[00:33] <raubvogel> Where can I get the rfc2307bis schema?
[00:53] <pmatulis> raubvogel: http://www.grotan.com/ldap/rfc2307bis.schema
[03:07] <Darkwing> jcastro: ping
[06:01] <bitmonk> hey guys, i see daily livecds of precise desktop (12.04), but none for server.  i tried building a vm with debootstrap, but it fails to start saying that there's no /bin/init .. anyone tried this?
[06:02] <twb> bitmonk: there isn't supposed to be a /bin/init; it's /sbin/init
[06:03] <bitmonk> right, that's probably just a mistake on my part in what i've said here.  i can verify.
[06:03] <EvilResistance> twb:  /query?
[06:03] <twb> EvilResistance: why?
[06:03] <EvilResistance> regarding the bug thing of TheEvilOne @ #debian-mentors @ OFTC (if you're the same one)
[06:03]  * EvilResistance is TheEvilOne over there
[06:03] <bitmonk> yah there is no /sbin/init
[06:03] <twb> bitmonk: also obviously the server install doesn't get a live CD
[06:04] <EvilResistance> otherwise you can ignore my request :P
[06:04] <twb> EvilResistance: sure wahtever
[06:04] <bitmonk> right, but there is no server iso yet at all afaict, which is not surprising, i only looked for it because my debootstrap failed.
[06:04] <twb> As to debootstrap, you probably don't have upstart installed
[06:04] <twb> bitmonk: if your goal is simply to get a working precise install, you should install 11.10 and d-r-u, or use one of the daily installer builds
[06:05] <bitmonk> hm.  ok.  i just took a script that worked for boostrapping lucid and pointed it at precise, i guess i need to tell it to include upstart.
[06:05] <bitmonk> well, i'm beginning to test it within our infrastructure, this is how we build vms, so if that changes for precise, i can update the chef recipe to get it working.
[06:05] <twb> archive.ubuntu.com/ubuntu/dists/precise/main/installer-{i386,amd64}/current/images/netboot/mini.iso
[06:05] <bitmonk> ah
[06:10] <bitmonk> i think that did it, just adding upstart to debootstrap --include
[06:11] <bitmonk> neat. :)
[06:12] <twb> I advise you to avoid deboostrap if possible, there are non-negligible edge cases that it doesn't hold your hand for (e.g. no /etc/hosts and /etc/hostnames and /etc/fstab, so e.g. some apache modules will mysteriously fail)
[06:12] <twb> They are surmountable but not worth your time learning to deal with
[06:14] <bitmonk> we write all of those out with chef.
[06:14] <twb> haha
[06:14] <twb> Well hopefully the chef client will have enough to start eh
[06:14] <bitmonk> no no, on the host machine
[06:14] <twb> Ah
[06:14] <bitmonk> we have a chef recipe for building vms
[06:15] <bitmonk> knife node list | wc -l : 320 ;)
[06:15] <twb> I just make a site-local template with everything set up, then (basically) dd that to provision a new host
[06:15] <bitmonk> it's an option we've considered, and which i've done in the past, but this seems to work fine.
[06:16] <bitmonk> we also use debootstrap on the physical machines via fai
[06:16] <twb> OK clearly you're not the noob I first took you for ;-)
[06:16] <bitmonk> it's super handy, i can install a machine in like 2 minutes.
[06:16] <bitmonk> heh
[06:16] <bitmonk> no worries.
[06:16] <bitmonk> i knew i should have upstart, i was just concerned that the init system had changed again..
[06:17] <bitmonk> i haven't tracked the non-LTS releases
[06:17] <twb> My main beef with fai (back in 2003) was that it wanted to be "self-hosting", as it were
[06:17] <twb> bitmonk: AFAIK Canonical isn't dropping upstart yet
[06:17] <bitmonk> we noticed today that karmic disappeared from the mirrors (we have <20 hosts, mostly old dev VMs) so i decided we should start testing the next LTS. :)
[06:17] <twb> karmic isn't LTS
[06:17] <bitmonk> right, we have mostly lucid
[06:18] <bitmonk> but it occurred to me..
[06:18] <bitmonk> if karmic just disappeared..
[06:18] <twb> Also LTS applies on a per-package basis, most packages are not 5 year support
[06:18] <bitmonk> lucid will at some point.
[06:19]  * bitmonk rubs his chin wondering where he's going to get chef debs for precise ;d
[06:20] <twb> Does chef have the same bullshit as puppet where they want you to install puppet and half of ruby from the latest bleeding-edge repo, instead of what shipped with the distro?
[06:20] <bitmonk> noop.
[06:21] <bitmonk> lucid ships with chef 0.9.6 or so, and we point at opscode's apt repo, and we get 0.10.whatever
[06:21] <bitmonk> standard ruby
[06:21] <twb> But it still needs ruby on both the master and slave nodes, right?
[06:21] <bitmonk> correct.
[06:21] <twb> Grumble grumble, ruby is overkill for some of my embedded boxes
[06:21] <bitmonk> which i'm not terribly concerned with, i think we keep perl, python, and ruby handy everywhere.
[06:21] <bitmonk> yeah i hear that.
[06:21] <bitmonk> it does not provide us with a solution for network gear.
[06:21] <twb> I wouldn't really care if ruby "infected" only the master
[06:22] <bitmonk> the server is being rewritten in erlang against mysql is my understanding, the client could probably be made smaller.
[06:22] <bitmonk> but i guess, yah, i mean.. chef's recipes are actually ruby code.
[06:23] <bitmonk> i liked that idea more a year ago than i do now.
[06:23] <Tm_T> rruubyyy... why it makes me think food now?
[06:23]  * bitmonk is hella hungry too
[06:27] <twb> Also because I don't really see a difference between perl, python and ruby, and ubuntu boxes already need the first two because that's what core debian and ubuntu stuff are written in
[06:28] <bitmonk> well, they're different because people wrote stuff in 'em ;d
[06:28] <bitmonk> i once wanted the whole world to be in python, but i gave up.
[06:28] <bitmonk> and damned if perl still ain't kickin.
[06:34] <bitmonk> wheezy chef repo seems to work
[06:36] <bitmonk> hm, or not.
[06:37] <twb> Mm, but you are using a random PPA for it rather than <what the distro ships>
[06:37] <twb> That kinda makes me nervous for configuration management
[08:04] <hallyn> SpamapS: bug 914155 fyi
[08:21] <hallyn> rbasak: stgraber: my take on lxc bugs to be fixed today or tomorrow: http://paste.ubuntu.com/799087/
[08:24] <stgraber> hallyn: looks good for the bugs. On the feature side, we still need to talk with the kernel team to get Daniel's patch applied and look at what we want to do with apparmor (ideally, draft a basic profile restricting as much of the bad stuff as we can)
[08:26] <stgraber> hallyn: for util-linux, the patch should be sent to lamont so it also gets into Debian (if not upstream)
[08:27] <hallyn> stgraber: i sent the reboot patch to the kernel team m-l, waiting for a response
[08:28] <hallyn> stgraber: for apparmor, yeah, we should go chat with jj at some point.  But I'd rather get the other bugs sorted first
[08:28] <hallyn> for util-linux, i was waiting for comment from you on whether my proposed patch is sane :)
[08:31] <gary_poster> hallyn, hi.  any availability for a conversation about the ephemeral script?
[08:31] <stgraber> hallyn: so the patch for util-linux is simply changing console.conf to start getty with "console" instead of "/dev/console" and changing logging from error to warning?
[08:32] <hallyn> stgraber: right.  two of the 3 s/err/warn/'s are in an upstream patch right now
[08:32] <hallyn> i assume the 3d needs to be sent upstream and argued
[08:32] <hallyn> gary_poster: can we do it this afternoon?  i'd like to work through these bugs first.
[08:33] <gary_poster> hallyn sure, cool
[08:33] <koolhead17> hi all
[08:33] <hallyn> http://paste.ubuntu.com/799098/ <- updated bug list for the morning
[08:34] <hallyn> stgraber: so you recommend just mailing lamont@debian.org directly (as opposed to opening a debian bug)?
[08:40] <stgraber> hallyn: sending to the Debian BTS is fine too as lamont is the maintainer he'll surely see it (especially as I already mentioned it to him and we've been mentioning his nick a few times on this channel :))
[08:51] <hallyn> stgraber: are you planning to dput the cgroup-lite I posted last night?
[08:56] <stgraber> hallyn: yeah, I'll do that this morning
[08:57] <hallyn> stgraber: great, thanks.  Do I u ndestand right taht updating libcgroup is not *too* urgent, since its version # is lower?
[08:57] <stgraber> hallyn: that's right
[09:00] <hallyn> cool
[09:01] <similian> hello
[09:01] <similian> i have an issue with my br0
[09:02] <similian> when i add tap interfaces
[09:02] <similian> the mac of my br0 is changed to the tap0 instead of eth0
[09:02] <similian> sometimes it stays on eth0
[09:02] <similian> bridging works if mac is eth0 but not if its tap
[09:03] <similian> i have 10.04 (LTS) Kernel  2.6.32
[09:10] <CasmoNL> Anyone here have any idea if there's something like iptables for Windows (Windows 7, I have to route a port from Windows to my Linux server)?
[09:14] <bitmonk> twb: right, but it's not random PPA, it's apt.opscode.com.  it's the vendor managed repository.  for testing sake, i think this will do.  for some reason, precise ships chef from like 2009, worse than lucid by a major version.
[09:14] <bitmonk> 0.9 vs 0.10 chef was important enough for us to use a vendor repo, but i'd love if precise shipped with a modern chef.
[09:20] <stgraber> hallyn: cgroup-lite should still depend on upstart-job and conflict with cgroup-bin right?
[09:20] <hallyn> stgraber: the conflicts is there, no?
[09:21] <hallyn> depend on upstart-job: sounds right
[09:21] <stgraber> hallyn: they were on the old binary package, not the new one. I added that
[09:22] <hallyn> stgraber: did you grab my new one from p.c.c?
[09:22] <stgraber> hallyn: probably not
[09:22] <stgraber> hallyn: same url?
[09:22] <hallyn> http://people.canonical.com/~serge/cgroup-lite.tar.gz
[09:23] <hallyn> (same name as yours was, not mine)
[09:23] <hallyn> i also addressed doko's comments in the other bug
[09:23] <hallyn> (but still dont' have the Depends: upstart-job)
[09:24] <alex88> update-rc.d vm_network_setup start 2 . -> error: expected runlevel [0-9S] (did you forget "." ?)
[09:24] <alex88> why?
[09:25] <stgraber> hallyn: looks pretty good. I'll just update your debian/copyright to have it cover everything, not just debian/ assuming the scripts are also GPLv2+
[09:26] <stgraber> hallyn: uploaded
[09:27] <hallyn> stgraber: \o/ thx
[09:51] <dob_> I created a raid5 array with mdadm. The build has been completed. There are no activities in top. The HDD activity LED is off, but i hear that the system is permanently accessing one or all hard drives. Like a permanent read.
[09:52] <dob_> Even if it's unmounted
[09:55] <_ruben> dob_: perhaps one of the disks is about to die ?
[09:55] <dob_> _ruben: Don't say such things ;-). I will check with smart.
[09:55] <dob_> Hoped there is another possibility....
[09:56] <_ruben> if the light stays off (and lights up when there's activity) during idle, i doubt there's anything reading from the disk
[09:58] <dob_> _ruben: Thank u. I will check that.
[10:18] <RoyK> dob_: smartctl -H /dev/asdf
[10:19] <RoyK> dob_: change asdf to whatever device name the disk has
[10:22] <RoyK> dob_: if that succeeds, try starting a selftest with smartctl -t short and list status for the test(s) with smartctl -l selftest
[10:26] <dob_> RoyK: I will try that later when I am back at the office with the server. Thank u!
[11:06] <chiggins> So I'm trying to enable HTTPS on one of my Apache virtual hosts, but after adding in the cert file, key file, turning the engine on and restarting Apache, I get an error in the log saying invalid request-uri, and chrome says error 107
[11:13] <dax_roc> Morning all
[11:19] <dax_roc> I'm switching nic configuration from round robin to lacp (switch upgrade), so I need ifenslave and just add the configuration to /etc/network/interfaces. I no longer need to edit /etc/modprobe.d/bonding.conf right ( ubuntu 10.04 )
[11:48] <patdk-lap> why would you need to edit that ever?
[11:54]  * andol is slightly confused...
[11:54] <andol> Despite USN-1322-1 having been sent out, the 3.0.0-15 kernel packages only seems to be availible in -proposed?
[11:55] <mdeslaur> andol: there was a snafu with the script that produced the USN email. The actual version and info for the USN is here: http://www.ubuntu.com/usn/usn-1322-1/
[11:56] <andol> mdeslaur: Ahh, thanks.
[12:20] <Tixos> Hi, can someone please discuss with me the correct user layout/permissions for ubuntu servers, and apache?
[12:21] <memoryleak> Hi, I've encountered following problem: http://pastie.org/3159887
[12:21] <memoryleak> I installed pkg-config
[12:26] <memoryleak> still i get these errors
[12:44] <tilen1232> Hi, how to configure multiple ftp accounts (that each has its own public_html) and are managed with apache?
[13:12] <chmac> I'm using the default /etc/postfix/main.cf for a smart-host system. But all mail to user is being expanded to user@mydomain.com and seems to ignore the /etc/aliases file.
[13:12] <chmac> I've run sudo newaliases and I can't see mydomain.com in the main.cf file.
[13:13] <chmac> Hmm, I see the channel has been quiet for a while... :-)
[13:27] <Tm_T> caribou: it happens occasionally, I believe office hours are closing for today in Europe, while US hasn't started yet
[13:30] <caribou> Tm_T: not sure I understand your answdr
[13:30] <caribou> answer
[13:31] <nandemonai> back
[13:36] <Tm_T> caribou: bah, failed with nick tabcompletion, meant chmac (:
[13:36] <Tm_T> sorry (:)
[13:36] <caribou> Tm_T: np :)
[13:51] <chmac> Tm_T: I wondered :-)
[13:59] <smoser> m_3, https://code.launchpad.net/~orchestra/orchestra/odev/ is where "odev" is.
[13:59] <smoser> RoAkSoAx, actually got me up to an (almost) deployed mysql<->wordpress juju
[14:42] <RoAkSoAx> smoser: ?
[14:44] <smoser> RoAkSoAx, here.
[14:45] <smoser> RoAkSoAx, the 'odev' is at a point where "should work" for juju deploy of charms on precise.
[14:45] <RoAkSoAx> smoser: cool
[14:45] <RoAkSoAx> will give it a try later today
[14:56] <afeijo> hi guys, I cant run mysql anymore, it raise an socket error
[14:58] <zul> Daviey:  horizon MIR is waiting on the MIR team now fyi
[14:59] <smoser> RoAkSoAx, thatd' be great. i'd like to see someone else having have done it.
[15:01] <RoAkSoAx> smoser: btw... did you guys starting discussing anything about orchestra?
[15:01] <RoAkSoAx> zul: ^^
[15:01] <zul> RoAkSoAx: i havent
[15:01] <RoAkSoAx> k
[15:03] <smoser> RoAkSoAx, well... there is meetings all day yesterday and today on maas.
[15:04] <RoAkSoAx> smoser: right, so by the new project created about maas, is this a replacement for orchestra, complementing it, a different prohject?
[15:05] <smoser> of those options, closest to "complementing it"
[15:05] <RoAkSoAx> smoser: so I'm guessing nothing has been discussed about extending cobbler?
[15:06] <RoAkSoAx> zul: nor cobbler and ocsinventory integration
[15:06] <zul> RoAkSoAx: nope I been doing oepnstack stuff
[15:06] <RoAkSoAx> k
[15:09] <hallyn> rbasak: ive been wasting time trying to get one patch working - im going to drop that and send you a debdiff in abt 30 mins
[15:10] <rbasak> hallyn: ok, thanks!
[15:17] <hallyn> gary_poster: came by to find you, failed.
[15:17] <gary_poster> hallyn, here now.  Maybe I was in restroom?
[15:17] <gary_poster> hallyn, I could come there if desired
[15:17] <raubvogel> quick check: If I am ssh'ing from A to B and want port 123 in A to be forwarded to 456 in B, I would do something like "ssh -R 123:localhost:456 user@B"
[15:22] <raubvogel> NVM. It's 456:localhost:123
[15:23] <hallyn> rbasak: ok, lxc-ps still needs some love, but I'm looking at http://people.canonical.com/~serge/lxc-rally-p.debdiff
[15:23] <hallyn> lxc-ps really needs to be re-written :(
[15:23] <hallyn> gary_poster: that'd be great, thx
[15:23] <gary_poster> hallyn ok coming over
[15:44] <lynxman> smoser: ping
[15:45] <smoser> here
[15:45] <stgraber> hallyn: did you move the MIR request for cgroup-lite from the cgroup-bin source package to the new cgroup-lite source package?
[15:45] <lynxman> smoser: I can see you from here
[15:46] <RoAkSoAx> lol
[15:46] <eagles051387|> !grub2
[15:47] <Deathvalley122> old
[15:47] <Deathvalley122> :/
[15:47] <Deathvalley122> it hasn't been  updated
[15:48] <eagles051387|> ok but still applies to grub 2 though
[16:19] <Birck> I'm trying to set up a cloud enviroment with OpenStack. I have a VM running in a machine located in a local network. I'm trying to ssh this VM from the gateway machine, but I'm getting this erro: Permission denied (publickey). I'm using a .priv archive with this command: ssh -i archive.priv ubuntu@10.0.0.2.
[16:19] <Birck> Anyone have a suggestion on how to solve this problem?
[16:26] <lborda> Birck, did you run euca-add-keypair and started the instance using that same keypair?
[16:27] <zul> Birck: is the permissions for the *.priv file correct: 600?
[16:31] <Birck> lborda: yes, i executed this command euca-add-keypair nuvemcpca > nuvemcpca.priv
[16:31] <Birck> lbroda: and created the instance with the same .priv file
[16:32] <Birck> lborda: with this command euca-run-instances  ami-00000002 -k nuvemcpca -t m1.tiny
[16:32] <Birck> zul: yes, i have changed that: chmod 600 nuvemcpca.priv
[16:34] <Birck> I'll try to terminate all the instances and generate a new .priv file and then create new instanves
[16:35] <pmatulis> Birck: well, did you properly place the appropriate public key on the server?
[16:38] <Birck> pmatulis: yes, because i already have a working VM with this key, now i'm just trying another network architecture, with 2 network interfaces in the gateway, one for the external and one for the local network
[16:39] <Birck> pmatulis: I mean i have another machines with the same installation process and the key is working fine
[16:41] <pmatulis> Birck: you must have missed something.  what about the server side permissions on ~/.ssh directory.  anyway, on client use 'ssh -vv ...' and on server you can start it in the foreground for debugging (see man sshd)
[16:43] <Birck> pmatulis: ok, I'll try to generate again the keys. If i can't succed then i will try what you are saying
[16:49] <lborda> Birck, try using another keypair
[16:49] <lborda> Birck, what do you get from euca-console-output?
[16:50] <gary_poster> hallyn, I had one more question for you.  It would be convenient to have lxc-tools (https://github.com/phbaer/lxc-tools/ , as linked by https://help.ubuntu.com/community/LXC#Host_configuration).  In particular, the graceful approach of lxc-halt would be nice for the people we are supporting.  lxc-status would be nice too.  The rest I've looked at are less interesting.  Are these packaged somewhere?  I looked and did
[16:50] <gary_poster>  not find.
[16:54] <hallyn> gary_poster: I don't know about them
[16:54] <gary_poster> hallyn, cool, np thanks
[16:56] <Birck> lborda: i should run ssh before euca-get-console-output ?
[16:56] <Birck> lborda: i didn't find anything important
[16:56] <Birck> lborda: I'm gonna paste bin the output
[16:57] <lborda> Birck, try euca-get-console-output >instance_id>
[16:57] <lborda> Birck, try euca-get-console-output <instance_id>
[16:58] <lborda> Birck, also ssh -vvv will tell more debug info
[16:59] <Birck> lborda: I didn't find anything about ssh on the console output. But look at this messages at the final of the output
[17:00] <Birck> lborda: follow the paste http://pastebin.com/HGisgi8q
[17:01] <lborda> Birck, i think iptables is playing a role in this case... did you say you changed your network config?
[17:03] <Birck> I'm working with a local network so i have to masquerade and enable ip forwarding on my gateway
[17:04] <Birck> then i can ping external network from my local network
[17:07] <GrueMaster> arosales: https://blueprints.launchpad.net/ubuntu/+spec/arm-p-revalidate-arm-server
[17:08] <arosales> thanks GrueMaster
[17:12] <Birck> lborda: that was the onl things i have changed
[17:59] <talntid> anyone know of any tools that can compare a current ubuntu install, to a base ubuntu install, and display the differences between the 2? (like installed software, modified files, etc)
[18:01] <glebaron> wondering if ubuntu 10.04 64bit server is capable of addressing 128gb of RAM. A quick google search did not tell me.
[18:03] <raubvogel> talntid: AFAIK you can list all the installed packages
[18:03] <raubvogel> So, a first idea is to list them in the two machines
[18:15] <RoAkSoAx> smoser: still around?
[18:16] <smoser> here
[18:16] <RoAkSoAx> smoser: bug #887186
[18:16] <RoAkSoAx> smoser: so according to http://www.squid-cache.org/Doc/config/cache_dir/
[18:17] <RoAkSoAx> smoser: the big directory is correctly using the directory
[18:17] <RoAkSoAx> as the 40000 stands for MB's that the dir can extend to
[18:17] <smoser> sure. thats fine.
[18:17] <RoAkSoAx> smoser: so the cache for that dir will grow up until 40000
[18:18] <smoser> but the intent of the 2 lines:
[18:18] <smoser> cache_dir aufs /var/spool/squid3/small 40000 16 256 max-size=40M
[18:18] <smoser> cache_dir aufs /var/spool/squid3/big 40000 16 256
[18:18] <smoser> is that small files (those under 40M) will go in 'small'
[18:18] <smoser> and large files in large
[18:18] <RoAkSoAx> right
[18:18] <smoser> and everything is going in large
[18:19] <RoAkSoAx> smoser: oh I see your point now
[18:19] <RoAkSoAx> smoser: i understood it differently :P
[18:21] <smoser> and, per the doc there it seems that we have ti right... i think
[18:21] <smoser> 	max-size=n, refers to the max object size in bytes this cache_dir
[18:21] <smoser> 	supports.  It is used to select the cache_dir to store the object.
[18:21] <smoser> 	Note: To make optimal use of the max-size limits you should order
[18:21] <smoser> 	the cache_dir lines with the smallest max-size value first and the
[18:21] <smoser> 	ones with no max-size specification last
[18:21] <RoAkSoAx> smoser: yes
[18:21] <smoser>  but experience shows otherwise
[18:22] <RoAkSoAx> smoser: maybe it should be 'cache_dir aufs /var/spool/squid/small 40000 16 256 max-size=40' instead of '... max-size=40M'
[18:24] <RoAkSoAx> or maybe even use 'MB' instead of 'M'
[18:24] <RoAkSoAx> smoser: cause otherwise, it would be a squid error per se
[18:25] <eagles0513875> are there any cobbler users in here?
[18:25] <RoAkSoAx> eagles0513875: yes
[18:25] <eagles0513875> in cobbler where does one specify a systems disk size ram etc
[18:25] <RoAkSoAx> eagles0513875: there's no place where you can do so
[18:25] <eagles0513875> then how do you do that
[18:26] <RoAkSoAx> eagles0513875: well, long story short, you can't just yet
[18:26] <RoAkSoAx> zu^^
[18:26] <RoAkSoAx> err zul not here
[18:26] <eagles0513875> was that for me that last long story short comment
[18:27] <RoAkSoAx> eagles0513875: yeah. Unfortunately we cannot specify the hw specification for a system in cobbler just yet
[18:27] <eagles0513875> RoAkSoAx: can that be done when using koan to provision the guest and get it going
[18:28] <RoAkSoAx> eagles0513875: not that I know of. Are you trying to specify how much disk you want to assign for each partition and stuff like that?
[18:28] <eagles0513875> RoAkSoAx: to this particular xen guest
[18:29] <eagles0513875> i have lvm setup on the system
[18:30] <RoAkSoAx> eagles0513875: if it is for virtualization then you DO can
[18:30] <eagles0513875> do can what
[18:30]  * eagles0513875 waits for RoAkSoAx's rest of the last sentence
[18:31] <eagles0513875> guess ill have to dig up the koan man page
[18:31] <RoAkSoAx> eagles0513875: sudo cobbler system edit --name <whatever> --virt-cpus=<xyz> --virt-ram=<xyz>
[18:32] <eagles0513875> thanks RoAkSoAx
[18:32] <eagles0513875> RoAkSoAx: thats all in the koan man page
[18:32] <RoAkSoAx> eagles0513875: cool then, but that command is done in the cobbler server
[18:32] <RoAkSoAx> I haven't actually tried to do it with koan
[18:33] <eagles0513875> RoAkSoAx: what ive seen with koan is it connects to the cobbler server to setup the guest
[18:33] <RoAkSoAx> eagles0513875: yes
[18:34] <eagles0513875> RoAkSoAx: let me look up the man page and ill let ya know if i can specify the guest requirements using koan instead :)
[18:36] <squidly> I'm trying to install cacti on my server but I get this. "no type given for question at /usr/share/perl5/Debconf/Question.pm line 22, <GEN1> line 2." I cant remove or install cacti
[18:36] <smoser> RoAkSoAx, well, i tried 40, and 400000
[18:37] <smoser> and stuff is still going into 'big'
[18:37] <RoAkSoAx> smoser: did you reconfigure squid?
[18:37] <kirkland> smoser: we're seeing some Ubuntu kernels in Amazon that don't have ecryptfs built in, or as a module: http://thecloudmarket.com/image/aki-8903cbe0--ubuntu-kernels-testing-ubuntu-lucid-amd64-linux-image-2-6-32-340-ec2-v-2-6-32-340-40-kernel
[18:37] <smoser> restart
[18:37] <smoser> yes
[18:38] <RoAkSoAx> smoser: squid3 -k reconfigure
[18:38] <smoser> surely restart re-reads config no?
[18:38] <kirkland> smoser: looks like these: http://thecloudmarket.com/search?search_term=2.6.32-340
[18:38] <RoAkSoAx> smoser: supposed to. I'm recreating the cache dirs and see what happens
[18:40] <smoser> kirkland, so you think that config for linux-ec2 kernels changed through the life of lucid-updates from lucid-release ?
[18:41] <kirkland> smoser: i'm trying to chase down what's changed exactly right now
[18:42] <smoser> kirkland, but fwiw, you almost certainly should not b e using that aki
[18:43] <smoser> you should be using pv-grub kernels
[18:43] <kirkland> smoser: that's a testing kernel, it looks like?
[18:44] <smoser> well, its a kernel that was published for a daily build
[18:44] <kirkland> smoser: okay, looks like those are just broken somehow;  i've confirmed that i can modprobe ecryptfs in 2.6.32-341-ec2
[18:44] <smoser> but ubuntu amis do not use those kernels
[18:44] <smoser> they use the pv-grub ones (after 2010-12 or so)
[18:45] <kirkland> smoser: thanks;  this isn't me, but rather a customer of ours
[18:45] <smoser> well they shouldn't be using that aki
[18:45] <smoser> for many reasons
[18:45] <smoser> a.) they should use pv-grub
[18:46] <smoser> b.) its listed as "testing", meaning it might "fall off" after no dailies use it
[18:46] <smoser> c.) its not even current
[18:46] <hallyn> rbasak: I've got another update to that debdiff I'm afraid :)  (to fix lxc-ls)
[18:50] <RoAkSoAx> smoser: i'm starting to think this is a bug in squid3
[18:51] <smoser> well, it was also resent in squid 2
[18:51] <RoAkSoAx> smoser: FATAL: Bungled squid.conf line 54: cache_dir aufs /var/spool/squid3/big 40000 16 256 min-size=41M
[18:51] <RoAkSoAx> smoser: I get that error if I try to restrict the min-size for that cache dir
[18:51] <smoser> hm..
[18:53] <RoAkSoAx> smoser: oh I see, maybe because of this:
[18:53] <RoAkSoAx> Changes in 3.2 cache_dir
[18:53] <RoAkSoAx> min-size option ported from Squid-2
[18:54] <RoAkSoAx> smoser: this strengthens my idea for which this is a squid bug
[19:05] <peppe9O> !list
[19:10] <eagles0513875> got a question for you guys what is the best way to change the default kernel grub 2 boots
[19:10] <eagles0513875> by default
[19:11] <RoAkSoAx> eagles0513875: /etc/default/grub
[19:11] <RoAkSoAx> eagles0513875: GRUB_DEFAULT=0 -> change the number in that option
[19:11] <RoAkSoAx> then update-grub
[19:11] <eagles0513875> ok here is something odd im not seeing a list of kernels available
[19:11] <RoAkSoAx> q!
[19:12] <RoAkSoAx> eagles0513875: maybe take a look at cat /boot/grub/grub.cfg | grep menuitem
[19:12] <eagles0513875> ok
[19:12] <RoAkSoAx> eagles0513875: maybe take a look at cat /boot/grub/grub.cfg | grep menuentry
[19:13] <eagles0513875> and they start at entry 0 onwards right
[19:13] <RoAkSoAx> eagles0513875: correct
[19:14] <eagles0513875> thanks :) that should solve the headache of trying to kvm to get to grub
[19:14] <eagles0513875> im at a total loss RoAkSoAx whats the best way when working with cobbler and koan to specify ram hard disk etc for a xen guest
[19:15] <RoAkSoAx> eagles0513875: go to the cobbler server, and edit the virt-* settings for the system
[19:15] <talntid> anyone know of any tools that can compare a current ubuntu install, to a base ubuntu install, and display the differences between the 2? (like installed software, modified files, etc)
[19:15] <RoAkSoAx> eagles0513875: or, from thje cobbler server, do: sudo cobbler edit --name <abcd> --virt-cpus <etc> --etc-etc-etc
[19:17] <pehden> i am on some random persons wifi
[19:17] <pehden> lol
[19:17] <eagles0513875> for modified source files there is diff not sure about compile stuff
[19:17] <RoAkSoAx> talntid: idk how to compare exactly, but you can get all the installed packages with sudo dpkg --get-selections
[19:17] <pehden> and cant use ff or it would drop the connection
[19:18] <RoAkSoAx> eagles0513875: or, from the cobbler server, do: sudo cobbler edit --name <abcd> --virt-cpus <etc> --etc-etc-etc
[19:19] <eagles0513875> RoAkSoAx: how come this cannot be done from the web interface
[19:19] <RoAkSoAx> eagles0513875: this *can*
[19:19] <eagles0513875> how can it be done from the web interface O_o
[19:19] <RoAkSoAx> eagles0513875: *edit* a *system* and go to "Virtualization"
[19:20] <RoAkSoAx> eagles0513875: where most of it should show "<<inherit>>", bnut you can just go ahead and change it
[19:20] <eagles0513875> ahh file size
[19:20] <eagles0513875> is supposed to be hard disk
[19:20] <eagles0513875> and will this work with lvm
[19:21] <RoAkSoAx> eagles0513875: I guess it should, haven't used xen
[19:21] <eagles0513875> does cobbler and koan work for provisioning lvm based xen guests
[19:21] <eagles0513875> yes it does actually :)
[19:21] <RoAkSoAx> eagles0513875: it should
[19:22] <hallyn> rbasak: http://people.canonical.com/~serge/lxc-rally-p6.debdiff
[19:22] <eagles0513875> viva la little ? marks on the side of the text boxes
[19:29] <eagles0513875> RoAkSoAx: in cobbler whats the difference between xenpv and xenfv
[19:47] <smoser> eagles0513875, i'm guessing its pvm and hvm
[19:47] <smoser> para-virtualization hand hardware virtualization
[19:47] <eagles0513875> smoser: all it says in cobbler xenpv and xenFV
[19:47] <smoser> its almost certainly pvm and hvm
[19:47] <eagles0513875> found out xenpv = paravirtualization and fv = full virtualization
[19:47] <eagles0513875> ya paravirtualization is enough for what i need
[20:06] <eagles0513875> hey guys i need some help with lvm
[20:07] <eagles0513875> i have an extened partition which is setup as linux lvm
[20:07] <eagles0513875> does that mean i need to add it to a volume group and what do i do if one doesnt exist
[20:16] <RoAkSoAx> smoser: btw.. why aren't you enjoying budapest?
[20:16] <smoser> dont know.
[20:16] <smoser> felt like owking tonight.
[20:16] <smoser> i'm poking at juju local provider and openstack. trying to apply adams' work and get it functional
[20:17] <smoser> did you test my odev ?
[20:17] <RoAkSoAx> smoser: i'm about to start doing that
[20:17] <RoAkSoAx> just got back from lunch
[20:18] <RoAkSoAx> smoser: this is based on the cobbler-devenv right?
[20:19] <smoser> yeah.
[20:19] <smoser> reanmed to odev
[20:19] <smoser> probably to be renamed to maasdev
[20:19] <smoser> ;)
[20:20] <RoAkSoAx> hehe
[20:25] <mgw> I'm trying to figure out how to push multi-homed routes via dhcp (we have several VLANs on each system)…. DHCP OPTION 121 (RFC1519) doesn't seem to me to be sufficient — any pointers?
[20:26] <eagles0513875> hey guys how can i check to see if xen is compiled into the kernel
[20:33] <smoser> eagles0513875, grep CONFIG_XEN /boot/config-$(uname -r)
[20:34] <eagles0513875> smoser: i forgot to hit update-grub after changing the default kernel to boot from
[20:34] <ChmEarl> eagles0513875, oneiric has xen inside with kernel 3.0+
[20:36] <eagles0513875> ChmEarl: i know but you still have to boot onto a special xen kernel from the looks of it
[20:37] <eagles0513875> ugh
[20:37] <eagles0513875> ok its not a problem with ubuntu
[20:38] <eagles0513875> anyone here used koan in conjunction with cobbler
[21:05] <roaksoax> smoser: still around?
[21:19] <eagles0513875> any koan and cobbler users online atm here
[21:20] <smoser> roaksoax, here
[21:20] <chrislabeard> when your running your own dns and virtual servers do the domains point at the public ip or the local ip?
[21:20] <chrislabeard> I'm alittle confused about that
[21:21] <eagles0513875> public ip chrislabeard
[21:22] <eagles0513875> private ips are never seen by the external network
[21:24] <smoser> adam_g, around ?
[21:25] <chrislabeard> eagles0513875: so if you were to dig wwmcd.org it would show the external IP of the server.
[21:25] <roaksoax> smoser: so how can I log into zimmer?
[21:25] <roaksoax> smoser: doesn't accept my ssh key, nor passw0rd works
[21:26] <smoser> ubuntu@192.168.123.2 should take 'passw0rd'
[21:26] <eagles0513875> chrislabeard: correct
[21:26] <eagles0513875> chrislabeard: yoru server running a 99 ip
[21:26] <chrislabeard> eagles0513875: yeah it is
[21:26] <roaksoax> smoser: does not
[21:27] <roaksoax> ssh ubuntu@192.168.123.2
[21:27] <roaksoax> Permission denied (publickey).
[21:27] <eagles0513875> :) chrislabeard besides dig you can also use nslookup and the domain
[21:27] <chrislabeard> eagles0513875: its weird when I change it to the public ip I can't get the subdomains to work
[21:28] <smoser> roaksoax, i dont know. you're sure its going to zimmer there ?
[21:28] <roaksoax> smoser: yes
[21:28] <smoser> maybe something went wrong in the building of it.
[21:28] <smoser> can i see your build-console.log
[21:28] <smoser> (or whatever that is called)
[21:28] <roaksoax> smoser: yeah, cause can't even use a recovery console given that it asks for root pass
[21:30] <smoser> roaksoax, can you pastebin the log?
[21:32] <roaksoax> smoser http://pastebin.ubuntu.com/799870/
[21:32] <eagles0513875> chrislabeard: you need to create another a dns entry for the subdomain pointing to the 99 ip address
[21:32] <chrislabeard> eagles0513875: yeah I have it in there if I dig test.wwmcd.org
[21:32] <chrislabeard> it shows up with the public ip
[21:33] <smoser> roaksoax, i think it didn't execute anything
[21:33] <smoser> the cloud-init creation of the iso failed
[21:34] <smoser> or something went wrong there at least
[21:35] <chrislabeard> eagles0513875: welp looks like it hasn't propagated yet , I'll wait lol
[21:35] <roaksoax> smoser: $ ( cd cloud-init/doc/ovf/ && ./make-iso ovf-env.xml.tmpl ../../../ud-build.txt -o - ) > build.iso
[21:35] <roaksoax> wrote iso to -
[21:35] <smoser> yeah, something went wrong there.
[21:35] <eagles0513875> chrislabeard:  you have done everythign right
[21:36] <roaksoax> smoser: redoing it causes the same thing
[21:41] <smoser> roaksoax, isoinfo -i build.iso  -x "/OVF_ENV.XML;1" | pastebinit
[21:41] <roaksoax> smoser: too late, I did this: ./make-iso ovf-env.xml.tmpl ../../../ud-build.txt -o ../../../build.iso from cloud-init/doc/ovf and now it was created correctly
[21:43] <smoser> i dont know what could have gone wrong there.
[21:43] <smoser> roaksoax, note, that you can in ud-build.txt set your ssh_import_id
[21:43] <smoser> and also a local proxy
[21:43] <smoser> (or other cloud-init stuff)
[21:44] <roaksoax> smoser: yeah
[21:45] <smoser> and this is what lead me to want to proxy in cobbler-orchestra
[21:45] <smoser> so i could use one from outside the libvirt network (my existing squid proxy)
[21:46] <John-_> Hello there
[21:47] <John-_> I'm trying to set a file to only allow access to one user
[21:48] <John-_> This is obviously easy in Windows but having trouble doing this on a ubuntu server
[21:48] <smoser> John-_, the user 'thatusername' is probably in a group called 'thatusername'
[21:49] <smoser> if thats the case, you can
[21:49] <smoser> chgrp thatusername the-filename
[21:49] <smoser> chmod 440 the-file
[21:51] <John-_> Yeah, I mean, I'm running Samba as a file and print server.. If 7 people log on and all try to change the same file we run into problems. It would be easy to just set only one user access at one given time but let the hole 7 users roam and see files
[21:51] <John-_> does that make sence ?
[21:52] <smoser> sorry. can't help with samba
[21:52] <John-_> This is at the base level though
[21:53] <roaksoax> smoser: right, yeah the proxy thing Is something I've been thinking about for a while. But, the easiest way to do it is having another debconf question
[21:54] <smoser> well..
[21:54] <smoser> you dont need debconf questions for everything
[21:54] <smoser> is there some place we can add cobbler kickstart variables ?
[21:54] <smoser> that would get read?
[21:54] <smoser> adn then we just use them (with defaults) from the other snippets
[21:54] <roaksoax> smoser: yes, but it is an initial config when setting up orchestra (i.e. WE can installa  proxy for you if you don't have one already)
[21:55] <smoser> i dont think that is all that important really.
[21:56] <roaksoax> smoser: i guess we could add some default kickstart variables, but then, we would have to make them configurable
[21:56] <smoser> orchestra-provisioing-server owuld just install one, and it go unused
[21:56] <smoser> if the gave you a different variable
[21:56] <smoser> even just as files in /etc/kickstart/
[21:56] <smoser> that were read and the user could overwrite
[21:56] <roaksoax> right, yeah I think there's still thinkgs to consider about the snippets and stuff
[21:57] <mgw> what controls the order in which the scripts in /etc/dhcp/dhclient-exit-hooks.d/ are run?
[21:59] <John-_> Would it be possible to set a script in a user dir that, if file4 opened in dir 1 lock file4 then when file4 closed unlock file4
[22:05] <John-_> Help, windows is going to win over one simple function !
[22:18] <erichammond> If it were possible to run a standard PV (non-HVM) Ubuntu AMI on EC2's cluster compute instance type (cc1.4xlarge) would there be any dangers or expected problems with this combination?
[22:36] <Patrickdk> not sure what you mean
[22:36] <Patrickdk> but I run normal ubuntu ec2 ami's on xen pv's on my own cluster
[23:40] <jason> I was wondering if anyone has used GAMIN package before and if it allows notifications via email when changes occur?