[00:36] <nOStahl> hi guys, i just installed dovecot-postfix from apt
[00:36] <nOStahl> looking for documentation on where to go from here
[00:36] <nOStahl> can't find anything online regarding the dovecot-postfix package
[00:41] <qman___> nOStahl, it's in the server guide
[00:41] <SpamapS> nOStahl: look in /usr/share/doc/dovecot-postfix
[00:41] <SpamapS> nOStahl: oh, and the ubuntu server guide too :)
[00:41] <nOStahl> ty
[00:41] <nOStahl> is it pre-configured for virtual hosts
[00:44] <qman___> if you're referring to virtual users, no
[00:44] <qman___> the default configuration is a mailbox for each local unix user, at the domain you configure during setup
[00:44] <nOStahl> ah ok.
[00:49] <nOStahl> there is nothing except change log in /usr/share/doc/dovecot-postfix
[02:31] <jMCg> Aaah... this is really starting to get old: http://dpaste.com/707037/ -- I think I've sold this problem some 6 or so times, but I  always forget what it was.
[02:34] <jMCg> (kvm stagnating at init-bottom)
[02:34] <adam_g> jMCg: does it eventually timeout and continue on?
[02:34] <jMCg> https://bugs.launchpad.net/ubuntu/+source/linux/+bug/818177 should be this bug, should be be solved already, but it doesn't seem like it.
[02:34] <uvirtbot`> Launchpad bug 818177 in udev "boot failures as /dev is not transferred to /root (because 'udevadm exit' times out waiting for a deadlocked worker)" [High,Fix released]
[02:34] <jMCg> adam_g: I'm too impatient to wait.
[02:34] <jMCg> But OTOH, I'm too tired, should goto bed and see what happens tomorrow.
[02:35] <jMCg> So, good night.
[02:35] <adam_g> jMCg: yeah, i was going to point you there. g'night
[02:35] <jMCg> o/~
[02:49] <tash> is there a default directory size limit in ubuntu server?
[03:00] <twb> Depends on the filesystem, but shouldn't be anything you'll notice unless it's FAT
[03:09] <qman___> yeah, it should be high enough that you won't hit it except in very unusual circumstances
[03:09] <qman___> full details for ext4: http://en.wikipedia.org/wiki/Ext4
[03:35] <erichammond> utlemming: You might consider hanging out on the (quiet) #ubuntu-cloud channel.  There are occasionally some EC2 AMI conversations there.
[06:25] <utlemming> erichammond: will do
[07:30] <uvirtbot`> New bug: #939290 in samba (main) "package samba 2:3.6.3-1ubuntu1 failed to install/upgrade: ErrorMessage: subprocess installed post-installation script returned error exit status 1" [Undecided,New] https://launchpad.net/bugs/939290
[07:30] <uvirtbot`> New bug: #939291 in keystone (universe) "package keystone 2012.1~e4~20120203.1574-0ubuntu3 failed to install/upgrade: subprocess installed post-installation script returned error exit status 1" [Undecided,New] https://launchpad.net/bugs/939291
[07:51] <Canadian1296> Okay, I just installed ubuntu server (yes, this question belongs here.) and I need to connect it to my wifi network to update the packages. It is a WPA network, and I can't seem to connect. I brought down eth0, and brought up wlan0. I read in a guide I needed to install wpaservices. As I have no connection I got the .deb on a memory stick, but I cant install it since it's dependancies need to be updated as well.
[07:53] <twb> Canadian1296: one moment
[07:53] <Canadian1296> twb: okay thanks :)
[07:53] <twb> http://paste.debian.net/157275/
[07:54] <twb> Note that you cannot perform the initial netboot/d-i install over a WPA network; WEP is OK, WPA is still in the works (at least as at 10.04)
[07:56] <twb> wpaservices is not a valid package as at Wheezy or Oneiric.
[07:56] <Canadian1296> twb: I have the wpasupplicant (sorry I used the wrong name in the first one) deb on a memory stick as I have no Internet connection to use apt-get, but wpasupplicant has unmet dependancies that I also can't get.
[07:57] <twb> Then you need to get them
[07:57] <twb> Either borrow an install CD or do apt-get install --print-uris wpasupplicant to find out what packages to install
[07:57] <Canadian1296> And yes I used the wrong name before. I'm running natty and it was actually wpasupplicant. Would the easiest way be to set up a unprotected network just to upgrade, then switch to WPA?
[07:58] <twb> The latter assumes you have a working package list already; you might not
[07:58] <Canadian1296> twb: the installation is brand new and has never had a net connection.
[07:58] <twb> Canadian1296: probably yes.  Easiest would be to just get a 20m ethernet cable and plug it in for a while
[07:58] <Canadian1296> Alright, I'll try that then. Thanks for your help :)
[08:00] <Canadian1296> I'm going to want a wired connection once I start using the server anyways, I was just going to use wifi so I can have it in a different room to set up.
[08:01] <twb> Canadian1296: what you should do is set it up enough to have sshd installed, then leave it and do the rest of the install remotely
[08:02] <twb> If you were still doing the install, that would be done by picking the "network-console" udeb when prompted in an expert (priority=low) install.  If you're already booting off the install, just apt-get install openssh-server.
[08:03] <Canadian1296> I already have the openssh server and lamp installed as I selected them during install.
[08:03] <Canadian1296> I just need a net connection to get things going
[08:40] <Canadian1296> twb: just thought I'd let you know I got it connected to my network, and everything is working now. Thanks :)
[09:18] <jamespage> morning all
[09:20] <jamespage> Daviey: is there a project for people to report bugs in the openstack packages provided in PPA's by upstream?
[09:20] <jamespage> I have one in the new queue and would like to direct it appropriately
[09:26] <mrevell> yo
[09:28] <Daviey> jamespage: 'no'
[09:35] <ohai___> Why should I use ubuntu server instead of debian?
[09:36] <greppy> ohai___: if you use the LTS release, you have a longer support cycle.
[09:36] <ohai___> greppy: longer than debian?
[09:36] <greppy> yes
[09:40] <a_ok> I have removed a non escential disk but did not remove the entry from fstab(as I will replace the disk tomorow). Why won't my system boot?
[09:50] <TeTeT> a_ok: it waits for the device to become available
[10:03] <lynxman> morning o/
[10:06] <uvirtbot`> New bug: #939361 in openjdk-7 (multiverse) "ec2-get-console-output fails with xfire native error message with openjdk-7" [Undecided,New] https://launchpad.net/bugs/939361
[10:22] <diogo_79> hi
[10:27] <diogo_79> i am having some trouble with my portuguese keyboard the alt gr key doest work in ubuntu server 11.10 64 bits dont know what to try next?
[10:40] <jMCg> Now, when did I goto bed? 3? Left the vm running, and it still didn't boot..
[10:40] <jMCg> 3:40.
[10:53] <rbasak> jamespage: wondering about pre-existing conflicts/replaces for upgrade paths for a different previous binary package arrangement for openmpi. I think I should leave the conflicts in place, but not sure about the replaces.
[10:54] <jamespage> its probably OK
[10:54] <jamespage> as it is
[10:54] <jamespage> give it a try
[11:12] <jMCg> Oh, cool. There's actual developers here.
[11:13] <jMCg> Is there any way in launchpad to see the bugs I've participated in?
[11:15] <jMCg> Oh, none obviously.
[11:16] <jMCg> So, anyway, I'm "suffering" once again, of this https://bugs.launchpad.net/ubuntu/+source/linux/+bug/818177 -- and wondering if it's gonna be released anytime soon.
[11:16] <uvirtbot`> Launchpad bug 818177 in udev "boot failures as /dev is not transferred to /root (because 'udevadm exit' times out waiting for a deadlocked worker)" [High,Fix released]
[11:17] <jMCg> ho-hum. It should be released for about a month now.
[11:36] <mealstrom> hi, how to start mdadm before nfs ?
[11:37] <mealstrom> im trying to export directory from softraid, but nfs starts before raid partition md0 is mounted
[11:37] <mealstrom> i need to start mdadm then nfs, how can i do this
[11:38] <jMCg> pretty sure that works for me.
[11:54] <mealstrom> what is right syntax to change priority in update-rc.d ?
[11:55] <mealstrom> to set S15 and K90 for some init.d process/
[11:55] <mealstrom> ?
[11:55] <melvincv> A doubt on Ubuntu's ssh client: My internet connection is unstable, so my ssh client shows an error and disconnects my session often. Error: "Write failed: Broken pipe" Any solutions?
[11:56] <mealstrom> ok iv got it, just add -f parameter
[12:04] <melvincv> Using screen is of any benefit?
[12:05] <mealstrom> melvincv: yes
[12:05] <melvincv> I need a how to
[12:05] <mealstrom> melvincv: you can run program via ssh from one PC and then after you have connected to this pc from another one -- resume screen
[12:05] <mealstrom> melvincv:  type screen
[12:06] <melvincv> ok
[12:06] <mealstrom> melvincv: then run some program like top, after this press 'ctrl+a' then 'd'  to  detach
[12:07] <mealstrom> melvincv: then  run "screen -r" to resume
[12:07] <melvincv> But will my ssh session disconnect if the Internet connection is unstable, even with screen?
[12:07] <mealstrom> screen will  detach
[12:07] <mealstrom> and after you reconnect you can continue to work
[12:07] <melvincv> ok, let me try, thanks mealstrom
[12:08] <mealstrom> you are welcome
[12:14] <melvincv> ok, got it!
[12:16] <greppy> melvincv: another thing to look into is tmux, a search for 'gnu screen tutorial' on google may provide some good info.
[12:18] <linocisco> hi
[12:19] <linocisco> i am almost done installing ubuntu server. It said GRUB is installed on bootloader or Not ? If I say NO, what would be?
[12:19] <linocisco> i am almost done installing ubuntu server. It said GRUB is installed on bootloader or Not ? If I say NO, what would be?
[12:22] <linocisco> how to configure console font using dpkg?
[12:30] <linocisco> hi
[12:36] <linocisco> hi all
[12:37] <linocisco> does "http://www.debuntu.org/2006/02/23/8-using-multiple-network-device-to-connect-to-the-internet" work for failover?
[12:40] <linocisco> hi
[12:40] <linocisco> all inactive here?
[12:44] <TeTeT> hi, but no experience with multi nic failover setups, sorry
[12:45] <TeTeT> if you say no to grub loader install, your system will most likely stop after the bios tries to launch a boot loader from disk
[12:46] <TeTeT> linocisco: sudo dpkg-reconfigure console-setup might help with reconfiguring the font
[12:47] <linocisco> thanks
[12:49] <linocisco> how to disable or stop sleeping monitor on ubuntu server?
[12:50] <linocisco> i dont want server's monitor to go into sleep mode after a period of time
[12:55] <TeTeT> linocisco: check if the dpms module is loaded and remove it if so
[12:55] <linocisco> how to check and how to remove perminently or temporarily?
[12:59] <jMCg> Anyone know if this fix has/will be released to oneiric? https://bugs.launchpad.net/ubuntu/+source/linux/+bug/818177 -- I can't build the preceise package.
[12:59] <uvirtbot`> Launchpad bug 818177 in udev "boot failures as /dev is not transferred to /root (because 'udevadm exit' times out waiting for a deadlocked worker)" [High,Fix released]
[12:59] <mdeslaur> SpamapS: if I want to test mysql upgrades, what would be the simplest lamp package in the archive to set up?
[13:01] <linocisco> TeTeT, how to??
[13:13] <med_> rbasak, I took a look at the Tilera approach last night. It's both a heterogeneous approach and a "no virtualization" approach to OpenStack on non x86 hardware. See this:
[13:13] <TeTeT> linocisco: lsmod | grep dpms
[13:13] <med_> http://wiki.openstack.org/HeterogeneousTileraSupport  and there are a number of other supporting blueprints/wiki pages.
[13:13] <med_> moreover, that code landed here:
[13:13] <TeTeT> hmm, there's no such module, bummer
[13:14] <med_>  nova/nova/virt/baremetal.
[13:14] <rbasak> med_: thanks - reading
[13:14] <TeTeT> linocisco: if there's no X running, I don't know, sorry
[13:15] <linocisco> TeTeT, I found no response running this command
[13:15] <TeTeT> linocisco: that's correct
[13:15] <linocisco> How to setup squid proxy server with Two Internet lines of different ISP?
[13:16] <TeTeT> linocisco: maybe vbetool can help, but i've never used it
[13:17] <linocisco> vbetool?
[13:17] <TeTeT> linocisco: help with the display, not the proxy
[13:17] <linocisco> ok
[13:17] <rbasak> med_: this looks great - the perfect integration point
[13:18] <rbasak> med_: it sounds like all the fundamental architectural changes are all done then, and an arm module just needs to be added. This method won't use LXC at all, then?
[13:18] <rbasak> zul: ^^
[13:19] <linocisco> How to setup squid proxy server with Two Internet lines of different ISP to failover?
[13:20] <rbasak> jamespage: I think I need to remove the Replaces lines. The Conflicts/Replaces presumably exist for an upgrade path. In the case that somebody is upgrading but requested libopenmpi1.5, I don't think we should consider that an upgrade path, and it should just break on the conflicts. So I think I should remove all the Replaces lines, but keep all the Conflicts, and add any Conflicts that aren't already there from the Replaces I'm removing. Does this
[13:20] <rbasak>  seem sane?
[13:41] <med_> rbasak, it would/could still use LXC. It's a solution for Redstone/Moonshot systems. It is still perfectly valid for ARM Server to be pure OpenStack with LXC.
[13:41] <rbasak> med_: so you're thinking of one nova-compute proxy per node, each of which deploy multiple LXC "bare metal" instances?
[13:42] <med_> truthfully I'm still noodling the idea.
[13:42] <med_> but a nova-compute proxy, yes.
[13:43] <rbasak> As an alternative, how about one nova-compute controller that deploys true bare metal non-LXC instances?
[13:44] <zul> yes please :)
[13:46] <rbasak> I think the integration point would be the same. Reasons are that I'm not sure about how to manage instances that don't want to share on LXC, that security between LXC instances and LXC-to-host is weak, and potential instance breakage because an LXC environment isn't quite perfect yet.
[13:46] <rbasak> Not that I have anything against LXC, but I'm not sure we need it in this case.
[13:46] <rbasak> I've been working on automated ARM metal deployment and I have that sussed.
[13:47] <Daviey> zul: What is the status of keystone redux, and in particular the MIR?
[13:47] <rbasak> So I suppose it's fewer unknowns for me.
[13:47] <zul> Daviey: im waiting for it to work, im hoping to throw up a new keystone tomorrow so the security guys can have a look
[13:48] <rbasak> One fundamental limitation though is that you're limited to the number of nodes. This may lead to under-utilisation which is one of the benefits of Openstack. I suppose it depends on expected workload and use cases.
[13:48] <zul> Daviey: what is in the archive right now is not redux
[13:49] <zul> Daviey: im hoping to get some movement on it next week
[13:52] <Daviey> zul: right, thanks
[13:52] <Daviey> zul: did you drop DEP-5 headers from the console patch in nova?
[13:52] <zul> Daviey: possibly
[13:53] <zul> i was trying to get ci moving again
[13:53] <Daviey> zul: cool.. fancy re-adding ? :)
[13:53] <Daviey> zul: also, what is a thinko?
[13:53] <zul> Daviey: thinko = stupid mistake
[13:53] <zul> Daviey: i guess i could re-add :)
[13:54] <linocisco> hi
[13:54] <linocisco> how to install squid proxy with two different ISP lines?
[13:56] <zul> Daviey: ill try to limit canadianisms from my changelog entries :0
[13:57] <Lazerath> can i resize the partition on here with that partition unmounted
[13:58] <Lazerath> with the server lice
[13:58] <Lazerath> with the server live
[13:58] <gary_poster> hallyn, do I remember correctly that lxc does not support nested containers at this time?
[13:58] <koolhead17> hi all
[13:59] <gary_poster> (and hi btw :-)
[13:59] <Daviey> zul: heh
[14:00] <Lazerath> can i resize the partition on here with that partition unmounted but with the server live... so that it ends on a cynderlic boundry>
[14:06] <jamespage> rbasak, sounds reasonable - I would probably still try a few scenarios to test tho
[14:06] <jamespage> (happy todo that BTW)
[14:11] <koolhead17> zul: any workaround for this  https://bugs.launchpad.net/horizon/+bug/934064
[14:11] <uvirtbot`> Launchpad bug 934064 in horizon "Installing openstack-dashboard on Precise removes Keystone package" [High,Confirmed]
[14:12] <koolhead17> Ursinha: hey
[14:12] <koolhead17> hi Daviey
[14:12] <Daviey> hey koolhead17
[14:12] <Ursinha> koolhead17, hello
[14:12] <koolhead17> zul: if you need a n00b helping hand i can help
[14:12] <zul> koolhead17: no it should get fixed when keystone gets fixed
[14:14] <koolhead17> zul: so. any idea how many more days/week i have to work?
[14:14] <zul> koolhead17: hopefully a new version of keystone will be uploaded tomorrow
[14:15] <koolhead17> zul: that is great news, so i can work on it from monday!! :)
[14:15] <zul> *hopefully*
[14:17]  * koolhead17 has his fingers crossed!! :)
[14:22] <linocisco> hi
[14:22] <linocisco> how to failover in proxy server with two different ISP link?
[14:36] <zul> Daviey: for keystone dbconfig-common stuff this is what im thinking so far: http://paste.ubuntu.com/854047/
[14:36] <koolhead17> uksysadmin: around?
[14:37] <uksysadmin> here, sir.
[14:37] <uksysadmin> ah, keystone and dbconfig....
[14:38] <koolhead17> uksysadmin: so whole keystone is re-writtten and thats why the bug.
[14:38] <Daviey> zul: does it work?
[14:38] <zul> Daviey: not tested yet..
[14:39] <uksysadmin> sounds like things are moving on the keystone front
[14:39] <koolhead17> as buddha said patience gets you success!! uksysadmin ^^
[14:40]  * uksysadmin is doomed for failure
[14:40] <uksysadmin> ;-)
[14:41] <hallyn> gary_poster: on precise, nested containers should work fine
[14:42] <gary_poster> hallyn, wow, ok, we'll give it a try.  Thanks!
[14:42] <Cryp71c> SpamapS, I'm back from yesterday...got SSH opened up so, yay.
[14:42] <hallyn> gary_poster: I fear that the apparmor mount restrictions may complicate that, but I'll just have to think of a way around that
[14:43] <hallyn> hm, desktop not behaving.  think i need an update+reboot.  bbl <he hopes>
[14:43] <gary_poster> hallyn, heh, ok.
[14:45] <hallyn> jjohansen: I don't want to bug you, but for the sake of having answers when asked, when do you expect (a) the apparmor lxc fix and (b) the apparmor mount stuff to hit the archive?
[14:46] <zul> Daviey: it should work
[14:51] <rbasak> jamespage: https://launchpad.net/~racb/+archive/experimental/+files/openmpi1.5_1.5.4-0ubuntu1.dsc
[14:51] <jamespage> rbasak, it would be good to get openmpi1.5 FFe approved and uploaded before beta freeze tonight
[14:51] <jamespage> bah
[14:52] <rbasak> jamespage: I've test built it, it's not built on the PPA yet
[14:52] <rbasak> (just uploaded)
[14:52] <rbasak> jamespage: I've also just tried a basic fresh install compile and hello world. I'll do some upgrade scenarios that occur to me.
[14:53] <jamespage> rbasak, ack - I'll put it through some testing as well
[14:53] <jamespage> then use that for the FFe
[14:53] <rbasak> I hope it's OK! I've not done this type of thing before. I used a new orig.tar.gz that's the same as the old one but it doesn't know that so had to upload it.
[14:54] <rbasak> And I went through the debian/ directory to try and make sure I got everything. Hope I haven't missed anything!
[14:55] <jamespage> rbasak, I'll review now
[15:04] <rbasak> I was pleased to discover that a debdiff between two "entirely different" packages does in fact work :)
[15:05] <jamespage> rbasak, made one minor tweak - folded the changelog entries for the last updates we did and this new branch.
[15:06] <[1]maarezende> hello everyone
[15:06] <rbasak> jamespage: OK
[15:07] <Cryp71c> I'm just learning to manage an ubuntu-server from ssh (trying to get to a point where I can turn x off entirely and feel confident in not messing anything up). I'm running a couple of (small) minecraft servers which run in their own terminal, how can I run each server so that I can return and type commands into each server's terminal (like, stop, to stop the server cleanly) ?
[15:07] <[1]maarezende> im having networking troublle with 10.04, iv migrated the server from a vmware server to anoter, now i cant get the network up. Iv already followed this steps: http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=1032790
[15:07] <jacobw> hi
[15:07] <jacobw> pam-auth-update
[15:07] <[1]maarezende> but went to something similar to this bug - https://bugs.launchpad.net/ubuntu/+source/sysvinit/+bug/440179
[15:07] <uvirtbot`> Launchpad bug 440179 in sysvinit "service fails to start/stop/restart networking daemon" [Medium,Fix released]
[15:08] <jacobw> why does it prime the stack for a positive response and how i can stop it from doing this?
[15:08] <[1]maarezende> iv tried many things till now, deleted the 70-persistent file, changed the interfaces file to an abitrary eth
[15:17] <jcastro> server users! I've just launched our first ever juju charm contest! http://cloud.ubuntu.com/2012/02/juju-charm-contest-help-bring-free-software-into-the-cloud/
[15:22]  * koolhead17 claps jcastro ^^
[15:22] <jcastro> :)
[15:23] <koolhead17> have you tweeted it as well
[15:24] <jcastro> yep
[15:24] <koolhead17> cool!!
[15:26] <rbasak> Cryp71c: sounds like you want byobu
[15:26] <Cryp71c> rbasak, byobu?
[15:26] <Cryp71c> sorry, misunderstood what you were saying.
[15:27] <kirkland`> Cryp71c: SpamapS suggested the same thing to you yesterday :-)
[15:27] <Cryp71c> rbasak, I've had others suggest something like screen (or, the more modern tmux)
[15:27] <rbasak> Cryp71c: https://launchpad.net/byobu
[15:27] <Cryp71c> kirkland`, sorry, must have missed that.
[15:27] <rbasak> Cryp71c: yes, you can use screen directly if you wish, or tmux. byobu is a nice front-end to it.
[15:28] <kirkland> rbasak: how's it working on ARM these days?
[15:28] <rbasak> kirkland: good, thanks! Lots of different aspects to it. I'm really enjoying it.
[15:28] <rbasak> kirkland: how's the new job?
[15:29] <kirkland> rbasak: having a lot of fun, actually!
[15:29] <rbasak> great!
[15:29] <kirkland> rbasak: we just got a panda board, btw... gonna start playing with ecryptfs there
[15:29] <rbasak> Cool!
[15:30] <rbasak> Give me a shout if you have any panda install issues :)
[15:30] <kirkland> rbasak: ;-)  will do
[15:30] <kirkland> Cryp71c: http://bit.ly/byobu-help
[15:31] <rbasak> I was getting usb booting packaged and integrated with udev to make it fully automatic but I got distracted
[15:31] <kirkland> heh
[15:31] <rbasak> It will happen soon though
[15:34] <SpamapS> mdeslaur: wordpress or phpmyadmin
[15:34] <mdeslaur> SpamapS: cool, thanks
[15:38] <jamespage> rbasak, looks OK - I want to make one other change - libopenmpi1.5 libopenmpi1.5-2 inline with the ABI bump versioning that Debian are using
[15:39] <rbasak> jamespage: ok. I fail to understand what Debian are doing with that.
[15:45] <rbasak> jamespage: warning: I think there was a mention of libopenmpi2 in debian/rules which will need changing then
[15:45] <rbasak> (in addition to the debhelper files)
[15:46] <jamespage> rbasak, ah - nice spot had missed that
[15:47] <jamespage> rbasak, have you submitted that change to re-jig the modules back to debian yet?
[15:48] <rbasak> jamespage: no. We want to submit your upgrade path fix too, right? And those two fixes are combined in your PPA version now?
[15:48] <jamespage> rbasak, ah - yes I remember
[15:48] <jamespage> although I see that the upgrade issue has been reported in Debian
[15:49] <jamespage> my fix was not working that well
[15:50] <Cryp71c> kirkland, is byobu for screen / tmux, or is it its own program that doesn't rely on the others to be installed?
[15:51] <rbasak> Cryp71c: the former. I'm not familiar with byobu and tmux, but on byobu with screen, it uses screen but configured nicely
[15:52] <rbasak> Cryp71c: I think in your case  byobu will do what you want and you won't really need to learn much. With screen or tmux you'd have to spend some time learning and customising a workflow to suit your needs, but then they'll do what you want. byobu just uses screen and tmux to do that step for you
[16:00] <zul> Daviey/adam_g: apparently this is the reason for nova-compute crapping out: https://bugs.launchpad.net/nova/+bug/934575
[16:00] <uvirtbot`> Launchpad bug 934575 in nova "notifier endless loops in is_primitive" [Medium,Fix committed]
[16:01] <Cryp71c> rbasak, thanks, loving byobu
[16:02] <Cryp71c> kirkland, SpamapS, thanks for your earlier help and the original suggestion of byobu.
[16:02] <Daviey> zul: hah
[16:12] <jamespage> Daviey: are you OK to review the openmpi1.5 FFe?
[16:14] <Daviey> jamespage: myself or infinity.
[16:31] <jamespage> rbasak, openmpi1.5 uploading now to NEW - just got an ack in -release
[16:33] <gondoi> i'm trying to build a correctly structured repo for some software, but i'm stock on the <release>-updates part.. what is the best channel to get help on this?
[16:34] <jamespage> rbasak, thanks for all of your work on openmpi - its been hard work but interesting!
[16:36] <gondoi> s/stock/stuck
[16:38] <rbasak> jamespage: no, thank you!
[16:38] <SpamapS> gondoi: you may just want to use reprepro instead
[16:39] <SpamapS> gondoi: for running your own apt repo.. its really the simplest way
[16:39] <gondoi> yeah that's what I'm using
[16:39] <gondoi> but here is the scenario
[16:39] <gondoi> i use unattended-upgrades
[16:40] <gondoi> and I want the updates to pull from <release>-updates, but i've currently only have serveragent setup..so my packages aren't getting auto updated
[16:40] <SpamapS> gondoi: I don't know anything about unattended-upgrades .. but I'm sure somebody here will
[16:41] <gondoi> I've got an idea of how to get it working... but I just want to make sure I'm doing it right
[16:42] <rbasak> The other day I found what is IMHO a simpler way. I dump the debs in a directory and then run "apt-ftparchive packages .|tee Packages|gzip>Packages.gz; apt-ftparchive release . > Release; gpg --yes -o Release.gpg -ab Release". I prefer it because the only state I have to maintain is the debs in the directory, whereas with reprepro I have to worry about the state it has in its database, write a config file and so on.
[16:43] <SpamapS> rbasak: thats "old school" :)
[16:43] <rbasak> SpamapS: it's fundamentally simpler :)
[16:44] <SpamapS> rbasak: old school is almost always simpler.. and preferrable. Yours is also idempoetent.
[16:44] <rbasak> Yeah, exactly!
[16:44] <rbasak> Though of course reprepro also has its uses.
[16:44] <caribou> would anybody have an idea to suggest on where to look when 'lxc-create' fails on trying to do 'apt-get install lxcguest', not being able to resolve the archives.ubuntu.com ???
[16:44] <SpamapS> I believe reprepro can be setup in a similar way... but I have not played with it much.
[16:44] <caribou> I'm in London and my setup, that works fine @home is not able to create new containerzs
[16:45] <SpamapS> rbasak: reprepro is good for prototyping and casual "oo I need that package too" work
[16:45] <caribou> FYI, exising lxc containers do work fine with the current network setup
[16:45] <rbasak> SpamapS: that's what I've been using my way for - eg. rebuild testing where packages should build-depend on the local copy of the package I've just built
[16:45] <SpamapS> rbasak: you really should turn that command you just mentioned into a single command and put it in apt-ftparchive so everyone can benefit. :)
[16:46]  * SpamapS notices that its 15 min till the exterminators arrive and he hasn't even moved the turtles outside yet. DOH
[16:46] <rbasak> SpamapS: http://paste.ubuntu.com/854222/ :-)
[16:47] <rbasak> SpamapS: and for sbuild, I create ~/repo/add as follows: http://paste.ubuntu.com/854224/. Then sbuild --chroot-setup=~/repo/add does the right thing
[16:47] <rbasak> err, --chroot-setup-commands
[17:33] <Cryp71c> I have an ssh connection to a server that's also running X. I'd like to turn off X without destroying my ssh connection, is this possible?
[17:37] <raubvogel> Cryp71c: X has nothing to do with ssh, so you should be able to turn it off
[17:37] <SpamapS> Cryp71c: your ssh connection to the other machine is not dependent on X
[17:37] <Cryp71c> "/etc/init.d/gdm stop" to stop X?
[17:38] <raubvogel> That should stop your window manager
[17:42] <Cryp71c> raubvogel, ...I'm a bit confused, WM is not X, is it? (though I think X is the graphical environment where your WM runs?)
[17:50] <JanC> Cryp71c: stopping the dm (display manager, e.g. gdm or lightdm or kdm or ...) should also stop X
[17:56] <Cryp71c> JanC, thanks, indeed it did
[17:57] <WALoeIII> I've got lucid server images with console-kit-daemon --no-daemon taking up a large amount of memory, seems like this would not be used on a server, can I remove the consolekit package?
[18:03] <Cryp71c> In htop I have a ton of processes that are all related to one another (threads of a parent process, perhaps)? is there a flag to consolidate some of those?
[18:06] <caribou> :2
[18:06] <JanC> Cryp71c: at the bottom left it says "F1 help"  ;-)
[18:07] <JanC> which shows you the keys to hide threads
[18:08] <JanC> or colapse parts of a tree (if you use that view)
[18:12] <Cryp71c> ah! :)
[18:14] <ericthesysadmin> I am trying to get ubuntu to connect to an LDAPS AD server and its not working, I have read that I need the security certificate from the windows server installed on the ubuntu server
[18:14] <ericthesysadmin> basically im trying to use php-ldap to authenticate a script to the ldaps server
[18:23] <Cryp71c> JanC, I see tree view, that helps, but I don't see how to collapse children of a tree entry.
[18:24] <JanC> Cryp71c: might depend on the version, but here it says you can use + & - for that
[18:25] <JanC> but to show/hide threads, it's just "H"
[18:28] <Cryp71c> H works, thanks.
[19:17] <SockPants> hello all
[19:18] <SockPants> i'm sick and tired of having to deal with different systems all the time for development, so i'd like to create a well-configured desktop environment on my server that I can connect to securely from any location over internet
[19:19] <SockPants> but my experience with vnc over ssh over internet and such aren't very good, the response time is usually unusable even with small distances where both parties are connected by a decent connection
[19:19] <SockPants> could that have been due to bad configuration / setup choices or is that generally the case? if so what should I do?
[19:20] <adam_g> zul: i think that creation of keystones admin users/tenant/role/etc shouldn't happen in postinst
[19:20] <Myrtti> I'd probably not add vnc in the equasion at all and just use X forwarding over ssh
[19:20] <Myrtti> but YMMV
[19:20] <zul> adam_g: i was looking at this morning and I happen to agree
[19:20] <adam_g> zul: or in packaging at all, it should be on the user to decide that stuff. and its not going to work from postinst without environment variables specific to (sometimes existing) configuration
[19:21] <zul> adam_g: agreed
[19:21] <rbasak> SockPants: you might look into NX. It's commercial but there's a free version too. I haven't tried it recently, but it was leagues faster than VNC the last time I looked
[19:21] <zul> keystone was breaking because of it when i was installing it this morning
[19:21] <SockPants> rbasak: is it secure enough to use  plain or would I still need ssh tunneling?
[19:22] <Cryp71c> SockPants, we used NX at our uni to standardize file locality and development environments...when working in the lab it was like working on a local system (though this was a local network environment)
[19:23] <rbasak> SockPants: it uses ssh. What I don't like about it is its expectation of taking over your machine with an "nx" account to mediate the connection, weakening the security. But nowadays you can replace with your own keypair so it's not so bad. I know that freenx were working on a way to eliminate the necessity for an nx account entirely, but I don't know if that's usable yet.
[19:23] <zul> adam_g: are we any closer to updating keystone in the archive yet?
[19:23] <SockPants> Cryp71c: sounds good. I -should- have a similar environment most of the time since my home is connected to ftth and the server is at uni on fiber not 10 km away, but so far it's not entirely as you would expect
[19:24] <zul> adam_g: i want the MIR process to get going again
[19:24] <SockPants> rbasak: i'm looking into it, thanks
[19:24] <rbasak> SockPants: they once used to provide test accounts where you could log in to their gnome or kde based server directly to see what it was like. I used it over a congested adsl link and it looked good. I don't know whether they still let you trial it that way or not, but the demonstration was very impressive.
[19:26] <SockPants> well it's good to hear of promising new software in remote desktoping because vnc was old and slow
[19:28] <hallyn> stgraber: RUUUDE!  I can get a fedora 16 container booted such that I can ssh in, but looking at the systemd source, there is simply no way to stop it from mounting devtmpfs over /dev (apart from apparmor forbidding it :)
[19:28] <hallyn> so, I'm leaving that be for now.  I'll give it one more shot with the apparmor rules
[19:28] <adam_g> zul: e4 is tuesday. i was told thats the feature freeze for keystone, can we wait until then? im trying to get this SQL catalog stuff proposed and merged before then
[19:29] <zul> adam_g: yeah i think so
[19:30] <stgraber> hallyn: and then systemd will explode ;)
[19:32] <SockPants> what's the best way to start setting up freenx on my server? it has no GUI yet so it would need to virtualize one, or I could just install it, or I could virtualize an entire OS with vbox
[19:32] <hallyn> stgraber: haha, fortunately, even though the caller of 'mount_one' checks its return value, *it's* caller ignores it :)
[19:32] <hallyn> so it should go fine
[19:32] <hallyn> (until they "fix that bug")
[19:32] <stgraber> ;)
[19:33] <hallyn> (src/selinux-setup.c and src/mount-setup.c in the systemd source...)
[19:33] <hallyn> all right my eyes hurt.  quick break, biab
[19:35] <rbasak> SockPants: personally, I'd virtualise an entire OS
[19:45] <Klavier> is there anyone who use 11.10 - orchestra (dns+dhcp)
[19:47] <adam_g> Klavier: yea
[19:48] <pmatulis> what's the reliability factor these days for shrinking an ext3 or ext4 filesystem?
[19:48] <patdk-wk> never had an issue
[19:49] <patdk-wk> never had an issue two years ago either :)
[19:49] <Klavier> adam_g, when i installing orchestra, i said yes  for DNS&dhcp installation
[19:49] <Klavier> but DHCP is not working correctly for PXE boot.. ,, PXE boot property not worked for node
[19:50] <Klavier> i added mac address correctly with cobbler
[19:51]  * roaksoax lunch
[19:51] <uvirtbot`> New bug: #939765 in lxc (universe) "lxc-clone should duplicate original size for lvm devices" [Medium,In progress] https://launchpad.net/bugs/939765
[19:52] <Klavier> adam_g,  logs: http://pastebin.com/4zWQ9dT3   can you help please,
[19:55] <adam_g> Klavier: can you paste output of 'cobbler system dumpvars --name=oneiric01.ubuntu.lan' ?
[19:56] <Klavier> iam at home now, system is up at office, :((
[19:57] <Klavier> i can send it tomorrow, with "cobbler system dumpvars --name="oneiric01.ubuntu.lan",, which vars do you want to check, why?
[19:59] <Canadian1296> Can anyone here help with an apache configuration problem? I'm only asking because no one is responding in #httpd
[20:00] <Klavier> Canadian1296, ask
[20:00] <Klavier> adam_g, how can i check DHCP daemon of orchestra running correctly or not?
[20:00] <adam_g> Klavier: you need to make sure the system is tagged as 'netboot enabled'
[20:02] <Klavier> adam_g, when iam installating system i install "only" openssh-server, than: apt-get update; apt-get dist-upgrade -y; apt-get install ubuntu-orchestra-server
[20:02] <Klavier> it is virtual enviroment
[20:02] <ch33z> mk
[20:02] <ch33z> I need help
[20:03] <Klavier> iam not installing any extra dhcpd or dns app,, iam just installating orchestra,, i will check 'netbook enabled' property,, in which configration file..? i want to be sure if orchestra dhcpd is running correctly or not,, how can i do that?
[20:05] <danp> any particular reason (besides possible HTTP pipelining issues) that the ubuntu mirrors meant for ec2 aren't on S3?
[20:06] <adam_g> Klavier: its handled by dnsmasq, which is installed as part of the orchestra package
[20:06] <adam_g> Klavier: if you're running this in virtual machines, also make sure kvm-pxe is insatlled on the host
[20:07] <Klavier> adam_g, it s default installation,,, when you installed orchestra, did you manually change 'netboot enabled' ??
[20:08] <Klavier> or did you manually enable netboot with cobbler_web?
[20:09] <Canadian1296> Klavier: I just figured it out. Thanks anyways
[20:09] <adam_g> Klavier: when adding the system into cobbler, you should also mark it netboot enabled.
[20:12] <Klavier> adam_g,  last question: when i run: cobbler check ->>  One or more repos need to be processed by cobbler reposync for the first time before kickstarting against them ...
[20:12] <Klavier> how can i fix it? i tried to disable cobbler repo "update" properties for each subject, but it didnt work
[20:13] <Klavier> i didnt understand why i see these, when i run: cobbler check
[20:16] <Klavier> :  -  ((
[20:18] <adam_g> Klavier: im not sure of that one, i dont think its anything to worry about
[20:31] <Klavier> adam_g, thank you very much for answers
[20:31] <Klavier> i dont know if there is a maillist or something for orchestra users questions
[20:31] <Klavier> so i asked here
[20:36] <uvirtbot`> New bug: #939795 in lxc (universe) "package lxc (not installed) failed to install/upgrade: subprocess installed post-installation script returned error exit status 1" [Undecided,New] https://launchpad.net/bugs/939795
[20:38] <uvirtbot`> New bug: #939774 in lxc (universe) "dnsmasq should run as lxc-dnsmasq user" [High,Confirmed] https://launchpad.net/bugs/939774
[20:40] <adam_g> Klavier: ubuntu-server-discuss is as good a list as any
[21:22] <jMCg> Still no volunteer to debug https://bugs.launchpad.net/ubuntu/+source/linux/+bug/818177 with me?
[21:22] <uvirtbot`> Launchpad bug 818177 in udev "boot failures as /dev is not transferred to /root (because 'udevadm exit' times out waiting for a deadlocked worker)" [High,Fix released]
[21:56] <SpamapS> kirkland: FYI, doing ctrl-A: set aggressive-resize off does not actually resize the viewport for me
[21:58] <kirkland> SpamapS: looking ...
[22:04] <kirkland> SpamapS: well, I have an answer, but I don't think you're gonna like it....
[22:06] <SpamapS> haha
[22:08] <hallyn> stgraber: I'm not thinking right about something.  If I want to have dnsmasq (from dnsmasq pkg) serve lxcbr0, by using a /etc/dnsmasq.d/lxc file for lxcbr0.  What do you think is the best way to handle that while accounting for LXC_BRIDGE beign configurable?
[22:09] <hallyn> just write the file at boot, unconditionally, every time in a job that starts on starting dnsmasq?
[22:10] <stgraber> pulling the full dnsmasq package (instead of just -base) will at least break Edubuntu and any LTSP system, so please don't
[22:10] <kirkland> SpamapS: tmux attach -d
[22:10] <kirkland> SpamapS: or, byobu attach -d
[22:10] <kirkland> SpamapS: basically, you have to detach all of your other clients
[22:10] <kirkland> SpamapS: looks like tmux forces the size to the smallest attached client
[22:10] <kirkland> SpamapS: http://stackoverflow.com/questions/7814612/is-there-any-way-to-redraw-tmux-window-when-switching-smaller-monitor-to-bigger
[22:11] <kirkland> SpamapS: it's a common enough request;  I'll try to make a keybinding for it
[22:11] <treeder> hey utlemming
[22:11] <treeder> travis from the G+ convo
[22:12] <SpamapS> kirkland: thats ok, thats what I used to do with screen, but it didn't work anymore...
[22:12]  * SpamapS tries again
[22:12] <utlemming> hey treeder
[22:12] <kirkland> SpamapS: well I just tested here and it works with tmux
[22:12] <hallyn> stgraber: d'oh.
[22:13] <SpamapS> kirkland: ahh, I didn't have the 'attach' command
[22:13] <SpamapS> kirkland: this is good, thanks. :)
[22:13] <kirkland> SpamapS: you bet
[22:14] <kirkland> SpamapS: would you mind filing a bug against ubuntu/byobu to track this?
[22:14] <kirkland> SpamapS: so that I can get it into precise?
[22:14] <treeder> hi all, I'm having some serious issues with apt repositories on ec2
[22:14] <treeder> Hoping someone can help us out
[22:14] <utlemming> treeder: what package are you trying to install?
[22:14] <treeder> apt-get update itself is one of the worst problems
[22:15] <hallyn> stgraber: i thought that based on comments in bug 928524 and bug 925511 that was actually the best way.
[22:15] <treeder> Then trying to install git is another big problem
[22:15] <Daviey> treeder: can you "sudo apt-get update | pastebinit" please?
[22:15] <uvirtbot`> Launchpad bug 928524 in lxc "lxcbr0 fails to come up when dnsmasq is installed" [Low,Confirmed] https://launchpad.net/bugs/928524
[22:15] <uvirtbot`> Launchpad bug 925511 in lxc "lxc init script should fail when it ... failed" [High,Fix released] https://launchpad.net/bugs/925511
[22:15] <utlemming> treeder: what zone are you in?
[22:15] <SpamapS> kirkland: sure. To be fair.. this is only an issue for me on my CentOS test box. ;)
[22:15] <kirkland> SpamapS: I've gotten the question a few times before though
[22:16] <hallyn> stgraber: i guess i punt on that for this release then.  Do you have any objections to making run as a new user called 'lxc-dnsmasq' ?
[22:16] <treeder> us-east
[22:16] <treeder> first off, here's our sources.list: http://pastebin.com/SX7qwS2M
[22:16] <utlemming> treeder: us-east-1{a,b,c,d}
[22:17] <stgraber> hallyn: what would be advantage of using lxc-dnsmasq instead of nobody?
[22:18] <hallyn> stgraber: sbeattie asked for it (for libvirt) with rationale in bug 938255
[22:18] <uvirtbot`> Launchpad bug 938255 in libvirt "libvirt runs dnsmasq as user nobody" [High,Confirmed] https://launchpad.net/bugs/938255
[22:18] <treeder> apt-get update error: http://pastebin.com/MUJRpJUs
[22:18] <treeder> us-east-1c
[22:18] <SpamapS> kirkland: bug 939870 file
[22:19] <uvirtbot`> Launchpad bug 939870 in byobu "With byobu 5.7 and tmux 1.5 cannot resize window without detacching all others" [Undecided,New] https://launchpad.net/bugs/939870
[22:19] <SpamapS> d
[22:19] <utlemming> treeder: what does your /etc/hosts look like?
[22:19] <treeder> i think our failure rate is nearly 100% now
[22:19] <SpamapS> we really need to fix the apt format so this can't happen
[22:20] <treeder> hosts: http://pastebin.com/4EPRKnHu
[22:22] <stgraber> hallyn: ok, if it's done for libvirt, I guess it makes sense to do it for lxc too
[22:22] <utlemming> treeder: how about the console output of "ec2-get-console-output <instance-id>"
[22:23] <hallyn> stgraber: one could argue it's less important for lxc, bc it's only exposed to contaienrs right?
[22:23] <hallyn> adn we don't trust containers yet anyway
[22:23] <hallyn> (whereas some fools do trust kvm :)
[22:24] <Canadian1296> What is the best way to allow authorized user to edit files on a server? (Like FTP, only secure).
[22:25] <Daviey> hallyn: i can't see that adding a user for lxc-dnsmasq is a feature...  why do you need to discuss it with the release team?
[22:25] <hallyn> Daviey: I wasn't going to
[22:26] <hallyn> question is more of just whether it should be done, or whether there are downsides i haven't considered
[22:27] <Daviey> hallyn: an unpriv'd user doesn't need to touch files owned by lxc-dnsmasq, does it?
[22:27] <utlemming> treeder: are these instances in the VPC?
[22:27] <hallyn> Daviey: no
[22:28] <Daviey> hallyn: unless stgraber or jjohansen raise objection, you should proceed :)
[22:29] <stgraber> hallyn: should be fine
[22:29] <treeder> one sec
[22:29] <hallyn> ok, thanks
[22:29] <stgraber> hallyn: though please note we're now frozen for beta1
[22:29] <Daviey> stgraber: not for universe we are not :)
[22:29] <hallyn> yup, i noticed Kate's email :)
[22:29] <hallyn> d'oh
[22:29] <stgraber> Daviey: we're for seeded universe
[22:29] <stgraber> Daviey: and lxc is seeded by Edubuntu
[22:30] <Daviey> stgraber: waaat.. since when?
[22:30] <treeder> and no, not VPC
[22:30] <stgraber> Daviey: since Edubuntu has been shipping Arkose and Arkose has been based on LXC, so since early Oneiric
[22:31] <Daviey> stgraber: ah, ok.
[22:31] <treeder> here's the console log: http://pastebin.com/ptrZwXT2
[22:33] <treeder> how can I use the s3 mirros?
[22:33] <treeder> mirrors?
[22:33] <treeder> we tried that yesterday and it didn't seem to work
[22:35] <utlemming> treeder: configure cloud-init with
[22:35] <utlemming> #cloud-config
[22:35] <utlemming> output: {all: '| tee -a /var/log/cloud-init-output.log'}
[22:35] <utlemming> apt_mirror: http://us-east-1.ec2.archive.ubuntu.com.s3.amazonaws.com/ubuntu
[22:35] <utlemming> ssh_import_id: [utlemming]
[22:35] <utlemming> arg....I meant to put: http://paste.ubuntu.com/854643/
[22:35] <Daviey> utlemming: not dug into it, but an openstack cloud instance upgraded from oneiric to precise.. rebooted gave me this: http://pb.daviey.com/nObX/
[22:37] <Daviey> utlemming: looks like bug 936667
[22:37] <uvirtbot`> Launchpad bug 936667 in upstart "Upstart early job logging causes boot failure for systems with no initramfs (error is "No available ptys")" [High,Confirmed] https://launchpad.net/bugs/936667
[22:39] <treeder> utlemming, will the s3 mirrors become default soon?
[22:39] <utlemming> treeder: hopefully
[23:02] <adam_g> win 6
[23:04] <Canadian1296> How do I set up an sftp server on my server to allow authorized users to access certain directories?
[23:07] <SpamapS> Canadian1296: sftp is built in to sshd
[23:07] <Canadian1296> SpamapS: But how do I configure it?
[23:08] <SpamapS> Canadian1296: its already setup usually. Just 'sftp user@host'
[23:09] <Canadian1296> SpamapS: How do I set which folders they can access?
[23:10] <Canadian1296> Can I set a root directory for them, so they can't leave that directory?
[23:10] <Canadian1296> I only want the user to be able to access /var/www/theirname, no other part of the system
[23:11] <SpamapS> Canadian1296: for that you have to chroot them
[23:11] <Canadian1296> How?
[23:11] <SpamapS> Canadian1296: http://solderintheveins.co.uk/2011/03/ubuntu-sftp-only-account-how-to/ pretty good HOWTO there.. I think.. haven't read the whole thing in detail but it gets all the basic concepts
[23:13] <Canadian1296> It looks like I can follow that and simply set their home to /var/www/theirusername (instead of in /home), since they can't leave their home or execute shell commands. Thanks :)