi
[23:52] <Daviey> We don't do a very good job at exposing tasks that are really easy for people to fix, based on stealing patches from other uploads.
[23:53] <RoAkSoAx> lifeless: howdy!
[23:53] <RoAkSoAx> lifeless: kirkland referred to me as I believe you provided the squid config for orchestra
[23:54] <kirkland> RoAkSoAx: sorry, no
[23:54] <kirkland> RoAkSoAx: he just provided squid :-)
[23:54] <RoAkSoAx> lol/win 24
[23:54] <lifeless> RoAkSoAx: I'm a squid upstream :)
[23:54] <RoAkSoAx> oh lol
[23:54] <RoAkSoAx> lifeless: cool!
[23:54] <lifeless> RoAkSoAx: amongst other things. A bit emeritus these days, but still tracking progress.
[23:54] <lifeless> how can I help ?
[23:55] <kirkland> lifeless: sorry, RoAkSoAx has a syntax problem, maybe in a squid.conf;  i figured you'd be online this time a day, so I pointed him your way
[23:55] <lifeless> kirkland: sure
[23:55] <lifeless> cache_dir ufs
[23:55] <lifeless> bad idea :)
[23:55] <lifeless> cache_dir aufs
[23:55] <lifeless>          ^
[23:55] <lifeless> the a is important
[23:55] <deltree_> hi everyone, I'm new here
[23:56] <Daviey> lifeless: Can i take this moment to moan again about squid + ssl support? :)
[23:56] <lifeless> Daviey: blame the GPL
[23:56] <lifeless> Daviey: or get us a patch for TLS
[23:56] <RoAkSoAx> lifeless: this might be just trivial
[23:56] <Daviey> lifeless: We had this discussion some time ago.. I just wanted to moan again.
[23:56] <lifeless> RoAkSoAx: so http_access allow <things>
[23:56] <RoAkSoAx> lifeless: but the line "http_access allow lan"
[23:56] <lifeless> the <things> have to be defined above it
[23:56] <lifeless> its a 1-pass parser
[23:57] <lifeless> you have no acl line defining the land
[23:57] <lifeless> e.g.
[23:57] <RoAkSoAx> lifeless: right, so there's no "default" acl called lan
[23:57] <lifeless> localnet on lines 5 through 7
[23:57] <lifeless> RoAkSoAx: there are no default acls at all
[23:57] <lifeless> well, I lie slightly, but you *really* don't want to know the truth here ;)
[23:58] <RoAkSoAx> lifeless: right so yeah, I should just change from "lan" to "localnet"
[23:58] <lifeless> yes
[23:58] <lifeless> also change the ufs to aufs
[23:58] <RoAkSoAx> lifeless: awesome!thanks for the help!
[23:58] <lifeless> why do you have the memory set to 512MB ?
[23:59] <RoAkSoAx> lifeless: idk :) kirkland was the one who did the config
[23:59] <lifeless> kirkland: ^
[23:59] <RoAkSoAx> lifeless: but we were experiencing issues with not being able to cache stuff like the kernel