[00:09] jamespage, i just pushed a change to https://code.launchpad.net/~ubuntu-server-ec2-testing-dev/+junk/ec2-automated-tests.smello for the ec2-automated-tests . it seems that maybe because its +junk that i can't propse it for merging [00:09] anyway, i propose that for merging. [00:10] and my plan is to just copy 'smello' binary to each release. [00:14] smoser: right, +junks can't use merge proposals [00:27] anyone have any ideas on why I can ssh into my server without a password but when trying to run rsync from same machine it gives me an error http://pastebin.com/11PypSGi [00:29] Daviey, updated https://wiki.ubuntu.com/PrecisePangolin/TechnicalOverview [00:30] hello all... any mdadm users here? [00:30] chrislabeard, no real help, only ftp://ftp.samba.org/pub/unpacked/rsyncweb/issues.html [00:30] and it suggests some things to check. [00:31] a little confusing as to why rsync runs fine when I don't use ssh [00:31] I guess I will look at these files [00:31] thanks [01:02] In sid I have a Unattended-Upgrade::Origins-Pattern -- but on my lucid box I only see Unattended-Upgrade::Allowed-Origins [01:02] Can I use the former? [01:03] I have an in-house apt repo that includes e.g. lucid openldap packages automatically rebuilt against openssl, and currently unattended-upgrades is upgrading from one of those to the newer one in lucid-security, thereby breaking sudo-ldap [01:03] So the simple solution is to add my PPA's origin pattern, except that I don't understand the matching style of the Allowed-Origins variable. [01:05] Hmm, I think it's just Origin and Suite, so I will try adding "Cyber IT Solutions lucid-cyber" to the list and see what happens. [01:14] Why does unattended-upgrades recommend mailx? AFAICT it just uses cron to send notification emails, and cron uses /usr/sbin/sendmail, not mailx [01:15] twb: perhaps because * add "mailx" to suggests (LP: #137994) [01:16] Oh WTF, there's an /usr/bin/unattended-upgrades *as well as* the /etc/cron.daily/apt that apt ships [01:17] OK, I see, /etc/cron.daily/apt does everything except the actual package upgrade, and calls u-a to do that. [01:18] I still don't see why u-a doesn't just emit to stdout and let cron take care of it :-/ [01:18] Or use sendmail directly [01:18] IMO that is wrong [01:19] If you're invoked via cron, bypassing cron's mail configuration is naughty [01:19] Otherwise e.g. if I change MAILTO=fred@telstra.com to /etc/crontab and you are still sending to root@localhost, I will lose your mail [01:20] wouldn't it be advisable to alias root's mail to an account that is watched? [01:20] I started at a company that hadn't done that for 4 years.. [01:20] SpamapS: yes, but that's orthogonal to my complain [01:20] I found SO much broken stuff going through and just running 'mail' as root [01:20] backups failing.. disks filling up... crazy [01:20] SpamapS: the one that was fucking me was... lemme find the ticket... [01:21] So can someone tell me if I have a snippet in my orchestra/cobbler preseed that is showing up in the d-i as $SNIPPET('snippet_name') does that mean it failed to compile? [01:21] http://bugs.debian.org/645286 [01:21] twb: But cron doesn't allow me to specify custom subjects which makes all cron emails look equally uninteresting. [01:21] erichammond: tough [01:21] erichammond: if you care enough fix that in cron, not by patching individual jobs [01:22] ipl31: yes most likely. Do you have other snippets working that way? [01:22] (FWIW the subject comes from the job, so the direct issue there is using cron.daily instead of cron.d) [01:23] twb: nice bug btw.. hah! [01:23] The fix is grotesque [01:23] s/fix/kludge/ [01:24] twb: looks like they fixed it upstream for you [01:24] twb: but then dropped it [01:24] SpamapS: This particular snippet is in last command and it is surrounded by two others that are working [01:24] SpamapS: yeah, I seem to have missed that email [01:25] Are there any code examples that show manipulating the interfaces file with snippets? That is essentially what I am trying to do [01:25] twb: best be subscribing to those next time. ;) [01:25] * SpamapS always forgets to subscribe to the bugs he files in Debian [01:25] ipl31: no but .. netcfg is pretty powerful [01:25] anyway, I'm late... [01:25] It ought to have subscribed me automagically to the one I filed, surely? [01:25] * SpamapS disappears [01:25] twb: no, Debian's BTS does not subscribe you to bugs you file [01:26] OH [01:26] yeah [01:26] So it's just normally people hit group reply [01:26] exactly [01:26] SpampS: does netcfg support bonding, vlans etc?? [01:26] SpamapS: does netcfg support bonding and vlans? [01:26] sorry about the double post [01:26] ipl31: not sure.. [01:26] anyway, I must be going [01:26] SpamapS: thanks [01:56] Sigh. [01:56] Unattended-Upgrade::Allowed-Origins "Cyber IT Solutions lucid-cyber"; fails, because unattended-upgrades(1) assumes that the Origin is a single word. [02:14] RoAkSoAx: eh? [02:16] zul: eh what? [02:17] Daviey: zul I think we'll also need to add user/password fields for each system so that users can change them [02:17] zul: cobbler system, add user/password fields for each of the systems so administrators can change admin user/password of the system they are deploying [02:18] im thinking im missing context here what was the problem [02:20] zul: someone was asking how they can change the user/password for the machine they are deploying [02:20] zul: and it makes no sense to manually edit a preseed file to do it [02:20] ah ok [02:20] yeah [02:20] zul: so we should add user/pass to cobbler systems and then add a snippet that will automatically do that [02:21] RoAkSoAx: probably worry about it tomorrow though :) [02:21] Bah, symmetric crypto blows [02:21] You need to add a public key infrastructure to it :-) [02:22] Say kerberos, that was you can put a big sticker on the front "compatible with AD" [02:22] zul: lol yeah it's should be fairly easy though [02:22] What could POSSIBLY go wrong [02:27] There should be a better way to handle dkms/m-a updates to a bastion where you don't want to install gcc [02:28] Specifically, I scripted an automated rebuild of xtables m-a whenever a new linux-generic update comes down the pipe, BUT, it won't be installed because there is no metapackage that depends on the latest xtables-addons-modules-NNN-generic. [02:28] And I am surely not the only one who wants to avoid putting gcc on an xtables-using firewall... === hopp3r is now known as st3fann === bladernr_ is now known as bladernr_afk [03:54] Hey, guys. I need help! === DanaG_ is now known as DanaG [04:30] anyone around? [04:46] !anyone [04:47] Stupid bot. [04:47] jjjrmy: what is your question? [05:03] hey guys, i was wondering if anyone knows what brand of cpu is typlically better for ubuntu? AMD vs intel it seems like ubuntu isnt optomized for an intel like windows is, but it also seems like AMD is lacking in some features. thoughts? [05:04] I am talking from an OS / schedualing view point. [05:06] T3CHKOMMIE: x86-64 CPUs it doesn't matter a damn to Ubuntu [05:07] T3CHKOMMIE: obviously both intel and amd make a few CPUs of different architectures (e.g. tegra) which is a whole other ballgame [05:08] twb, really? i was looking at i7 vs bulldozer... seems like the way windows and ubuntu sheudal the threads on the hardware can either help or hinder performance based on number and load of threads... [05:08] twb, do you know of any cpu features that dont work yet on the linux kernel? [05:08] What I mean is that Ubuntu binaries are compiled to a baseline x86-64 ABI. So anything "cool" about a particular CPU is not going to be Ubuntu-specific. [05:08] for example, turbo clocking with intel etc. [05:09] Regarding specific features: I have no idea. [05:09] Generally I buy intel because an AMD-based box is more likely to contain nvidia components, and nvidia is actively hostile to the FOSS movement [05:10] foss? [05:10] And generally I'm building things like file/print servers, so anything newer than a Pentium 4 will be perfectly fine [05:10] T3CHKOMMIE: Free and Open Source Software [05:10] gotchya. [05:11] T3CHKOMMIE: i.e. the ideology behind the software [05:11] its been my experience that nvida drivers work better on ubuntu than AMD or ATI. is that yours aswell? [05:12] If you're talking about GPU drivers, I do not run non-Free GPU drivers, so I have no idea. [05:12] ya, gpu. thats interesting. why dont you use non free gpu drivers? and how do you get "free" gpu drivers? [05:13] Free GPU drivers ship with your system [05:14] The reasons to use them are the same as reasons to use FOSS in general. Try http://fsf.org/philosophy [05:14] Sorry, I meant http://www.gnu.org/philosophy/. The FSF one seems to be some horrible blog thing. [05:15] (Stupid johnsu01...) === Kutakizukari is now known as Guest31917 === DanaG_ is now known as DanaG === himcesjf1 is now known as himcesjf [07:36] hi [07:37] my ufw not activated on boot. How to do that? === smb` is now known as smb [08:33] Good day [08:34] I'm using Ubuntu-server 10.04 and trying to configure mail server with Dovecot as SASL server for postfix [08:38] I'm using standard manual, but postfix on port 25 doesn't show AUTH string [08:58] hi all [08:59] hi [09:05] [09:24] morning' [09:31] * uksysadmin tips hat to Daviey [09:40] morning all === palt_ is now known as palt [09:58] good morning :) === sanderj is now known as Sander^work [11:01] hi all [11:26] New bug: #898583 in nova (main) "nova_sudoers allow unused commands" [Undecided,New] https://launchpad.net/bugs/898583 [11:44] jamespage: morning! [11:44] Ursinha: bom dia! [11:45] morning lynxman [11:45] lynxman: dia! [11:46] hellos lynxman :D [11:47] koolhead11: ello o/ [11:48] so how much your going to pay!! lynxman :D [11:48] koolhead11: oh hell, don't even remind me about it :/ [11:48] lynxman: :D [11:49] hola Ursinha Daviey [11:49] man, there not a single dlna client in ubuntu [11:49] Ursinha: there's a couple java based ones I reckon [11:49] Ursinha: not Ubuntu specific though [11:49] lynxman: if it runs in ubuntu that's fine [11:49] problem is I just can't find them [11:49] Ursinha: it does :) [11:49] I'm using my android phone for that [11:49] Ursinha: http://en.wikipedia.org/wiki/List_of_UPnP_AV_media_servers_and_clients [11:50] AH wikipedia [11:50] Ursinha: the source and failure of all information :D [11:50] helo evrybody i was installing ubuntu server 10.04 in my machin but the there is no graphical interface is that psible to make having graphical interface please help me? [11:50] binyam: yes sir, it's possible, it's just not installed as default :) [11:50] binyam: it is. :) [11:51] binyam: installing ubuntu-desktop package is one way of getting it.. [11:51] * koolhead11 wants to modify the bot to answer these questions!! [11:51] !ping [11:51] another contentless ping... sigh... [11:51] :O [11:51] Ursinha: xfce is better way [11:51] !ping ubottu [11:51] ubottu: that's rude! you should read Ubuntu CoC [11:51] Ursinha: I am only a bot, please don't think I'm intelligent :) [11:52] ursinha: iam trying in instlalling ubuntu-desktop but is ther any disadvantage installing ubuntu-desktop? [11:52] binyam: ubuntu-desktop is basically unity and all related stuff.. but you might want something lighter for a server, like koolhead11 suggested, as xfce [11:52] koolhead11: so he might want to install xubuntu-desktop? [11:53] Ursinha: not sure but binyam google is your friend don`t forget :D [11:54] I know, I just don't like to point people directly to google if I can remember stuff from the top of my head :) [11:54] sure, google knows it all :) [11:55] Ursinha:thanks anyway [11:55] can someone add !google kind feature to give result url here :D [11:56] !google [11:56] While Google is useful for helpers, many newer users don't have the google-fu yet. Please don't tell people to "google it" when they ask a question. [11:56] lawwal [11:56] yeahhhhhhhhhhhh [11:56] * Ursinha hugs ubottu [11:57] ubottu needs a friend [11:57] lynxman: I am only a bot, please don't think I'm intelligent :) [11:57] that's the spirit :D [11:57] I'm ubottu's friend :P [11:57] * Ursinha loves bots [11:57] #foreveralone [11:57] Ursinha: lol! [11:58] haha [11:58] Ursinha: defender of lost causes and advanced scripts ;) [11:58] hehe [11:58] binyam: let us know if you need any more advice :) [11:58] lynxman: yes I am [11:58] defender of the lost causes [11:58] Ursinha: ;) [12:05] Koolhead11:u seems bone head not kool am i right/ [12:05] hi, im currently using the ec2 natty ami, and im curious how its been setup. theres a 300GB partition mounted at /mnt, but its not showing up as an ebs volume. any ideas what this is? [12:05] did you install the ebs ami? [12:06] binyam: indeed i am at times http://www.ubuntugeek.com/how-to-install-gui-on-ubuntu-11-04-natty-server.html [12:06] patdk-lap: yep [12:07] i just did "gui ubuntu server" [12:07] that is the normal local disk then, that would have been your primary drive, if you didn't do ebs root [12:07] Ursinha:and is that possible to acesss out side of any country to my server is there any kind of confiugration that can i make alrady i have installed ssh [12:07] patdk-lap: the main os is on an ebs volume mounted at /dev/sda and when i snapshot it the core filesystem is there, just not anything in /mnt [12:07] wierdly i can write to it fine [12:07] otaku_coder, yes, it works fine, but it's kindof like a ram disk [12:08] binyam: so, it depends :) [12:08] when you destroy the instance, it's gone [12:08] but when you restart the instance, it will be there again, but blank [12:08] binyam: if you want to access your server from anywhere inside the same network, so you just need to make sure sshd is running [12:08] but as long as the instance is running, it's just a normal local disk [12:08] patdk-lap: right, bit confusing as its a farily large storage space [12:08] basically just think of it as a higher speed temp drive [12:08] binyam: if you want to access your server from anywhere else in the world, you need to have a valid ip to connect to [12:08] well, local storage that isn't highly redundant is cheaper to make larger :) [12:09] patdk-lap: ok, ive got mongoDB writing its data there, looks like i need to move it to an ebs volume fast! [12:09] binyam: so you need to make sure the router or whatever your server is connected to is configured to let you reach the server from outside [12:09] otaku_coder, depends [12:09] maybe leave it there and do hourly backups out of it [12:09] patdk-lap: well I intended to use ebs snapshots as a way to back it up [12:09] or whatever your application requires [12:09] binyam: here at home I have an irc proxy machine (among other stuff), so I configured my router to redirect ports [12:09] all depends on what you require [12:10] I store lots of stuff on ramdisks, and only back it up hourly or daily [12:10] Ursinha:so how can i makesure wether it is connected or not ? [12:11] binyam: not sure what you mean :) you need to understand what's the path to get from your computer to outside world [12:11] binyam: here at home my computer is connected to a router, which is connected to the world [12:11] so to reach my computer I need to do the inverse path [12:11] binyam: got it? [12:12] Ursinha:yeah i gote it let me ask some qoistion later [12:13] Ursinha:the installation of the ubuntu-desktop is alrady finish but nothing display/ [12:13] patdk-lap: how do you normally back it all up? especially with databases where its difficult to keep a consistent state [12:14] binyam: so, it's no magic :) you need to ask it to [12:17] why is it difficult? [12:17] how is a database that isn't consistent useful to use at all, if you can't back it up? [12:18] Ursinha:How coud i ask it? [12:18] lsb_release -a [12:18] binyam: so, you're installing a server, you need to understand at least a bit of what you're doing, otherwise you can get yourself in trouble [12:19] Ursinha:yeah but iam not familar with ubuntu thats why iam asking you? [12:19] binyam: to be really honest with you, ubuntu isn't much different of other distros when it comes to which tools to use to do specific things [12:20] and the concepts are the same [12:20] concepts are the important things [12:20] Ursinha:so what supoose to do now? [12:21] binyam: understand what you are doing :) [12:21] binyam: you need to run X, I'd start lightdm [12:21] binyam: sudo service lightdm start [12:21] X should start and then you can just login [12:22] Ursinha:ok [12:24] Ursinha:itsays t o me lightdm :unrecognized service [12:24] Patrickdk: false alarm, im using a replicaset so can pause one of the slaves to take a backup. :) [12:32] binyam: try to run startx and see what happens [12:42] Ursinha: thanks igot it what i need thanks very much [12:45] binyam: no problem :) next time, if you need more help regarding server, you can check here: https://help.ubuntu.com/11.10/serverguide/C/index.html [12:47] Ursinha: thanks [12:50] !serverguide [12:50] The Ubuntu server guide may be found at http://help.ubuntu.com/10.04/serverguide/C/ [12:50] thanks ubottu :D [13:12] koolhead11: that is the natty serverguide, we might want to update that :P [13:12] natty? no, maverick? [13:12] Ursinha: will do that when precise comes [13:13] koolhead11: why is that? [13:14] Ursinha: i think there is some rule/way/practise to keep LTS guide as url [13:14] koolhead11: ahh, I see [13:27] !servergui [13:27] Ubuntu server does not install a desktop environment or X11 by default in order to enhance security, efficiency and performance. !eBox provides a GUI system management option via a web interface. See https://help.ubuntu.com/community/ServerGUI for more background and options. [13:27] Ursinha: ^^ [13:28] Ursinha: We've purposely left ubottu and the /topic pointing to the latest supported LTS for server related thigns. [13:30] Pici: cool :D [13:35] thanks koolhead11 and Pici :) [13:50] good morning [13:55] zul: good moaning [13:55] hey lynxman === bladernr_afk is now known as bladernr_ [14:36] Daviey: yay, netcf is in unstable! [14:43] * ahs3 ^5s hallyn [14:43] ahs3: :) === bladernr_ is now known as bladernr_afk === bladernr_afk is now known as bladernr_ [15:15] zul, Did we not have some bug report request to add modprobing xen_gntdev to the init scripts (at least for precise). Finally understood what that is good for and it might be nice to have for Oneiric, too... [15:15] smb: i think we did wouldnt it be easier just to build it into the kernel as well? [15:16] less delta with debian if we did that as well [15:17] zul, On one side yes, on the other I am not yet sure a) whether it may have effects on normal installs b) increases the kernel size without it being boot essential (sort of) [15:17] k neither am i but yeah we can do that for precise i think [15:17] im also concenered if there is a modprobe order [15:18] What I found nice about it being in the init script was that I can dual boot into the no-xen side and all the xen modules are not loaded [15:19] zul, Seems the gntdev is only used by userspace driver to foreign domain communication. So I would say no there [15:20] zul, Right now I was experimenting and loaded it after anything else but before starting a new domU and it was ok [15:21] smb: ok ill take a look at it [15:22] zul, ok. well I may try to get it prepared and ask you to sponsor. I just for heck cannot find that bug report again [15:22] I am sure I saw it fly by [15:22] cool [15:24] * smb is waiting for a number from zul... ;) [15:24] smb: damn hold on [15:26] smb: i could have swore there was one as well [15:26] * RoyK gives smb a 42 [15:26] zul, heh, ok. I am pretty sure there was, as well but lp search... [15:26] bug 42 [15:26] Launchpad bug 42 in launchpad "Bug description listed in task is not the correct description" [Medium,Fix released] https://launchpad.net/bugs/42 [15:27] bug 1 [15:27] Launchpad bug 1 in ubuntu "Microsoft has a majority market share" [Critical,In progress] https://launchpad.net/bugs/1 [15:27] * smb watches lp explode [15:27] bug -1 [15:27] bug sqrt(-1) [15:27] bug `ls /` [15:28] RoyK: trying to exploit the system? ;) [15:28] lynxman: http://xkcd.com/327/ [15:29] hello zul :) [15:29] hi koolhead11 [15:29] RoyK: supp [15:29] zul: anything for me today? [15:30] koolhead11: no [15:30] ok sir. : [15:30] RoyK: ah little Johnny tables [15:31] http://xkcd.com/984/ yesterday's xkcd was rather cool too :) [16:03] soren: ping what happened to https://github.com/openstack/openstack-integration-tests [16:13] smoser: ping [16:14] smoser: bug #898697 [16:14] Launchpad bug 898697 in orchestra "automatic importing of isos should be configurable" [Undecided,New] https://launchpad.net/bugs/898697 [16:14] smoser: there's now a debconf question that asks whether to run it or not during install [16:14] smoser: and if you decide not to, you can configure what you want to import or not in the config file [16:14] smoser: so I'm closing this bug report, is that ok with you? [16:18] RoAkSoAx, how do I get to try this? This affected me. [16:20] is there any IRC server in main? [16:21] uksysadmin: this is released in Precise unfortunately. I cannot backport to oneiric as there's been changes on squid that would make orchestra fail in oneiric [16:21] kirkland: no [16:21] uksysadmin: i'll probably add backwards compatibility to be able to backport [16:26] RoAkSoAx, that's ok - I'm up for the challenge - I'll grab a build of Precise... I'm still at getting the thing working and April will probably be a time when we'll be looking at using this anyway... [16:27] uksysadmin: cool, if you are gonna rebuild, what you probably need is just change the dependency from squid to squid3 in debian/control [16:27] if i update the RAM allocation in KVM, and restart the VM, why wouldn't it show up in top? [16:27] ok [16:28] guest top that is === Ursinha is now known as Ursinha-lunch [16:52] RoAkSoAx, how does that work ? [16:52] the postinst runs 'runone' immediately [16:53] what debconf question were you talking about ? [16:54] smoser: Precise [16:54] smoser: http://pastebin.ubuntu.com/756138/ [16:55] carp. [16:55] i was reading out of date trunk/ source [16:55] smoser: its not in PPA though because the transition to squid3 was made and it will break oneiric [16:55] not in PPA for oneiric* [16:57] i was just reading out of date source [16:57] suck [16:57] sorry for bothering you. [16:58] hi guys [17:04] smoser: no worries :) it happens [17:04] i just marked that as a dupe of bug 892328 [17:04] Launchpad bug 892328 in orchestra "debconf question to run orchestra-import-isos or not during installation" [Medium,Fix released] https://launchpad.net/bugs/892328 [17:05] ok ;) [17:13] smoser: have you seen the cartoon animaics? [17:14] yeah. [17:14] smoser: they had a bit called good idea or bad idea, along the same lines would you say this is a good idea or a bad idea: people.canonical.com/~chucks/add-nova-connect.patch [17:15] hello nurse [17:15] good idea: pasting links with the protocol included [17:15] bad idea: pasting links you can't copy [17:15] http://people.canonical.com/~chucks/add-nova-connect.patch [17:16] http://paste.ubuntu.com/756170/ [17:16] boto's ec2 connection is harder to use than it should be [17:17] yeah [17:17] thats why i wrote a shortcut [17:17] or at least it should have a generic one that takes an endpoint rather than a region item. [17:18] next to "connect_walrus", it seems to make sense. [17:18] but i'd suggest a generic 'connect_ec2_endpoint' or something [17:19] so you are syaing you are ok with that but it would also be nice to have a generic one? [17:19] well i think its worth sending upstream [17:20] but a generic one makes sense to me. [17:20] k [17:20] thanks [17:20] i dont think i have the engergy to write a generic one [17:22] smoser: are you using boto for the ec2-image build process? or is this something else? [17:23] zul brought this up. and its for nova (openstack) [17:23] smoser: k, thanks -- just want to make sure we're not overlapping [17:23] utlemming: no im using it write some perfomance metrics stuff for nova [17:23] * zul lunches [17:24] zul, http://smoser.brickies.net/git/?p=xc2.git;a=blob;f=xc2_util.py;h=fc4878beb8922cf8ea4b8d2971c2adfc6e78377a;hb=HEAD [17:25] does something like the generic one i was talking about [17:25] if region == none, then it parses the url given (in ec2_url) [17:31] smoser: cool thanks [17:48] zul, http://paste.ubuntu.com/756203/ [17:48] something like that. [17:49] nifty [17:50] * RoAkSoAx redhat cluster breakge coming today :) [17:52] RoAkSoAx: w00t [17:53] Daviey: ehehehe its been tested by ivoks , everything works as expecgted, by given the upstream changes things will definitely break on upgrades from lucid -> precise [17:53] Daviey: as configuration methods have changed, binaries have changed, etc [17:55] rocking :) [18:01] any kerberos experts here? [18:04] hi all :) [18:57] adam_g: Have you been able to look at the keystone charm yet? [19:01] Daviey: not yet, no. i wanted to poke at openstack for the first time yesterday but sidetracked on bug 883988 [19:01] Launchpad bug 883988 in glance "db migration failing when upgrading glance - trying to create existing tables" [High,Confirmed] https://launchpad.net/bugs/883988 [19:07] adam_g: Great :) [19:10] adam_g: did you test squid3/orchestra in oneiric? [19:11] RoAkSoAx: it appeared to be working fine out of the box when i submitted the change. is it not? [19:11] adam_g: just wondering cause I wanna change the Dependency from squid to squid3 for oneiric backports [19:12] adam_g: that's why I was wondeirn if you tested this in oneiric too [19:12] Hmm [19:12] hi there.. i have a karmic-server which i would like to update and potentially upgrade, where can i find the repos ? [19:12] Something isn't quite working right for me with that. [19:12] But it might be unrelated. [19:13] Daviey: maybe the wrong network for squid? [19:13] though, it shouldn't [19:14] pukeko: Take a look at https://help.ubuntu.com/community/EOLUpgrades [19:14] thanks [19:14] RoAkSoAx: no, i haven't on oneiric [19:14] SpamapS: I just got zookeeper 3.3.4 uploaded into Debian unstable; includes upstart configuration for Ubuntu; tests to support MIR are currently disabled; they do run but not reliably... [19:15] adam_g: ok [19:20] jamespage: NICE .. I think the juju team is excited about 3.4 ... [19:20] jamespage: maybe we can patch up the tests and try to get fixes upstream [19:21] SpamapS: well to be honest I think its the build environments borking more than anything else [19:21] I can run them fine on my laptop in a clean chroot for either precise or sid [19:21] my sponsor had issue tho === gema is now known as gema_afk [19:22] SpamapS: 3.4.0 is appealing but still considered beta; I have a branch I'm working on for it; might get it uploaded to experimental/PPA [19:38] Daviey: so freeze is over. Did you want to push the etherboot/ipxe bits from me? :) [19:46] jjohansen: stgraber: can we pick a time tomorrow to talk about lxc and apparmor and security (to make Daviey happy)? [19:47] hallyn: sure, I good with any time you want [19:47] 17:00 utc? [19:48] (11am central, 9am pacific, i think) [19:49] cool, thanks, i'll ping tomorrow at that time and hope for the best :) [19:49] I have an ARB review/packaging shift tomorrow until 19:00 UTC tomorrow, will be around after that [19:50] so do you prefer 20:00 utc for a break? [19:51] New bug: #898363 in samba (main) "Command: echo hold is not working" [Undecided,New] https://launchpad.net/bugs/898363 [19:51] yep 20:00 should work fine [19:51] stgraber: cool, thanks. talk to you then. [19:51] jjohansen: ^ [19:52] (now i'm going for a walk toclear my head before some scary uploads) [19:52] hallyn: works for me too [19:56] New bug: #790863 in linux (main) "Unable to start lxc container after update to 2.6.32-32" [High,Confirmed] https://launchpad.net/bugs/790863 [20:13] so I just installed ubuntu-orchestra-server on a fresh oneiric server. why didn't that also install isc-dhcp-server ? [20:13] it asked me about address ranges. [20:19] utlemming, http://paste.ubuntu.com/756362/ [20:19] do you know anything about that? [20:20] tgardner, it does not assume that it is the owner of your dhcp infrastructure. it can live without it. [20:21] smoser, yeah, I just figured out that it install dnsmasq by default [20:22] ah. yeah, cobbler works with dnsmasq. [20:22] smoser, tgardner: cobbler will also work with isc-dhcp-server - but by default orchestra recommends dnsmasq and provides suitable config for it [20:23] hmm: maybe that should be dnsmasq | isc-dhcp-server.... [20:23] jamespage, I'm trying to figure out if dnsmasq is smart enough to assign specific MAC addresses to DNS names (and IP addresses) [20:25] tgardner: probably - /usr/share/doc/dnsmasq/examples/dnsmasq.conf.example has lots of options with descriptions [20:25] it can assign based on the hostname presented from the client via DHCP - that can be managed through cobbler [20:26] jamespage, I'll check it out. thanks. [20:26] tgardner, np [20:27] jamesone other question, what is the cobbler web URL ? [20:27] tgardner, cobbler_web [20:27] hostname/cobbler_web [20:28] smthats the one. thanks === Resistance is now known as EvilResistance === medberry is now known as mka [20:37] evening people [20:38] I _think_ I might be having apparmor troubles with bind - I'm setting up my slave server, and I'm getting this when I do a rndc reload: http://pastebin.com/60QX0WAM [20:38] could anyone help me find out for sure? === mka is now known as medberry [20:58] mjau^: apparmor would be logging denials in /var/log/syslog [20:58] SpamapS: ah, so it's not apparmor giving me a hard time then? [20:58] oh. found what was wrong. [20:59] hadn't created the foo.se. dir. [21:01] utlemming, would you be willing to review https://code.launchpad.net/~smoser/ubuntu/precise/python-boto/2.1.1/+merge/84002 ? [21:02] the merge is fairly straight forward if you ignore the file-id stuff (ie, you can see smaller changes by 'bzr export' of the tree for the from and to versions and diff -Naur. [21:10] Daviey: I'm going ahead and pushing ipxe (with a small change) [21:10] Daviey: i'm going to wait on etherboot until tomorrow [21:24] gah - ipxe upload rejected [21:28] Guess i'll be waiting for daviey after all [21:29] daviey: http://people.canonical.com/~serge/ipxe-2.debdiff fwiw === james__ is now known as blitzkrieg3 [21:57] hallyn: Great! [21:57] Will grok it first thing in the morning [21:58] daviey: thx, i just moved the rules steps to override_dh_install as that's more appropriate for that than override_dh_auto_install [21:59] groovy [22:09] zul: do you have an oneiric pandaboard powered on and handy? [22:10] if so any chance you could veify sru fix for 884407? (it should be quick) [22:33] anyone know much about vnc? if i run a kvm instance launched with vnc running on port "0.0.0.0:1" , that port "1" - what is that? That can't be tcp or udp is it? [22:39] Daviey: still around? [22:43] RoAkSoAx: no. :P [22:44] help please dovecot postix Error: ssl_key_file: Can't use /etc/ssl/private/ssl-cert-snakeoil.key: Permission denied have googled made new certs and checked permissions any advice ?? [22:44] Daviey: lol, so I'm still not sure whether we want a default *preseed* or default *profile* to do the enlist [22:45] Daviey: lol, so I'm still not sure whether we want a default *preseed* or default *profile* pointing to the enlist preseed ... to do the enlist [22:46] RoAkSoAx: right [22:46] RoAkSoAx: benefits of profile? [22:47] Daviey: having a separate profile that is defaul in the PXE Menu [22:48] Daviey: once the machine is enlisted, then it will just PXE boot from the pxefile generated for that particular system [22:48] Daviey: in comparison if we use a default preseed for *all* profiles, that means that we need to find a way to 1. use the enlist stuff. 2. once it has been used, then remove that from the preseed, or replace the preseed [22:48] Daviey: which seams a bit more complicated [22:49] RoAkSoAx: right, but do we want to encourage people doing 'default preseed' installs to systems which are not enlisted? [22:50] erm [22:50] RoAkSoAx: we do that already [22:50] Daviey: if we don't have system, there won't be a default preseed because it will do the enlist stuff [22:50] Daviey: if we don't have system, there won't be a default preseed install because it will do the enlist stuff [22:50] RoAkSoAx: we want all machines to enlist, right? [22:51] Daviey: yes, [22:51] one they enlist, with what i showed you, they will not re-enlist uless the 'system' is deleted in cobbler [22:51] Daviey: if the machine is already enlisted, then it could use *another* preseed to deploy [22:51] it /must/ use a different preseed [22:51] the default one will not be used on an already enlisted system [22:52] Daviey: right, how do you change that? [22:52] RoAkSoAx: do you have ipv6? [22:52] Daviey: no I don't [22:53] ok, nevermind [22:53] RoAkSoAx: steps: [22:53] 1 - default = enlist [22:53] .. a mac address preseed is auto generated [22:53] .. machine will boot from that rather than defaut [22:54] when we want to install, it puts a proper preseed and boot in there. [22:54] When we just want to localboot, the mac address preseed provides a localboot directive [22:54] the machine should only ever use the 'default' one once. [22:55] If i go to the ui and delete the system, the machine should boot into enlist again to re-enlist. [22:55] Daviey: right, but 1.default = enlist means => precise-i386 pointing to default.preseed, right? [22:56] Well... this blocks arm. [22:56] Daviey: now, when you add the system, it will point to profile precise-i386, which will point to default.preseed, see my point? [22:56] erm, no - not directly using precise-i386, using the same base system, but with an extra cobbler-enlist preseed values. [22:57] anybody know if there's a way to invoke "kdb5_util create" non interactively? [22:58] Daviey: what's the "base system" [22:59] RoAkSoAx: so i am actually currently using an enlist profile with that as the default timeout :) [22:59] Daviey: in order for it to pxe boot, you need to add a "profile" that wil tell where the preseed is [23:00] Daviey: right, you are modifying the *default* file under pxelinux.cfg/ [23:00] Daviey: but instead of manually modifying files, isn't it better to do it on top of cobbler [23:00] Daviey: cause this changes mean changes to *stock* cobbler [23:00] RoAkSoAx: well, i added a 'profile' in the webui [23:00] which created, http://pb.daviey.com/0RxC/ [23:01] then manually changed, ONTIMEOUT enlist [23:01] zul: It got renamed. [23:01] zul: To... er.... [23:01] zul: t-something. [23:01] zul: Tempest! [23:01] Daviey: right, which is basically what I was saying, add a *default* *profile* not a preseed :) [23:02] RoAkSoAx: Heh, works for me :) [23:03] Daviey: well that's what I was saying, we need a default *profile*, which means, if it cannot find the pxe file under pxelinux.cfg/01-, then use pxelinux.cfg/default that will run the enlist *profile* [23:04] RoAkSoAx: That sounds clean enough, it's how i'm doing it locally. [23:04] RoAkSoAx: I don't know how best to auto create a new profile? [23:04] Daviey: thought, I'm also saying that we can avoid the wait of seconds until it TIMEOUT's and defaults to *enlist-profile* by creating a system called *default* that will replace pxelinux.cfg/default and could tell it to enlist [23:05] RoAkSoAx: Also, how can we care for arm with this aswell? [23:05] That is a hard problem. :( [23:06] Daviey: what's broken from the ARM stuff? [23:06] Daviey: i'll do the autoconfiguration stuff in orchestra [23:07] Daviey: ipxe? [23:09] RoAkSoAx: ipxe can't work on arm [23:09] RoAkSoAx: the problem is, currently we are talking about making enlist use precise-i386 [23:09] this breaks arm enlistment [23:10] RoAkSoAx: uboot could be set to tftp fetch 'default-arm' ? [23:10] Daviey: you were TIL for mysql-cluster-7.0 ... I'm going to import 7.1.17 and fix bug #660379 .. mmmkay? [23:10] Launchpad bug 660379 in mysql-cluster-7.0 "Make mysql-cluster depend on libmysqlclient16 from mysql-5.1 (rather than conflict)" [Wishlist,Confirmed] https://launchpad.net/bugs/660379 [23:11] Daviey: Idk actually [23:13] SpamapS: i never object to people hijacking my TIL's :) [23:14] Daviey: I promise I'll treat your TIL nicely [23:15] * SpamapS shoves Daviey's TIL in the back of the van and speeds away [23:16] SpamapS: thanks :P [23:18] Daviey: well so I guess we can either make the administrator select enlistment for either i386/x86_64 or arm [23:18] RoAkSoAx: i think you can config what file gets pulled in uboot.. i'd need to experiment [23:18] Hey guys. What do you think the role should be if you're highering someone to specialize in deploying high availability in web servers and maintain them [23:18] RoAkSoAx: At the moment, all arm reference hardware we have doesn't have native PXE booting. This means that /some/ config via uboot can be expected. [23:19] system administrator or systems engineer seems to generic or ambiguous [23:19] Daviey: cause either way, arm is not imported automatically [23:19] RoAkSoAx: right, we should probably address that aswell. [23:20] jetole: I'd say it depends how deep you are expecting the person to be involved. [23:20] Daviey: right, but the way how arm is handle is quite different from the way a normal ISO is handled [23:20] jetole: if it requires extensive knowledge, i'd say engineer.. if it's keeping it ticking over, it's an admin :) [23:20] jetole: I've seen similar jobs by "High Availability X,Y,Z" [23:21] Daviey: I meant a more descriptive, less broad title then systems administrator or systems engineer [23:21] ah [23:21] RoAkSoAx: High Availability Manager? [23:21] jetole: High Availability Engineer? [23:21] ha deployment engineer? [23:21] RoAkSoAx: Perhaps we should get importing working, then work out how to enlist enable it? [23:21] RoAkSoAx: ok [23:22] Daviey: importing is the least of the issues really [23:22] I'm preparing to write the help wanted ad and want a title that will scare off the people I would waste time interviewing [23:22] Linux Cluster Specialist? [23:23] jetole: that too [23:23] ok. Thanks. I was hoping you guys all knew some real professional sounding title already [23:23] lol [23:24] Daviey: i personally think we need to get this working first with i386/x86_64 and then hanlde arm [23:24] RoAkSoAx: ok [23:24] jetole: either linux cluster engineer or HA engineer === erichammond1 is now known as erichammond [23:36] * Daviey goes afk === erichammond1 is now known as erichammond