=== skaet_ is now known as skaet [00:38] I was curious if anyone had any idea why ubuntu did not have a make.conf, for compilation settings+c-flags, etc... For example more BSD-like systems, i.e. Gentoo & Archlinux have make.conf or makepkg.conf's [00:38] I just was not able to find anything on the net about it. [00:40] CyclicFlux: because we are a binary distribution and we only offer pre-compiled packages. [00:40] CyclicFlux: we do have dpkg-buildflags for example [00:41] xnox, Yes I know, with that said, there are utilities to compile from source/etc... i.e. apt-build, which I am playing with right now, I mean the system comes with gcc [00:41] So I was just curious where one would find that info/information, aside from /cat/proc/cpu [00:42] sorry *'cat /proc/cpu' [00:42] lol [00:43] But what I may not have been so clear on above is that ubuntu now has a utility, with a few helper utilities for compiling code from source, apt-build. [00:44] and I was setting the make options for it, etc... [00:45] compiling packages with specific cpu optimisations makes the resulting deb packages not portable to other machines. hence we compile to the lowest/common denominator [00:49] xnox, I def. know that I am not really into portability for my own system software. However for c, c++, java, python, etc... I def. go for portability. I am trying to optimize some of my system packages [00:50] sounds like you want gentoo :) === tgm4883` is now known as tgm4883 === Lcawte is now known as Lcawte|Away === n0ts_off is now known as n0ts === n0ts is now known as n0ts_off === cpg is now known as cpg|away [04:23] hmm i think cause iam using apt-fast .. the firewall isnt closing the connections quick enough.. and hence its getting stuck === jtv1 is now known as jtv === n0ts_off is now known as n0ts === n0ts is now known as n0ts_off [05:07] how do i reduce the number of connections apt-fast makes ? === dendrobates is now known as dendro-afk === cpg|away is now known as cpg [06:40] hi === smb` is now known as smb [07:15] Hi [07:15] I'm having a problem doing ./Config for anope... [07:15] even as root [07:15] i dont have permission to do it. === deegee__ is now known as drussell [07:47] Just installed 12.04 on a new server. Having some boot problems. When I boot, it gets past grub, but then nothing. Blank screen, not online (can't ping server). If I move the selector on the grub screen, but still choose the top option, it then boots. [07:48] Well sort of, I have another problem after that. [07:54] The other problem I have is when I do get past grub, it boots to ramdisk and I have to type exit at this prompt for it to continue. [07:54] Could be connected [08:04] Ah, never mind, found solution here: http://ubuntuforums.org/showthread.php?t=1976617 [08:29] Morning all [08:31] I've done a test upgrade on a VM instance from 10.04 to 12.04 and performance has become slugish at times. There's very little wa and no swapping going on. It's under minimal load. [08:32] *slugish, lags for upto 40seconds === Lcawte|Away is now known as Lcawte === railsraider_ is now known as railsraider === mrevell_ is now known as mrevell [09:26] I'd like to encrypt a text file in a simple way, what should I use ? [09:27] maxagaz: openssl, gnupg, ... [09:28] jpds: I don't want to have to use some key file to do it [09:28] just a passphrase [09:29] maxagaz: Those both have options just for that. [09:29] jpds: and both can easily be decrypted from ios ? [09:29] from some ios code [09:29] maxagaz: ios, as in Cisco, or Apple? [09:29] apple [09:29] maxagaz: No idea. [09:50] Daviey, hey, is there a guide for using cloud-init? i am using openstack but doing things "the hard way" and would like to find some examples or the like of how we are expected to do things === dendro-afk is now known as dendrobates === n0ts_off is now known as n0ts [10:02] apw: there is i suppose [10:03] https://help.ubuntu.com/community/CloudInit [10:07] koolhead17, ta [10:10] apw: if that doc isn't suitable, feel free to shout back for more [10:11] apw: what are you trying to outcome? [10:11] Daviey, thanks, i think now i have spotted just giving it #!/bin/bash lets me run my own code, i can just use that [10:11] apw: ok, if you are doing anything funky, would be interested in hearing out it. :) [10:11] Daviey, initialise an instance doing data handling, right now i start the instance, wait for it to start, pull the IP addy, ssh in, run random things to init it [10:12] apw: switching kernel? [10:12] Daviey, clearly the kind of shenanegans that cloud-init is designed to avoid one needing to do [10:12] Daviey, nope just a stock instance, all very simple, install two packages, bzr checkout some stuff, add cronjob, done [10:13] apw: sounds good.. if you get success, blog posts for this are always appreciated :) [10:13] Daviey, ok :) [10:15] Daviey, how big can your cloud-init data be ? [10:15] before one needs to host it [10:17] apw: 16K for the whole user-data [10:17] Daviey, ok so pretty big [10:18] apw: you can also gzip [10:30] * Daviey screams at keyrings. === dendrobates is now known as dendro-afk === cpg is now known as cpg|away === Assid|2 is now known as Assid [11:51] New bug: #1038044 in munin (main) "munin-node: apt_all does not Grok Ubuntu (possible regression)" [Undecided,New] https://launchpad.net/bugs/1038044 === dendro-afk is now known as dendrobates [12:52] Daviey, is there any way to reboot in cloud-init, as you offer update which may pull in a new kernel, does that reboot automatically? [12:56] apw: it does not [12:57] Daviey, that seems like something it should grow really [12:58] apw: yeah, smoser might like that suggestion [13:11] apw, you're saying you want first boot to say "oh, i need a new kernel", "apt-get update && apt-get install && reboot" ? [13:12] smoser, well i suppose i am saying i want "when i seleect update/upgrade mode that if a new kernel was installed to reboot and continue" [13:12] or something [13:13] smoser, the kernel does drop something to tell you, as the notifiers notice it [13:13] that was my next question. [13:13] /usr/share/update-notifier/notify-reboot-required [13:13] hm.. [13:13] thats a bad location for it [13:13] is it not? [13:14] something has to clean that up on next boot? [13:14] /run would seem better. [13:14] i think that is an _input_ to update-notifier rather than an output from the kernel [13:14] and indeed /run would seem more appropriate these days [13:14] well, there was /var/run before. [13:14] :) [13:15] so, apw, open a bug. i dont think its unreasonable. i would'nt make it the default though. [13:15] i think its a good suggestion [13:15] thank you. [13:15] smoser, no a nice new option perhaps .., against ? [13:16] cloud-init [13:16] rbasak, did you make my apt massively parralelizable yet? [13:17] (downloading) [13:17] although, if rbasak made installation parallelized too, that'd be cool [13:17] smoser: openssl emits notify-reboot-required aswell [13:18] emits ? [13:18] as in an upstart event? [13:19] Daviey, i assume my user init stuff is only run the first boot? nothing later ? [13:19] apw: correct [13:19] apw, right. [13:19] and it would makr that the upgrade had already been done [13:19] smoser: no, sorry.. creates the file [13:19] this is half way through the normal update. [13:19] er... hte normal boot. [13:19] smoser: fwiw, byobu also looks for that file [13:19] so some things would occur the "first time" on the "second boot" [13:20] i wonder what cleans it. [13:20] smoser, probabally the silly notifier thing, as it didn't go away on reboot :/ [13:20] anyway. this is doable. the biggest issue is that cloud-init gets ticked off when 'reboot' occurs while its running. [13:20] apw, i can just have cloud-init do: [13:20] reboot && rm -f that-silly-file [13:20] or the reverse order. === n0ts is now known as n0ts_off [13:21] apw: err, are you sure it;s /usr/share and not /var/run? [13:21] ahh [13:21] yes [13:22] /usr/share/update-notifier/notify-reboot-required is the script.. /var/run/reboot-required is the file to touch [13:22] so [ -f /var/run/reboot-required ] reboot [13:22] Daviey, oh, we run that ... doh [13:24] zul: the precise-staging uploads haven't been through CI yet, right? [13:24] .. and still no way to point CI to a PPA branch, is there? [13:25] s/branch// [13:25] Daviey: no...you would have to talk to adam_g about that [13:27] ta [13:33] dea?/w 156 [13:34] * apw screams at launchpad, timeing out when filing bugs today [13:40] apw: we found, our bug stats improve if the bug tracker doesn't function. The intention is, that you'll walk away, and we'll have less bugs raised.. improving our stats [13:42] Daviey, its working === smartboyhw is now known as ubuntustudio [13:44] apw: dammit [13:45] Daviey, am talking to launchpad peeps, probabally DC related === ubuntustudio is now known as smartboyhw [14:00] smoser, bug #1038108 [14:00] Launchpad bug 1038108 in cloud-init "cloud-init should offer an option to reboot following updates if the kernel was changed" [Wishlist,New] https://launchpad.net/bugs/1038108 [14:00] if i've just done sudo apt-get update can i now see a list somehow of what it's going to upgrade? [14:00] apw, gracias. [14:01] AdvoWork, you can type 'apt-get dist-upgrade' [14:01] and it will prompt you [14:01] before you continue (ie, that is safe) [14:01] you can also : apt-get dist-upgrade --dry-run [14:03] smoser, i dont mean for a dist upgrade,i mean just general new packages etc? [14:03] that is what dist-upgrade is [14:04] dist-upgrade is not "upgrade me from 11.10 to 12.04" [14:04] (that is a common source of confusion) [14:05] AdvoWork: dist-upgrade just takes you the latest updates in your current distro version [14:05] do-release-upgrade (not apt-get) takes you to next distro [14:05] next release, even [14:05] or next LTS release if you were running one [14:06] New bug: #1038108 in cloud-init (main) "cloud-init should offer an option to reboot following updates if the kernel was changed" [Wishlist,New] https://launchpad.net/bugs/1038108 [14:08] ahh ok i see, so dist-upgrade is just listing whats available since doing udpate? [14:08] *update [14:11] AdvoWork, update is only "download the listings of what is available" [14:11] "upgrade" is "apply them" [14:11] ('dist-upgrade' is just a more complete version of 'upgrade') === n0ts_off is now known as n0ts [14:22] smoser, thanks [14:22] I have a further query, if doing a symbolic link(so at the moment nothing exists in /auto/) is it ln -s /etc/myconfig.cfg /auto/ so that will make auto/myconfig (which links to /etc/myconfig.cfg) ? [14:26] AdvoWork, i think the answer is yes. [14:26] (as in if i understand the question properly) [14:27] but now I've got: lrwxrwxrwx 1 root root 14 Jun 20 15:54 test -> ../cfg/test (and the first test is blue). I've just done: ln -s cfg/new auto/ which has made it, but new is red. the permissions are the same, any ideas what the colour difference means? replacing myconfig.cfg with "new" [14:34] fixed it === n0ts is now known as n0ts_off === Assid|2 is now known as Assid [15:04] smoser: ping did the config-drive v2 make it in yet? [15:04] zul, yes. [15:04] smoser: cool just updating blueprints === Assid|2 is now known as Assid [15:42] New bug: #1028711 in glance (main) "glance-api.conf contains incorrect sqlite connection string path" [Undecided,Fix released] https://launchpad.net/bugs/1028711 [15:42] New bug: #1030152 in glance (main) "missing dependency on python-jsonschema" [High,Fix released] https://launchpad.net/bugs/1030152 [15:59] Daviey: f3 is uploaded [16:14] hallyn: ok so perhaps I am running into the umask problems... [16:15] hallyn: re the "failing to open /etc/sudoers" problem I was whining about yesterday [16:16] SpamapS: i didn't see that whine. [16:16] SpamapS: i've grepped and am not sure where the umask is being set. though i may have been looking at the wrong juju version [16:16] (since only some seem to have the probem) [16:16] hallyn: you did you're just forgetting I think [16:17] heh could be [16:17] /usr/bin/lxc-ls: line 36: cd: /sys/fs/cgroup/cpuset///lxc: Permission denied [16:17] thats also kind of weird [16:17] isn't it? [16:17] SpamapS: *that* is umask problem at lxc-start [16:18] lxc calls mkdir() and passes in the righ tmode to create the lxc dirs [16:18] yes, i suppose lxc could save the umask, set its own, do mkdir, then reset umask, but that's kind of silly [16:19] hallyn: so is that possibly caused by umask tightening on sudo run commands from 0002 to 0022 ? [16:19] i'd prefer to have lxc-start and lxc-create fail on bad umask [16:19] iirc juju was actually doing 0077 umask [16:19] $ sudo bash -c 'umask' [16:19] 0022 [16:20] edit lxc-create to save its umask to a tmpfile, create a container with juju, see what it sa\ys [16:20] we always call lxc through sudo so i don't think so [16:20] utlemming, i just pushed fix for growpart and nbd or loop devices [16:20] really? cool, I'll take a look...does it apply only to lucid? [16:21] smoser: ^ [16:22] i think i dont understand thw question. [16:22] smoser: where is the commit? [16:22] cloud-utils upstream just never worked for /dev/nbdX or /dev/loopX [16:22] ah, okay, now I understand what you're saying [16:23] smoser: I'll address the build system then for it [16:23] hallyn: ok confirmed that the umask is 0077 when the template runs [16:23] * SpamapS tracks down "why" [16:23] utlemming, right. [16:23] smoser: what did we decide to do with lucid full disk images? drop them or we're we going to choose an arbitrary size for the disk images? [16:23] we could sru that easily enough. but not really necessary forthis. [16:24] i dont know what to do there. [16:24] anything over 5G seems "too big" for some use cases. [16:24] and its a PITA to resize down [16:24] anything under 5G seems "too small" [16:24] so i really dont know what to do [16:24] maybe 10G isn't too small. i dont know. [16:24] smoser: right...could we SRU the cloud-init-ramfs that does that? [16:25] smoser: I know that's a long shot, but it seems like with openstack and others that we might be able to make a compelling case for doing so [16:25] no. we can't really sru that. [16:25] its a new package. [16:25] i dont know. you could ask what sru team thought about it [16:25] it is very stable and very limited use case. [16:25] right [16:26] but it'd mean getting cloud-utils growpart back to lucid (not hard), *and* cloud-initramfs-utils a new package added to lucid. [16:26] (and in main) [16:26] i personally would'nt want to waste my effort. [16:26] ugh, yeah, that wouldn't work [16:27] so for lucid on openstack, right now you have 2 less than ideal things: [16:27] so I'm leaning towards make arbitrary 10G images and just saying that's a limitation of lucid [16:27] a.) the uec images: these use kernel and initramfs directly, meaning kernel upgrades dont work [16:28] b.) the "full disk images". these support kernel upgrades, but without manual user action, the root filesystem is 1.4G , unless we choose a different size for them. [16:28] and 'a' has a work around, the loader kernel (http://people.canonical.com/~smoser/lucid-loaders/) [16:31] utlemming, checking resultant size of 10G [16:32] I resized one yesterday -- the qcow2 size is only slightly bigger [16:38] yeah. for the one i tried i see [16:38] $ ls -lh /archive/mirrors/uec/lucid/current/lucid-server-cloudimg-amd64-disk1.img disk1-10G.img [16:38] -rw-r--r-- 1 mirror mirror 216M Aug 17 01:51 /archive/mirrors/uec/lucid/current/lucid-server-cloudimg-amd64-disk1.img [16:38] -rw-r--r-- 1 root root 217M Aug 17 12:34 disk1-10G.img [16:38] 216M -> 217M [16:38] i would have expected more dirty-ness inserted [16:38] so maybe, yeah, 10G images sound resonable. [16:39] but we should send a mail to ubuntu-cloud announcing them and their issues. [16:39] okay, I'll do that [16:40] utlemming, we/you are targetting next wednesday for your cloud-init image creation changes, right? [16:40] that is FF. [16:40] smoser: yes sir [16:40] ok. [16:40] harlowja and i put this together for cloud-init work [16:41] http://pad.daviey.com/cloud-init-for-quantal [16:45] zul, where are we with releasing an openstack based on 2012.1.2 (and sorry if you've been asked this already) ref: http://lists.openstack.org/pipermail/openstack-announce/2012-August/000020.html [16:45] med_: we are going to start looking at it next and it will be in -updates when its finished [16:46] nod. [16:58] utlemming, ok. so i just tested natty [16:59] and i hit a race condition i think [16:59] race where? [17:00] i think its this race: [17:00] https://bugs.launchpad.net/ubuntu/+source/cloud-initramfs-tools/+bug/937352 [17:00] Launchpad bug 937352 in cloud-utils "root partition may not be grown" [High,Fix committed] [17:00] but the result is different. [17:00] root was resized, but then mount failed [17:01] in 2 instances i launched, i saw it once [17:02] http://paste.ubuntu.com/1152972/ [17:02] * SpamapS gets closer to chasing down this 0077 umask [17:04] smoser: is this a udev race? from the log you can't tell whether it /root or ../cloudimg-rootfs that is missing [17:05] utlemming, what happens is this: [17:05] * root is mounted [17:05] * growroot runs, and calls growpart $ROOT_DEVICE [17:05] * growroot unmounts $ROOT_DEVICE [17:06] * growroot mounts root device [17:06] (i'm not usre about the order on those in natty) [17:06] but the issue is after unmounting and at the instance when the mount tried, udev had deleted the devices and not yet recreated them. [17:07] (oneiric likely suffers from some similar issue, although i've not really seen it) [17:07] that's what I saw going on [17:11] i think for natty the best thing to do is just not create the -disk1.imgi files. [17:21] New bug: #460925 in lxc (main) "Kernel log message corruption due to incomplete /proc separation" [High,Triaged] https://launchpad.net/bugs/460925 === cpg|away is now known as cpg === matsubara is now known as matsubara-lunch [17:45] adam_g: what would it take to be able to deploy openstack in ci, from a PPA? [17:46] adam_g: there seems to be two obvious ways that jump out at me.. either the dput support jamespage was working on, to dput to jenkins, and test the package, then forward it to the PPA [17:46] or the second. byhand say.. "test this whole PPA" [17:46] what do you think? [17:54] Daviey: you just set a flag in each charm's config that points to a PPA [17:54] Daviey: we already do that, and point it at the local CI reprepro archive [17:55] Daviey: http://paste.ubuntu.com/1153070/ [17:56] Daviey: that will install all components from the trunk testing ppa === cpg is now known as cpg|away === matsubara-lunch is now known as matsubara [18:07] adam_g: right.. but how hard for someone to ssh in, run ./test_from_staging.sh :) [18:08] Daviey: ideally nobody would be doing that [18:09] Daviey: you'd login to jenkins and hit a button instead :) [18:09] adam_g: sounds super... have you done it? :) [18:11] Daviey: for that specific PPA, no [18:11] Daviey: but thats essentially how all of our testing works [18:11] adam_g: right.. i get that.. but this differs. [18:11] Daviey: how so? [18:12] adam_g: just that it is on-demand, rather than a trigger really [18:12] but hey, you know this more than me anyway :) [18:13] sounds like you have it in hand. [18:14] Daviey: right, so there'd be a job there that only triggers when someone hits the button. i have a similar job for installing everything from precise-proposed [18:16] New bug: #1038197 in samba (main) "samba changes in system-config-samba or /etc/smb.conf do not affect server" [Undecided,New] https://launchpad.net/bugs/1038197 [18:16] Daviey: but that dput support sounds really cool and better, for testing certain things. deploying the entire cluster and debugging after its broken is a really expensive way to find trivial packaging mistakes [18:17] adam_g: you put in a new openvswitch in 12.04.1 didnt you? [18:18] zul: it looks like it never made it out of proposed === cpg|away is now known as cpg [18:18] tho clint mentioned in bug that it was due out last monday [18:18] SpamapS: ^ [18:20] bug #1021530 [18:20] Launchpad bug 1021530 in openvswitch "[SRU] update to include stable fixes for OVS 1.4" [Medium,Fix committed] https://launchpad.net/bugs/1021530 [18:21] adam_g: yeah! [18:21] New bug: #1038199 in bind9 (main) "Bind9 9.8.1 high CPU utilization" [Undecided,New] https://launchpad.net/bugs/1038199 [18:31] adam_g: for 12.04.1 they locked down precise-updates thats the only reason it hasn't been published [18:31] adam_g: and now its Friday, so we'd rather not publish on a friday [18:31] New bug: #1034946 in dnsmasq (main) "dnsmasq and network manager broken if dnsmasq package has 'enable-dbus'" [High,In progress] https://launchpad.net/bugs/1034946 [18:39] adam_g: bah [18:55] hallyn, ping [18:55] do you have a simple example of working libvirt xml for a domain on virbr0 [18:57] sure, does http://people.canonical.com/~serge/cdboot.xml suffice? [19:10] hallyn, thanks. thas similar to what i have that wasn't working :-( [19:32] hallyn, ugh. [19:32] thi sis frustrating [19:33] http://paste.ubuntu.com/1153232/ [19:34] this works: sudo kvm -drive if=virtio,file=/tmp/lp-1035279/disk.img -curses [19:35] this works: sudo kvm /tmp/lp-1035279/disk.img -curses [19:35] but via libvirt it acts like it can't find anything to boot off of [19:35] sits and spins trying to boot network, disk, network, disk [19:38] smoser: you are trying to play sol.exe arent you [19:39] well, yes, but only because i just beat http://www.erikyyy.de/invaders/ [19:40] smoser: i dunno, should work... does it work without the 'boot order', or with order set to 0? [19:42] smoser: youdon't have the boot dev=hd tag that i have in cdboot.xml [19:42] (part of the tag) [19:45] GAH! [19:47] hallyn, i blame you [19:47] bbl [19:47] you are the reason we can't boot in 512k of memory anymore. [19:48] clearly, 64k *should* be enough for anyone. [19:48] but not even grub [19:49] hah [19:51] smoser: it's because i'm always pushing for all that eyecandy and video === mcclurmc_away is now known as mcclurmc === cpg is now known as cpg|away === cpg|away is now known as cpg [20:16] New bug: #1031761 in python-warlock (universe) "[MIR] python-warlock" [Undecided,Fix released] https://launchpad.net/bugs/1031761 [20:22] New bug: #1031757 in python-django-compressor (universe) "[MIR] python-django-compressor" [Undecided,Fix released] https://launchpad.net/bugs/1031757 [20:22] New bug: #1032242 in python-cliff (universe) "[MIR] python-cliff" [Undecided,Fix released] https://launchpad.net/bugs/1032242 [20:26] New bug: #1031759 in python-django-appconf (universe) "[MIR] python-django-appconf" [Undecided,Fix released] https://launchpad.net/bugs/1031759 === cpg is now known as cpg|away [20:38] My MaaS / JuJu deployment has an issue after a 'juju destroy-environment' did not complete successfully. I no longer have an environment but nodes are still marked as deployed. [20:39] Any ides on how I would "release" them? === cpg|away is now known as cpg [20:58] hallyn: hey, KSM should allow greater density of concurrent VM's right? [21:07] New bug: #1036101 in postfix (main) "package postfix 2.9.3-2~12.04.2 failed to install/upgrade: le sous-processus script post-installation installé a retourné une erreur de sortie d'état 75" [Undecided,New] https://launchpad.net/bugs/1036101 [21:07] New bug: #1036817 in postfix (main) "package postfix 2.9.3-2~12.04.2 failed to install/upgrade: subprocess installed post-installation script returned error exit status 75" [Undecided,New] https://launchpad.net/bugs/1036817 [21:09] ^^^ someone please go explain that the error message actually explains you have to fix your configuration if you would actually read the error message that is. [21:09] ScottK: Error: "^^" is not a valid command. === cpg is now known as cpg|away [21:22] I have a problem with kerberos: "kinit: Client not found in Kerberos database while getting initial credentials" all I did was follow the Kerberos guide on two VMs but on the client kinit doesnt seem to work - see http://pastebin.ca/2195577 -any ideas? [21:26] Daviey: uh, well, yes, adn by quite a bit, but of course that just means that when things go wrong you'll get slammed all the harder [21:29] hallyn: right, just checking i wasn't going nuts. [21:33] Daviey: now if you're using larger pages, it's possible that ll of the pagecache wins you should gain are lost because of small offset differences... not sure. i've not really measured that [21:34] Anybody on able to to assist with MaaS ? Following along here: https://wiki.ubuntu.com/ServerTeam/MAAS/Juju I get stuck at juju bootstrap. === Haruspex is now known as Shehrazad [21:39] moep: Your pastebin looks like you're running as root first, then as a user... perhaps the user you're running kinit from is not registered? http://linux.die.net/man/1/kerberos [21:42] seijirou: doesnt matter if I use root or a user, always the same [21:44] moep: Hmm. Well according to the kerberos man page, that error means you haven't been registered as a Kerberos user. Maybe the guide is missing a step. [21:47] seijirou: accoring to wireshark that is a server reply that the principal I want to use does not exist, however listprinc tells me otherwise [21:53] Sorry i'm not having any brilliant ideas [22:41] . === bitmonk_ is now known as bitmonk === cpg|away is now known as cpg === n0ts_off is now known as n0ts