[00:44] <seicherlbob> hi there! I'm running a lucid server and i'm trying to install sun-java6 packages. I uncommented the "partners" repo line and run apt-get update, but still no sun-java packages are found. what am i doint wrong?
[01:23] <kklimonda> seicherlbob: oracle no longer grants distributions a license to redistribute "sun" java packages. You'll have to download it from their site
[01:23] <seicherlbob> thanks, i just read the "news"
[01:31] <twb> Or use an open alternative
[02:16] <inka> Hi, I wasn't sure what the best channel would be to ask this sort of question.
[02:17] <twb> !ask
[02:17] <inka> I'm currently investigating several different cloud-computing solutions
[02:18] <inka> I plan to develop a social mobile app that utilitizes Facebook for authentication
[02:19] <inka> I've been looking primarily between Amazon EC2, Google App Engine as solutions that would allow me to scale-out nicely
[02:20] <inka> I like that Google App Engine essentially scales automatically but I hear a lot of talk about being "Locked In" but I haven't really found too much that elaborates on this issue
[02:21] <inka> I saw some stackoverflow questions on this but much of them were dated back to 08-09 so I'm not sure how relevant they still are
[03:37] <jknight> Hi
[03:37] <jknight> I have a question
[03:37] <jknight> does anyone know why my cron job is not running if the user is not logged in?
[03:49] <qman___> jknight, do you have encrypted home directory enabled for that user?
[04:02] <jknight> yes I do
[04:02] <jknight> I think that's the reason why it's not working
[04:03] <jknight> but why is it behaving this way?
[05:12] <linocisco> hi
[05:16] <linocisco> hi all
[05:16] <linocisco> Zentyal has failover feature with 2 ISP links . It is ubuntu based. why ubuntu server could not do like that?
[05:26] <Q_Continuum> So.  Landscape.  I can't afford it for my home machines, but need that capability to keep using it.  Are there any Free self-hosted alternatives, or do I go over to CentOS for my Linux non-Desktop use?
[05:28] <twb> Q_Continuum: landscape server side is not free software, and there is no compatible alternative.
[05:28] <twb> Q_Continuum: if you replace both server and client side, you could deploy puppet or chef or cfengine
[05:28] <twb> Q_Continuum: IMO these have enough grief that they aren't worth considering unless you have at least ten hosts on a LAN.
[05:31] <Q_Continuum> So Canonical/Ubuntu has no alternative for those unwilling to fork out $100-$300/seat, or for those wanting to learn the system :-/  I would've hoped they'd come up with an alternative to putting Enterprise features out of reach of everyone, something they promised NOT to ever do.
[05:31] <Q_Continuum> Guess I'll switch, since I heard puppet is a PITA to learn - but once you know it it rocks.  Only recently heard of chef, haven't looked at it, and no idea on cfengine.
[05:33] <Q_Continuum> Presently I have 2 Ubuntu Server VMs, but will be adding more.  Personal home server, for various tasks and learning more stuff.  Wanted to play with enterprise-level capabilities, like I can with Windows via AD, but guess I don't get to, unless I have 5+ and want a 60-day trial that then ends :-(
[05:34] <Q_Continuum> CentOS it is, I guess.
[05:36] <twb> AFAICT chef has basically all the same faults as puppet
[05:37] <Q_Continuum> Figures.
[05:39] <andol> Q_Continuum: While Puppet sure has its limitation I wouldn't call it as PITA to learn. At least basic usage of Puppet, which still can be a big gain, is fairly trivial.
[05:40] <Q_Continuum> I like the concept of a web-interface though.  Easy to do from anywhere, without needing a terminal
[05:43] <andol> twb: By the way, I am happy running (a masterless) Puppet just for my personal computers, which are well below ten.
[05:43] <Q_Continuum> I run Windows mostly, but Linux for a few things (Minecraft server, irc shell, webserver stuffs)
[05:44] <Q_Continuum> but am looking to expand, hence the want for some easy managability
[05:45] <twb> andol: I was unimpressed in particular by puppet because 1) the wire protocol isn't versioned, and the devs expect you to install the latest bleeding-edge version on all your hosts; 2) it requires ruby on the client side as well as the master; and 3) anything remotely complicated often ends up being done by poorly-tested Exec grep/sed lines, which IMO is no better than if you weren't using puppet in the first place.
[05:46] <twb> e.g. for (3) I ran into it in about ten minutes when I wanted to manage sysctl.conf
[05:47] <twb> And writing a parser for that, which should've been straightforward, was inhibited by my inexperience with Ruby and partly by there it not being having to first write puppet rules to distribute the new parser to all the puppet clients.
[05:47] <twb> Bottom line is that the puppet people felt like fly-by-night cowboys rather than professionals
[05:48] <koolhead17> twb: what would you suggest other than puppet then :)
[05:48] <Q_Continuum> Yeah, I want stuff that is more polished and easy-to-use than that.
[05:48] <andol> twb: Yepp, agree Puppet isn't the magical bullet, and it does have plenty of limitations. Still for me the gain for what I am using it is still a net gain.
[05:48] <twb> Oh, I forgot to mention that the puppet file syntax is obtuse and unhelpful and silly
[05:49] <twb> koolhead17: unfortunately there is no real competition except cfengine (which is differently bad) and chef, which AFAICT is not really any better
[05:49] <koolhead17> andol: use bash :)
[05:49] <RoyK> twb: I attended puppet training a few months back, and that helped a lot ;)
[05:50] <twb> What I am actually doing in practice right now is to create a template container from a script, and then derive specific containers by copying the template and applying a second container-specific script.
[05:50] <koolhead17> twb: am still surviving on bash, but planning to learn puppet soon
[05:50] <twb> RoyK: I was working with Daniel Pittman, whom thingo Labs currently employs...
[05:50] <RoyK> koolhead17: bash really isn't a replacement for puppet :þ
[05:50] <koolhead17> is there no python based config management
[05:50] <twb> RoyK: I'm still having trouble with it.  Oh, also I read the whole Apress book
[05:50] <twb> koolhead17: AFAIK no
[05:51] <koolhead17> RoyK: :(
[05:51] <twb> IMO "not really any better" applies for python->ruby too
[05:51]  * RoyK prepares to go to work and not sitting around on irc all day
[05:51] <RoyK> s/day/morning/
[05:51] <koolhead17> twb: i wonder setting up a puppet server/config will be  a pain in itself
[05:52] <twb> I would be a LOT more confident about puppet if it had a plain C client side (so it can fit on embedded hosts), and it had a purely declarative SANE file syntax, and a versioned wire protocol that was clearly documented.
[05:53] <twb> I also seem to recall puppet's use of TLS was mainly self-signed certs, to avoid having to bootstrap the cert trust chain before you could run puppet
[05:53] <twb> koolhead17: you can get up a trivial instance pretty quickly
[05:53] <twb> koolhead17: being fancy takes more time
[05:55] <koolhead17> i just hope learning puppet won`t turn into a new expedition 4 me :)
[05:55] <andol> twb: Of those things you listed, I would probably rank versioned wire protocol the highest.
[05:55] <twb> andol: 100% agree
[05:55] <twb> koolhead17: IMO do it later
[05:56] <koolhead17> twb: but i have to play with it :P
[06:00] <andol> twb: And as a side note, for recention version of Debian/Ubuntu I would rather put my own changes under /etc/sysctl.d/ rather than modify /etc/sysctl.conf. Not that the initial assumption necessarily applies to your case.
[06:01] <koolhead17> twb: so the who schema/way of writing changes every version of puppet?
[06:02] <andol> koolhead17: It doesn't have to change, it is more that it can change.
[06:03] <twb> andol: that is a good point; this was for an older system (8.04 IIRC)
[06:03] <andol> koolhead17: You might very well be able to run separate Puppet versions along side each other, as long as you verify that being the case, that your recipies works for both versions, etc.
[06:04] <koolhead17> andol: that is confusing :(
[06:13] <andol> koolhead17: It is an imperfect world :)
[07:06] <koolhead17> andol: :P
[09:05] <randomDude> hey so i have this ubuntu server, and postgress seems to be borked... : $ sudo service postgresql restart; Error: /var/lib/postgresql/9.1/main is not accessible or does not exist
[09:05] <randomDude> 11.10
[09:05] <randomDude> i can't sudo apt-get install postgres --reinstall either
[10:00] <lynxman> morning o/
[11:48] <M0nt7> where is the rsync.conf file in ubuntu?
[11:48] <M0nt7> anybody on this channel?
[11:52] <andol> M0nt7: rsyncd will read /etc/rsyncd.conf by default, but you will have to create that file yourself.
[11:52] <andol> M0nt7: See also /usr/share/doc/rsync/examples/rsyncd.conf
[11:53] <M0nt7> andol:  what would be a basi configuration for rsyncd.conf?
[11:54] <andol> M0nt7: Well, base or not, you will still have to provide explit shares. The rsyncd.conf(5) does provide some good examples for different situations.
[11:55] <M0nt7> andol:  ryncd.conf (5) is the man page right?
[11:56] <andol> yepp, search for EXAMPLES
[11:56] <M0nt7> andol: I get no manual entry rsyncd.conf
[11:57] <andol> ...and you have the rsync packages installed?
[11:58] <M0nt7> andol: yes indeed
[11:59] <andol> odd
[12:20] <Amin0v> hello all
[12:20] <Amin0v> how can I create a repository in my svn server and then ask client to checkout from a subdirectory in this repository
[12:20] <Amin0v> let's say the repository is in /home/user
[12:20] <Amin0v> and that people must check from /home/user/web
[12:21] <Amin0v>  /s/check/checkout
[12:33] <jMCg> wi Leseb
[12:34] <Leseb> jMCg: ?
[12:36] <jMCg> Leseb: missed the /
[12:38] <Leseb> jMCg: what do you mean?
[12:39] <jMCg> Leseb: /wi Leseb --> /whois Leseb
[12:40] <Leseb> Oh ok, and why do you whois me?^^
[12:57] <dnivra> hello. i just installed ubuntu-server 11.10 and installed selinux. After installation, I reboot and get "grub rescue>". what could I do to get grub back?
[12:58] <dnivra> i booted a live cd and selected "Reinstall grub boot loader" in rescue mode and tried installing grub to all partitions but no luck.
[13:22] <memoryleak> I've installed ubuntu server in a vm using paralells - but I can't mount CDROM drive. How I can list cdrom devices ?
[13:31] <uvirtbot`> New bug: #941905 in keystone (universe) "keystone user should not have a primary group of nogroup" [Undecided,New] https://launchpad.net/bugs/941905
[13:32] <uvirtbot`> New bug: #931893 in swift (main) "swift service scripts will not start unless /etc/swift/*-server/* exists" [Undecided,New] https://launchpad.net/bugs/931893
[13:38] <zul> good morning
[13:39] <rbasak> Daviey: what are the different mechanisms for chain loading on Intel? pxelinux.cfg localboot, what's the other chain loader you mentioned? And kexec of course. Anything else?
[13:39] <rbasak> hey zul
[13:39] <zul> ipxe
[13:43] <Daviey> rbasak: chain.c32
[13:43] <rbasak> Thanks!
[13:43] <Daviey> ^^ post reliable and recommended soltuion on traditional arch
[13:43] <uvirtbot`> Daviey: Error: "^" is not a valid command.
[13:51] <roaksoax> morning
[13:54] <smb> Daviey, Using do-release-upgrade is the preferred way to run an upgrade on server? If so, is there some preferred (and preferably simple) way to make it non-interactive. The wild world wisdom has some scary looking suggestions (including expect scripts)...
[14:04] <Daviey> smb: expect scripts?!
[14:04] <Daviey> crikey.
[14:05] <Daviey> smb: I don't think i'd recommend doing it completely hands off tbh.
[14:07] <smb> Daviey, Not sure any admin having to do a number higher that a few will agree. :) Hence those "work-arounds". Also it would be nice to have for an automated test. ;)
[14:08] <Daviey> smb: yeah.. there ARE automated tests.. i suspect they just sed -i 's/oneiric/precise/g' /etc/apt/sources.list
[14:09] <smb> Daviey, Right, and with that you can have some control over frontend and dpkg options when running apt-get dist-upgrade.
[14:10] <smb> Just thought about having the propagated way doing the same. But it does not offer much options (not even the frontend choices are really explained).
[14:11] <Daviey> smb: Yeah.. Sorry, not much assistance i can offer with that... Perhaps we need to discuss it at UDS. :)
[14:12] <smb> Daviey, We surely can and should. For precise it will be too late now anyway. So I guess people will go with dist-upgrade or expect scripts. ;)
[14:14] <Daviey> smb: sad. :(
[14:16] <Jeeves_> plz click affects me: https://bugs.launchpad.net/ubuntu/+bug/939300
[14:16] <uvirtbot`> Launchpad bug 939300 in ubuntu "precise 12.04: consider adding Apache 2.4 ?" [Wishlist,Confirmed]
[14:51] <uvirtbot`> New bug: #941955 in cloud-init (main) "util.islxc() method is broken without is-lxc-container" [Undecided,New] https://launchpad.net/bugs/941955
[15:01] <uvirtbot`> New bug: #941968 in lxc (universe) "lockfile-create hangs inside lxc containers" [Undecided,New] https://launchpad.net/bugs/941968
[15:11] <uvirtbot`> New bug: #941983 in freeradius (main) "freeradius refuses to run on Ubuntu 11.10" [Undecided,New] https://launchpad.net/bugs/941983
[15:36] <uvirtbot`> New bug: #941993 in php5 (main) "package libapache2-mod-php5 5.3.10-1ubuntu1 failed to install/upgrade: ErrorMessage: subprocess installed post-installation script returned error exit status 1" [Undecided,New] https://launchpad.net/bugs/941993
[15:43] <Amin0v> when I do an svn list like in this example :  http://pastebin.com/8Er4DNaT
[15:43] <Amin0v>  I find my commited files
[15:43] <Amin0v> when I ckeck the svn project I don't find this files
[16:31] <uvirtbot`> New bug: #942061 in cloud-init (main) "need support for maas" [Undecided,New] https://launchpad.net/bugs/942061
[16:57] <zul> smoser: http://paste.ubuntu.com/859404/ and the output http://paste.ubuntu.com/859407/
[17:00] <smoser> hm..
[17:00] <smoser> do you need the console=ttyS0 ?
[17:04] <mdeslaur> SpamapS: FYI, I've uploaded my preliminary mysql updates here: https://launchpad.net/~ubuntu-security-proposed/+archive/ppa/+packages
[17:06] <SpamapS> mdeslaur: great. Will poke a few people to test
[17:07] <mdeslaur> SpamapS: awesome, thanks
[17:16] <zul> smoser: yes otherwise it wont work
[17:16] <smoser> zul, what does it do wht that ?
[17:16] <smoser> ah...
[17:16] <zul> smoser: i need to clean that patch up but it works
[17:16] <smoser> ah, yeah, plymouth and upstart do some stuff baed on that on the kernel command line
[17:17] <smoser> but... strange
[17:17] <smoser> wait.
[17:17] <smoser> i dont beleive you
[17:17] <zul> smoser: fine try out the patch for yourself then
[17:17] <smoser> the only place upstart would go looking for 'console=' parame would be in kernel cmdline
[17:17] <smoser> i would  not hav thought it would look on its own command line
[17:18] <zul> smoser: right but libvirt-lxc starts without cmdline
[17:19] <smoser> zul, i'm confused.
[17:19] <smoser> libvirt-lxc starts init without parameters
[17:19] <zul> right
[17:19] <smoser> but you added some parameter "console=ttyS0"
[17:19] <zul> right
[17:19] <smoser> which i can only assume means that its going to append that to /sbin/init
[17:19] <smoser> which i dont think will do anything
[17:20] <zul> http://libvirt.org/drvlxc.html
[17:21] <zul> smoser: and from my log file: http://paste.ubuntu.com/859436/
[17:23] <smoser> zul, so you're asserting that upstart reads that
[17:23] <zul> smoser: it *does* work provided you have the console fix as well: http://libvirt.org/git/?p=libvirt.git;a=commit;h=9130396214975ba2251082f943c9717281039050
[17:23] <zul> smoser: yes
[17:25] <zul> smoser: i have 5 instances running all with a working euca-get-console-output
[17:25] <smoser> hallyn, ^ stgraber ^ is that true
[17:25] <smoser> $ grep -r "LXC_COMMANDLINE" . || echo NOT HERE
[17:25] <smoser> NOT HERE
[17:26] <smoser> oh. and that is in a checkout of lp:ubuntu/upstart
[17:27] <zul> smoser: LIBVIRT_LXC_CMDLINE is a libvirt thing
[17:27] <hallyn> is what true
[17:27] <smoser> yes, but upstart is an upstart thing
[17:27] <smoser> and upstart is what determines where console output goes
[17:28] <smoser> (other than kernel messages)
[17:28] <smoser> zul, i'm asserting that your "console=ttyS0" is doing nothing but causing confusion.
[17:28] <hallyn> pretty sure upstart listens to console= kernel cmdline
[17:28] <smoser> yes
[17:29] <smoser> the kernel command line it does, hallyn (or rather at least reads it and it and plymouth decided on some stuff together based on it)
[17:29] <smoser> but zul is suggesting that his patch at http://paste.ubuntu.com/859404/
[17:29] <hallyn> dude i'm just trying to get reboot to work in $*)(&%$* libvirt-lxc...
[17:29] <smoser> needs the "cmdline" tag
[17:29] <smoser> s/tag/hunk/
[17:29] <smoser> but i dont understand how that would do anything
[17:29] <cwillu_at_work> not to jump in the middle of anything, but console= is interpreted by the kernel as well
[17:29] <smoser> eys
[17:29] <smoser> cwillu_at_work, right.
[17:29] <hallyn> lxc doesn't run a kernel though :)
[17:30] <smoser> i'm just trying to figure out why adding 'cmdline' in that patch there would do anyting.
[17:30] <cwillu_at_work> ah, I see
[17:31] <hallyn> which patch
[17:31]  * cwillu_at_work decides to remain happily oblivious
[17:31] <jhobbs>  /wg 9
[17:32] <zul> hallyn: http://paste.ubuntu.com/859404/
[17:34] <zul> smoser: its all doug henningish
[17:49] <zul> smoser: any more comments?
[17:52] <zul> hallyn: i have a libvirt-lxc patch for you after beta
[17:53] <kirkland> smoser: hey, are you experiencing any unexpectedly high load in precise/byobu?  I just upgraded on Friday, and load seems way high
[17:54] <kirkland> but I also started trying unity 3d for the first time in ~3 months, so that's probably it
[17:55] <hallyn> zul: ok.  feel free to stash it in ubuntu:libvirt (preserving my delta), else i'll do it.
[17:55] <zul> hallyn: ack
[17:55] <hallyn> i just can't get reboot to work.  only thing i can figure is sigchld must be killing the parent before i get a chance to catch it
[18:14] <fixxxermet> I'm using Oneiric and I am trying to determine if the orchestra package, which requires puppet, has any type of built-in integration
[18:15] <fixxxermet> Does anyone know if there is any type of integration, or can they point me to any docs?
[18:29] <sidnei> hallyn, hola, looks like lxc-create -t ubuntu -- -b <username> doesn't work anymore if <username> has uid/gid 1000, the ubuntu user created in the lxc has the same uid/gid
[18:30] <hallyn> sidnei: gosh, good point.  could you open a bug on that?
[18:30] <sidnei> hallyn, oki
[18:31] <hallyn> stgraber: ^ i will make lxc not create the ubuntu user if -b is specified, and add the bound user to sudo/admin group instead.
[18:31] <hallyn> stgraber: ^ lemme know if there's a reason that's a bad idea
[18:31] <hallyn> jsidney: thanks
[18:31] <hallyn> gah
[18:31] <hallyn> sidnei: thanks
[18:31] <sidnei> hallyn, np
[18:33] <stgraber> hallyn: sounds good
[18:36] <sidnei> bug #942144
[18:36] <uvirtbot`> Launchpad bug 942144 in lxc "Using bindhome option of ubuntu template conflicts with ubuntu user" [Undecided,New] https://launchpad.net/bugs/942144
[18:36] <sidnei> hallyn, ^^
[18:36] <hallyn> thanks, sidnei.  Note we're in a freeze, so mar 3 is when the fix can go in for this (iirc)
[18:37] <sidnei> hallyn, ok.
[18:45] <uvirtbot`> New bug: #942144 in lxc (universe) "Using bindhome option of ubuntu template conflicts with ubuntu user" [High,Confirmed] https://launchpad.net/bugs/942144
[19:03] <smoser> hallyn, ping
[19:06] <zul> smoser: interesting http://paste.ubuntu.com/859574/
[19:08] <hex20dec> Best free webmail client for my server?
[19:09] <Pici> hex20dec: I've heard some good things about squirrelmail
[19:09] <hex20dec> Pici, I've used it a while back.
[19:09] <hex20dec> I hate it.
[19:09] <hex20dec> No offense.
[19:09] <smoser> whats interesting there zul?
[19:10] <Pici> hex20dec: none taken. I don't use it myself.
[19:10] <zul> smoser: it stripped out the "\" which i expected
[19:10] <Myrtti> hex20dec: some places where I've used webmail in have usually had both squirrel and later on roundcube
[19:11] <hex20dec> Myrtti, Yeah, I used them both on my old shared hosting.
[19:11] <smoser> zul,  your shell stripped out the '\' i suppose
[19:12] <zul> smoser: perhaps
[19:12] <smoser> $ cat /tmp/my.sh
[19:12] <smoser> #!/bin/sh
[19:12] <smoser> for a in "$@"; do printf "%s\n" "$a"; done
[19:12] <smoser> $ /tmp/my.sh foo \xasdf
[19:12] <smoser> foo
[19:12] <smoser> xasdf
[19:13] <hex20dec> What do you guys use for mail client?
[19:15] <Pici> gmail <.<
[19:18] <hex20dec> Pici, you forward all your emails to Gmail?
[19:18] <hallyn> smoser: .
[19:19] <Pici> hex20dec: Actually, yes. But you could always use Google apps for domains, although I don't think its complete free (as in beer) anymore.
[19:21] <smoser> hallyn, https://bugs.launchpad.net/cloud-init/+bug/941955
[19:21] <uvirtbot`> Launchpad bug 941955 in cloud-init "util.islxc() method is broken without is-lxc-container" [Medium,Triaged]
[19:22] <smoser> http://paste.ubuntu.com/859594/ is what my /proc/1/cgroup looks like
[19:22] <smoser> given that there are 8 things in which i might be in the root group of, which should i look at for this
[19:23] <jjohansen> hallyn: ping
[19:23] <hallyn> stgraber: ^ did you change the lxc-is-container logic at all?  (i'm fetching the pkg)
[19:25] <stgraber> hallyn: lxc-is-container no longer exists
[19:25] <hallyn> smoser: i can't find any code that gives a rip about /proc/1/cgroup involved
[19:25] <stgraber> hallyn: you need running-in-container
[19:25] <hallyn> stgraber: yes, smoser says that still doesn't work
[19:25] <stgraber> hallyn: which uses the container-detect upstart job
[19:25] <hallyn> jjohansen: (i'm here :)
[19:26] <smoser> stgraber, cloud-init code that you tol dme to write looks at that.
[19:26] <smoser> hold on
[19:26] <smoser> for link
[19:27] <jjohansen> hallyn: okay, so apparmor clear FFe on friday, /me doesn't know when that means it will show up.  But I don't expect the kernel will have the patches in beta1, they hadn't been pulled in last I checked
[19:27] <jjohansen> s/clear/cleared/
[19:28] <jjohansen> hallyn: /me still has built kernels with the patches if you want
[19:29] <hallyn> jjohansen: yes, i certainly hope i'l lhave time to experiment with the moutn rules before bet1 is released, so test kernels woudl be great, thanks
[19:30] <smoser> hallyn, http://bazaar.launchpad.net/~cloud-init-dev/cloud-init/trunk/view/head:/cloudinit/util.py
[19:30] <jjohansen> hallyn: http://people.canonical.com/~jj/linux-image-3.2.0-17-generic_3.2.0-17.26~aamount_amd64.deb
[19:30] <smoser> see islxc there.
[19:31] <hallyn> smoser: why not ditch taht first part?
[19:31] <hallyn> the running-in-container is now (i *think*) intended to be a definitive test
[19:32] <stgraber> smoser: right, call running-in-container, if return value is 0 you're in a container (it's also going to tell you want kind of container on stdout)
[19:32] <stgraber> smoser: that's assumin this python script is running after upstart is started though
[19:32] <hallyn> else yes, you'd have to do the equiv of 'head -1 /proc/1/cgroup | awk -F: '{ print $3}'' - as several lxc tools had to do
[19:33] <stgraber> (running-in-container depends on /run/container_type that's generated by the container-check upstart job (start on starting and mounted MOUNTPOINT=/run)
[19:33] <stgraber> )
[19:33] <hallyn> stgraber: i wonder if lxc should ship a symlink from is-lxc-contaienr to running_in_container, since is_lxc-container was shipped in lxcguest
[19:34] <stgraber> hallyn: having it ship in lxc wouldn't really solve the problem as we don't install lxc where we used to have lxcguest
[19:34] <hallyn> stgraber: true
[19:34] <stgraber> hallyn: and we should be able to catch and fix all cases by looking at lxcguest's rdepends (which AFAIK we did)
[19:35] <hallyn> jjohansen: wget'ed that kernel.  if this laptop should die, i may end up having to ask you wehre it is again :)
[19:36] <jjohansen> hallyn: np, just ping me if you need anything, /me should be around most of the day (just have to run an errand at lunch)
[19:38] <smoser> stgraber, well, i was calling lxc-is-container before, and i really tihnk that you should fix that to be backwards compatibile
[19:39] <smoser> i will change to (also) use running-in-container, but i need a path that works wen that command is not available also.
[19:39] <stgraber> smoser: that command is part of upstart, so it should always be there
[19:39] <stgraber> smoser: and it's in /bin so even with a separate /usr it's going to be available
[19:40] <smoser> stgraber, yes, it will always be there on ubuntu >= 12.04
[19:40] <smoser> but thats not good enough
[19:40] <hallyn> jjohansen: libvirt is bogging me down today (i blame smoser), so i won't be bugging you today.  thanks again.  ttyl
[19:41] <stgraber> smoser: hmm, ok, then check for either is-lxc-container or running-in-container or copy/paste the logic from container-detect.conf
[19:44] <smoser> stgraber, well, copy and paste that wont work all the way
[19:44] <smoser> as cloud-init is not init and is not privey to init's environment variable
[19:44] <smoser> regarding LIBVIRT_LXC_UUID
[19:44] <smoser> but i will use similar logic to that
[19:44] <stgraber> smoser: you can grep /proc/1/environ though
[19:45] <smoser> yeah. i guess i could do that.
[19:45] <stgraber> or source init's environment but that's kind of ugly ;)
[20:07] <pythonirc101> Any ideas how to  boot into single user mode on this ubuntu box.  There does not seem like a kernel line in my grub 1.99 "edit" view.
[20:16] <arosales> jcastro: m_3: Travel safely  and kill it in Strata!
[20:16]  * jcastro isn't going to strata
[20:16] <jcastro> but mims is!
[20:18] <arosales> jcastro: Sorry, https://juju.ubuntu.com/Events had you listed
[20:18]  * jcastro fixes
[20:21] <arosales> jcastro: thanks :-)
[21:25] <wonderman> can someone help me diagnose server freezing please
[21:26] <wonderman> i can reboot it remotely from my control panel with hosting provider
[21:26] <wonderman> it stays up for 5 min then goes down
[21:26] <guntbert> wonderman: what do the logs say?
[21:26] <wonderman> i cant get in to see
[21:26] <wonderman> in the 5 min i had, i checked syslog
[21:26] <wonderman> and messages
[21:26] <wonderman> only a few relating to NTPD
[21:27] <guntbert> wonderman: logs survive a reboot
[21:27] <wonderman> also this <>  Feb 27 15:00:29 server1 kernel: [   36.173580] possible SYN flooding on port 80. Sending cookies.
[21:27] <wonderman> yes i know, but i got 5 min before maybe it happens again!
[21:27] <wonderman> i will stop apache etc
[21:27] <wonderman> as soon as it connects
[21:28] <wonderman> ok im back in, how can i find mate?
[21:28] <wonderman> i did 'cat /var/log/syslog | less'
[21:28] <guntbert> wonderman: good, although less /var/log/syslog is better
[21:28] <wonderman> ok, dates arent in order?
[21:29] <wonderman> let me try your way
[21:29] <guntbert> wonderman: and tailf /var/log/syslog will keep you informed until it crashes
[21:29] <wonderman> yea i did that, but only had one window and changed !
[21:29] <wonderman> lemme grab another open
[21:30] <wonderman> Feb 27 15:30:28 server1 named[1181]: client 200.90.132.200#64938: query (cache) 'ns1.serv1mail.com/AAAA/IN' denied
[21:30] <wonderman> dropping that
[21:32] <wonderman> guntbert: why would syslog msgs not be in the correct order?
[21:32] <wonderman> one minute its 11am, then 9pm, then 3pm
[21:32] <guntbert> wonderman: your time setting - do you use ntp?
[21:32] <wonderman> yes there are msgs relating to this
[21:33] <wonderman> i dont remember installing it if its not default
[21:33] <wonderman> maybe i did
[21:33] <wonderman> this is it crashing?
[21:33] <wonderman> Feb 27 15:21:59 server1 kernel: [ 1321.502878] Pid: 2604, comm: processes Not tainted 2.6.32-38-generic #83-Ubuntu
[21:33] <wonderman> Feb 27 15:21:59 server1 kernel: [ 1321.502881] Call Trace:
[21:33] <wonderman> shall i pastebin?
[21:33] <guntbert> wonderman: try
[21:34] <wonderman> try pastebin? sorry i am impatient, panicing a bit :P
[21:40] <guntbert> wonderman: sorry, didn't want to upset you still more - I wanted to say: give it a try, maybe someone spots something
[21:47] <adam_g> lifeless: ping
[21:52] <wonderman> erm
[21:52] <wonderman> !
[21:53] <wonderman> guntbert: what was you saying about NTP?
[21:53] <wonderman> should/can i remove it?
[21:55] <guntbert> wonderman: no, it should be enabled, or else the system time would certainly drift away, but those big steps are weird, can you pastebin the syslog? use !pastebinit
[21:55] <guntbert> !pastebinit | wonderman
[21:55] <wonderman> ok thanks man, sec
[21:55] <wonderman> its yet to crash, with apache off
[21:56] <lifeless> adam_g: hi
[21:57] <wonderman> guntbert: how can i copy from 'less' ? or cant ?
[21:57] <wonderman> i cant scroll in terminal and copy at the same time
[21:58] <guntbert> wonderman: on the server:   apt-get install pastebinit       then      pastebinit  /var/log/syslog          it will tell you an URL
[21:58] <wonderman> ok sorry
[21:59] <adam_g> lifeless: hey, i was looking at Bug #924739, trying to come solution. wondering if you had any input, in terms of the features that are not yet ported to squid3
[21:59] <uvirtbot`> Launchpad bug 924739 in squid3 "after upgrade from oneiric to precise, previous squid config unused, cannot be used when relocated" [Critical,Triaged] https://launchpad.net/bugs/924739
[22:00] <guntbert> and wonderman: please use my nick when talking to me - else I might overlook your response :)
[22:00] <wonderman> guntbert:  http://pastebin.com/686SyAzn
[22:00] <wonderman> i had already started
[22:01] <wonderman> guntbert: i copied a section to where i think it 'crashed' just before i rebooted which you can see from the paste too, and the time is totally different... the reboot time is UTC
[22:03] <guntbert> wonderman: look at line 106: out of memory kills apache - some process hogs your memory and then the kernel decides to kill one (big) process
[22:03] <wonderman> sedc
[22:04] <wonderman> wtf
[22:04] <wonderman> it swapped?
[22:05] <wonderman> hmm
[22:05] <wonderman> so this is a RAM causing the crash?
[22:06] <wonderman> let me start apache again, and take a look
[22:06] <guntbert> wonderman: no, the system has definitely not enough memory (probably some process eating it up) - please read up on OOM  - its the first one I have seen yet
[22:06] <wonderman> OOM?
[22:06] <wonderman> i have apache max clients set to 1024, but its been fine for many many days
[22:06] <wonderman> i think maybe its 'awstats'
[22:07] <wonderman> as there is a CRON running quite close to the crash?
[22:07] <wonderman> i will start apache and monitor it quickly?
[22:09] <guntbert> wonderman: OOM=out of memory
[22:09] <wonderman> i am tailing syslog, and its trying to run a load of crons
[22:09] <wonderman> because it cannot run them, it keeps trying
[22:10] <wonderman> how can i stop this? because upon reboot it will try to run all crons at once...
[22:10] <wonderman> think thats why its crashing
[22:10] <wonderman> upon reboot, i cant see it using up 4GB of RAM in 5minutes
[22:12] <guntbert> wonderman: I don't think so - the RAM usage goes quickly up to nearly 100% - thats ok, please search for    linuxatemyram  - a very instructive article
[22:12] <wonderman> i will monitor it, as it hasnt happened for days, i dont know why it would be now
[22:13] <wonderman> ill see what happens when i turn apache on
[22:13] <wonderman> as atm its fine...
[22:14] <guntbert> wonderman: Good luck :-) I'm off to bed
[22:14] <wonderman> ok pal thanks
[22:32] <randomDude> I need some advice on how to purge a broken postgres install : http://dpaste.com/708653/
[22:35] <SpamapS> randomDude: interesting, can you try editing /var/lib/dpkg/info/postgresql-9.1.prerm and add '-x' to the #!/bin/sh line at the top? Then re-run.. it will give details as to what failed
[22:36] <randomDude> SpamapS, as a caveat, /var/lib/postgres-9.1/main doesn't actually exist.
[22:37] <SpamapS> randomDude: that shouldn't preclude you from being able to remove.
[22:37] <randomDude> SpamapS, http://dpaste.com/708655/
[22:38] <randomDude> http://dpaste.com/708656/
[22:38] <randomDude> gah
[22:38] <randomDude> http://dpaste.com/708657/
[22:38] <wonderman> can anyone tell me why syslogs have different 'timestamps', ie they are not in order
[22:38] <randomDude> uh ok, it's not pasting everything
[22:39] <SpamapS> randomDude: consider using pastebinit ;)
[22:39] <SpamapS> randomDude: sudo apt-get remove postgresql-9.1 2>&1 | pastebinit
[22:40] <randomDude> essentially it says : /var/lib/postgresql/9.1/main is not accessible or does not exist
[22:42] <SpamapS> randomDude: that sounds like a bit of a bug in the prerm or init script, as it shouldn't care about that.. but anyway, perhaps to make things simpler.. just 'sudo mkdir -p /var/lib/postgresql/9.1/main ?
[22:43] <randomDude> ok
[22:43] <randomDude> SpamapS, that may have fixed the issue
[22:44] <randomDude> SpamapS, thank you
[22:44] <SpamapS> randomDude: seems like its a worthwhile bug to report
[22:45] <randomDude> yeah, one would assume that since that highly essential directory is missing that the server process is not actually running
[22:47] <SpamapS> randomDude: I think the only bug is that it didn't explicitly tell you why it failed
[22:47] <SpamapS> randomDude: because of the removal of an essential directory, things were in an inconsistent state.. I don't think it should ignore that, but I do think it should tell you.
[23:06] <talntid> Gr. I'm getting my butt kicked by Apache2. I installed it, and it is looking for /etc/apache2/htdocs regardless of there being an entry in sites-enabled, pushing to /var/www
[23:17] <jMCg> talntid: you're doing it wrong, then.
[23:18] <jMCg> Well, jMCg that wasn't really helpful.
[23:18] <jMCg> talntid: can you please pastebin your apache2.conf - and your vhost, also: apache2ctl -S
[23:36] <talntid> jMCg, actually right now, I'm trying to purge the very existance of apache2, but it's fighting me.
[23:36] <talntid> apache2.2-common : Depends: apache2-utils but it is not going to be installed
[23:36] <talntid> E: Broken packages
[23:37] <jMCg> talntid: aptitude purge apache2.2-common apache2-utils etc..
[23:37] <jMCg> talntid: just get rid of the whole she-bang.
[23:38] <jMCg> talntid: also, sfritsch and daemonkeeper promised me the 2.4 version of httpd will not suck stinky donkey balls in hell.
[23:38] <jMCg> 00:44:25 <+fajita> Please read http://wiki.apache.org/httpd/DebianDeb0rkification
[23:39] <jMCg> talntid: ^ a wiki page we have over at #httpd  for people comming with Debuntu specific problems.
[23:39] <jMCg> most of this is already fixed.
[23:39] <jMCg> I think.
[23:39] <talntid> my issue is more...
[23:40] <talntid> http://pastebin.com/gqDqMGbU
[23:40] <hallyn> stgraber: do you think we should hack lxc-start to, for every contaienr that is running, keep a file open called '$rootfs.lock' (so long as [ ! -b $rootfs])?
[23:40] <hallyn> stgraber: to keep the container from remounting the fs readonly on shutdown
[23:41] <hallyn> bc there won't be a kernel fix for that, at least for precise (maybe ever)
[23:44] <stgraber> hallyn: yeah, I guess we don't really have a choice
[23:44] <hallyn> i guess i'll open a bug for it.  thx
[23:54] <talntid> I took the Windows way out, since I wasn't very invested in the install.
[23:54] <talntid> all who wish to throw rocks, go ahead.
[23:55] <wonderman> qman___: can i holla?
[23:56] <uvirtbot`> New bug: #942325 in lxc (universe) "keep container from marking shared rootfs readonly on shutdown" [Undecided,New] https://launchpad.net/bugs/942325