[00:26] <MTecknology> How can I resize a volume group to match the size of an increased physical volume?
[00:35] <rockets> Nothing makes server maintenance go more quickly like joy division
[00:37] <twb> MTecknology: pvresize?
[00:37] <MTecknology> twb: I did pvresize to get that bigger, but now I can't get the volume group to see that it's bigger
[00:38] <MTecknology> frick
[00:38] <MTecknology> twb: i was looking at the wrong volume group.... sorry
[00:56] <seas> Is there a way to get gfs2 working alongside pacemaker in 12.04? dlm_controld.pcmk and gfs_controld.pcmk seem to be gone and pacemaker refuses to start with cman+corosync (https://bugs.launchpad.net/ubuntu/+source/pacemaker/+bug/887165)
[01:06] <marcoceppi> Is there a 12.04 server install out yet? Or will that land when 12.04 goes live?
[01:26] <twb> You can do one at any time using auto-built installer media
[01:26] <twb> Obviously until 12.04 is released, it will only be a alpha or beta install
[01:27] <twb> e.g. http://archive.ubuntu.com/ubuntu/dists/precise/main/installer-amd64/current/images/netboot/mini.iso
[01:28] <twb> Note that's effectively the same as the "alternate" installer, the main difference between the alternate and server install media is what tasksel prompts you for, whether the language-pack packages are installed, and what packages are cached on the CD.  The last of these doesn't apply for the mini ISOs.
[01:28] <adam_g> zul: ping
[01:32] <pabelanger> marcoceppi: not sure what you mean, 12.04-beta1 is out. For desktop / server / etc
[01:32] <marcoceppi> pabelanger: I can't find the ISO for it, I guess it was just a lazy way of saying "Where can I grab a beta ISO" other than install 11.10 and upgrade -d
[01:33] <pabelanger> marcoceppi: https://wiki.ubuntu.com/PrecisePangolin/TechnicalOverview/Beta1
[01:33] <pabelanger> links listed
[01:33] <Luisinho> PESSOAL EU QUERO APRENDER A NUKAR UM SERVIDOR OQUE EU FAÇO ?
[01:33] <marcoceppi> pabelanger: Thanks o/
[01:34] <twb> !pt
[01:36] <marcoceppi> So, I've never done a bare server install. I've got a HP Proliant that I'm going to be using to run OpenStack on. I'm assuming it's just something as simple as "Install Ubuntu Server, Install OpenStack, Profit!"?
[01:37] <pabelanger> marcoceppi: you forgot step 3... ???
[01:37] <pabelanger> :)
[01:37] <pabelanger> but basically, yes
[01:38] <marcoceppi> But it's so easy, you don't need a step 3 \o/
[01:39] <twb> Installing a base Ubuntu Server is reasonably straightforward.  The main consideration is how to partition / raid / lvm your disks.
[01:39] <marcoceppi> Well, I take that back. I've done a few Ubuntu Server installs, but they were on smaller machiens
[01:40] <marcoceppi> This is the first time we'll be racking something this big with the express purpose of virtualzing machines
[01:40] <twb> IIRC proliant's aren't exactly big iron
[01:40] <qman__> hardest part is if you care about HP's management software
[01:40] <qman__> have to source debs for it
[01:41] <qman__> otherwise it's just the normal deal
[01:41] <marcoceppi> I don't particularly care for that
[01:42] <qman__> only reason I do is I have to be in the same room as it when I use it, and that's the only way to slow the fans down
[01:42] <qman__> but mine's an old one
[01:42] <qman__> newer ones that may not be the case
[01:43] <qman__> mine's a dual P3 model, and the HP software brings it from 'hearing damage' to 'vacuum cleaner'
[01:43] <marcoceppi> twb: Sorry, 8 cores and 20GB of RAM are big for us
[01:43] <marcoceppi> :)
[01:43] <twb> marcoceppi: it's still just a beefy x86 workstation, architecture-wise
[01:43] <marcoceppi> qman__: This will be racked in a DC soon, so I don't have to worry about noise too much
[01:44] <qman__> yeah
[01:44] <adam_g> zul:  i sent two packaging merges to ubuntu-server-dev for keystone + nova. i'd take a look at keystone, i dont think your fix resolved the uninstall bug.  i have nothing propose for other packages. i think glance is mostly quiet other than the debian merge stuff, which hasn't been merged and tested in -propsed yet
[08:11] <linocisco> hi anybody
[08:12] <linocisco> anybody who has configured server and clients on vbox as guest?
[08:19] <linocisco> my squid server on guest never work for my client on Guest
[08:31] <linocisco> hi i have followed all squid documentation, I was never ok
[09:00] <Daviey> Gooooood Morning people!
[09:01] <linocisco> hi
[09:01] <linocisco> all
[09:02] <linocisco> who could help me with squid?
[09:05] <ikonia> linocisco: depends what you want/need
[09:06] <linocisco> i just want pure squid first with the need of client to point squid's server IP and port 3128, before I can test transparent squid.
[09:06] <lynxman> morning o/
[09:06] <linocisco> ikonia, I have two NICs on ubuntu server
[09:06] <linocisco> ikonia,  one is for WAN and one is for LAN
[09:06] <lynxman> Daviey: does bug #950183 deserve some attention? afaict the package shouldn't gracefully recover from a previous config misconfiguration right?
[09:10] <ikonia> linocisco: ok, so what are you stuck with ?
[09:12] <Daviey> lynxman: If you want to prepare a debdiff, i'll certainly sponsor
[09:12] <Daviey> (precise)
[09:12] <Daviey> lynxman: Can you take the lead in helpig to close the delta with Debian?
[09:12] <lynxman> Daviey: there's already a merge request for another bug and the puppetlabs merge
[09:12] <lynxman> Daviey: sure! I'll be glad to do so
[09:13] <Daviey> \o/
[09:13] <Daviey> lynxman: So, point me to the puppet merge!
[09:13] <lynxman> Daviey: was just wondering if this one merited patching or not, since it parts from a broken premise
[09:13] <lynxman> Daviey: https://code.launchpad.net/~lynxman/ubuntu/precise/puppet/puppetlabsfixbug12844/+merge/96391
[09:14] <Daviey> lynxman: Right, but the issue is that it's not exit 1'ng on failure
[09:14] <Daviey> right?
[09:14] <lynxman> Daviey: hmm I reckon so
[09:15] <Daviey> Well it *should* exit 1 on failure :)
[09:15] <Daviey> or at least non-0
[09:15] <lynxman> Daviey: fair enough :)
[09:15] <lynxman> Daviey: I'll have a look after my GP visit
[09:15] <linocisco> iknon, i never got internet page
[09:15] <Daviey> lynxman: the package maintainer scripts will barf if a command exit's non-zero and mark the package as not configured
[09:16] <Daviey> lynxman: Have fun :)
[09:16] <lynxman> Daviey: heh, thanks :)
[09:20] <Gallomimia> can someone quickly remind me how i should install the tool to do-release-update?
[09:32] <koolhead17> hi all
[09:36] <Daviey> Gallomimia: update-manager-core
[09:41] <uksysadmin> morning koolhead17
[09:41] <uksysadmin> (and others)
[09:41] <koolhead17> hey  uksysadmin. wassup
[09:42] <uksysadmin> yeah good. just doing the usual: install ubu, run install of OpenStack, rinse and repeat.
[09:42] <uksysadmin> and decided that I'm actually fed up of doing the basics of that and decided on looking at Orchestra again
[09:43] <uksysadmin> so may actually get around that that promise of that partitioning ;-)
[09:43] <eagles0513875> ikonia: are you around im still having issues doing a net installation for some reason
[09:44] <ikonia> I am here yes
[09:51] <KM0201> when you're installing ubuntu server 11.10, what is the "ubuntu-desktop-usb" package that you can select to install? (along w/ openssh, samba, etc..)
[09:52] <koolhead17> uksysadmin: adam_g answered some of your queries i suppose :)
[09:52] <lynxman> Daviey: already fixed the postinst script, it's available for review at the merge, if you need a debdiff otherwise let me know :)
[09:52] <Daviey> lynxman: do you have a debdiff handy?
[09:53] <lynxman> Daviey: hmm let me prepare one quickly
[09:58] <iclebyte> i've been searching for days and I cannot find a single concise document on how to create your own SIGNED repo. I've built repo's which work fine using various methods but I need it signed so that I can install custom packages via puppet without them failing. Can anyone point me in the right direction?
[10:01] <lynxman> Daviey: debdiff attached to bug #950183
[10:01] <lynxman> Daviey: also there's a branch merge with this fixes for bzr extra goodness
[10:01]  * lynxman runs to the doc
[10:02] <Daviey> thanks lynxman
[10:03] <Daviey> lynxman: your debdiff includes .bzr :o
[10:03] <iclebyte> nothing?
[10:04] <Daviey> iclebyte: reprepro
[10:04] <Daviey> It's the most elegant small apt repo manager, includes built in signing support
[10:04] <iclebyte> i've built it using that.
[10:05] <Daviey> great
[10:05] <iclebyte> when I go apt-get update on the other node it says 'The following signatures couldn't be verified because the public key is no avilable: NO_PUBKEY A95723464'
[10:06] <_ruben> iclebyte: i sign my internal repo manually, and have a post-installation script that imports its key into apt
[10:06] <Daviey> Well, that is a different issue iclebyte
[10:06] <iclebyte> _ruben, that is exactly what I have been looking for
[10:06] <Daviey> iclebyte: you need to import the key.
[10:06] <iclebyte> how?
[10:06] <iclebyte> with apt-key some how?
[10:06] <_ruben> yes
[10:07] <twb> 21:07 <dpkg> gpg --keyserver keys.gnupg.net --recv-keys import deadbeef ; gpg --export import deadbeef | sudo apt-key add -
[10:07] <Daviey> hmm
[10:07] <_ruben> $WGET -q -O - $KEYURL | apt-key add -
[10:07] <iclebyte> apt-get add 291F3DF9 ?
[10:07] <Daviey> Or... wget htp://path/to/key/on/mirror - | sudo apt-key add -
[10:07] <twb> (It would be nice if ubottu had more of dpkg bot's entries)
[10:08] <Daviey> (you won't need the sudo if you run it as a post_install option)
[10:08] <twb> Personally I use apt-ftparchive because reprorepo was overkill and also hard to type three times fast
[10:09] <Daviey> twb: Yeah, but you must admit that rerepro is more gentlemanly
[10:09] <Daviey> (although, didn't support udeb's OOTB last time i checked.)
[10:09] <_ruben> i use mini-dinstall myself
[10:09] <twb> Well, sure, it's all slicked up, and not like us cowboys
[10:10] <iclebyte> okay, lets go the download the key from my own repo path. I need to export the key to somewhere avilable by the webserver. using gpg --export > file.gpg ? ?
[10:11] <iclebyte> okay got it exported. gpg --export -a > mykey.gpg
[10:12] <iclebyte> wooohoo! =)
[10:12] <Daviey> cool
[10:12] <iclebyte> I couldn't find anything which explained the process that simply.
[10:13] <iclebyte> so I just need to use puppet to download and import that key on each node then add my custom repo to /etc/apt/sources.list and i'm cooking with gas.
[10:13] <Daviey> iclebyte: Fancy writing up a how-to, and we'll include it in the server guide?
[10:13] <iclebyte> Daviey, yea okay. I'll have to document it today anyway
[10:14] <Daviey> super!
[10:15] <iclebyte> I'm interested, how can you manually sign the repo? i.e. if I didn't want to use reprepro and I'd created my Packages.gz manually using 'dpkg-scanpackages . /dev/null | gzip -9c > Packages.gz' ?
[10:17] <Daviey> it's just signed..
[10:17] <twb> http://paste.debian.net/159093/
[10:19] <Daviey> twb: why use that?
[10:19] <iclebyte> but which file do you sign? Packages.gz ?
[10:19] <twb> Daviey: he asked "how do you do it manually"
[10:19] <twb> That's how I do it manually
[10:19] <twb> iclebyte: Release
[10:19] <iclebyte> okay. thanks.
[10:19] <Daviey> twb: you are nuts :)
[10:19] <twb> iclebyte: look at the pastebin
[10:19] <twb> Daviey: I was lazy
[10:20] <twb> Daviey: doing it by hand is straightforward; learning to use a "helpful" tool was going to take all day
[10:20] <twb> When I have >>1 arch or >>1 site I will probably bother
[10:20] <Daviey> twb: right.. but i don't grok why people would us apt-ftparchive
[10:20] <twb> Because it's simple and obvious and easy to spell
[10:21] <Daviey> twb: do you have to support multiple releases?
[10:21] <lynxman> Daviey: reuploaded, sorry, I'm in a hurry to get to the GP (I'm late already)
[10:21] <twb> Daviey: nope
[10:21] <Daviey> lynxman: eeek
[10:21] <Daviey> twb: ah, ok - no real benefit from a pool i guess.
[10:21] <twb> Daviey: that cron job literally just adds some apt metadata to a single dir, no pool/ even
[10:21] <Daviey> right
[10:22] <twb> http://paste.debian.net/159094/
[10:22] <twb> That's what's in there atm
[10:23] <Daviey> twb: i'd love to see the diff's yo apply to ubuntu packages.
[10:23] <twb> the ldap one just uses openssl because TLS + ppolicy is broken
[10:24] <eagles0513875> ikonia: for some reason citrix xen server is giving me issues even when specifying what version to pull the necessary files from
[10:24] <Daviey> twb: ahh
[10:24] <ikonia> you need to explain more than that
[10:24] <eagles0513875> ikonia: in the process of getting the error msg
[10:24] <twb> The mutt one basically tells mutt not to believe dovecot when it says the folder is read-only, because it *is* read only, but it will still remember read vs. not read
[10:24] <ikonia> eagles0513875: are you using a cyrtix xen server, or just xen
[10:24] <eagles0513875> citrix xen server the free license
[10:25] <twb> ikonia: here I have slapo-ppolicy(5) enabled and require all LDAP traffic to go over the ldaps port.  When linked with gnutls, it just silently fails to work, when linked with openssl it works fine.
[10:26] <twb> I can't remember which bit breaks specifically
[10:26] <ikonia> twb: sorry, have we been discussing this before ? I have no context in what you are talking about / why you are singling me out
[10:27] <twb> Ah, sudo-ldap doesn't work at all, and using passwd to reset your own password completely fails (over the wire w/exop)
[10:27] <twb> ikonia: I must have misread 21:24 <ikonia> you need to explain more than that
[10:27] <ikonia> twb: sorry, was talking to eagles
[10:27] <twb> Ah, I'm ignoring him, sorry about the noise
[10:27] <ikonia> twb: was lacking context on your problem, hence surprised on your question to me
[10:27] <eagles0513875> ikonia: yes this is in relation to yesterday when i asked you about creating a vm and the url i should point it too
[10:28] <ikonia> eagles0513875: again, but are you using the cytrix out of the box product or the linux xen components
[10:29] <eagles0513875> ikonia: citrix out of the box
[10:30] <ikonia> eagles0513875: comercial product - not ubuntu
[10:30] <ikonia> eagles0513875: contact cirtix
[10:30] <eagles0513875> ok
[10:30] <ikonia> eagles0513875: I'm assumign it's xenserver your using ?
[10:31] <eagles0513875> correct
[10:31] <ikonia> yeah, it's got excellent comercial support, if you've bought it they will really hold your hand and walk you through things
[10:32] <ikonia> it's very very well supported by cirtix, even for quite complex "away from the norm" type situations, you'll have no problem with them
[11:16] <ikonia> that's a pretty obvious bug
[11:16] <ikonia> if you are using nis maps for auto mounter, than of course nis must be started and bounc
[11:16] <ikonia> bound
[11:42] <Dioxin> is it possible to get a little assistance with PXEBooting an Ubuntu-Server?
[11:43] <Dioxin> I'd like to be able to boot almost completely disklessly
[11:43] <ikonia> ok
[11:44] <Dioxin> I have PXEBoot configured already, the issue I have is that I'm not sure how to generate the file system correctly to boot the way I'm intending
[11:44] <Dioxin> my intention is to have no dependence on the network once its booted
[11:45] <Dioxin> and the target system wont have any local disk, so I'm hoping to use some form of RAMDisk for system files
[11:46] <ikonia> Dioxin: have you looked at the ltsp project
[11:46] <ikonia> Dioxin: while this isn't exactly what you want, it does have good docs on creating rambased disk installs
[11:46] <Dioxin> I think so, let me pop off and remind myself, one second
[11:47] <koolhead17> !ltsp
[11:48] <koolhead17> Dioxin: :P
[11:49] <ikonia> Dioxin: it's not the same process, but the concept of booting into ram is in their docs
[11:50] <ikonia> Dioxin: hopefully enough to get you moving, then we can look at specific things
[11:51] <Dioxin> cheers, I'm just bouncing round the wiki
[11:52] <ikonia> Dioxin: or we can just dive right in, I've not done %100 diskless ram machines for a while, maybe intersting
[11:52] <Dioxin> :)
[11:52] <Dioxin> let me just get some of the issue in my head, so I don't have a brainfart
[11:54] <ikonia> they key part is making the boot image as I recall
[11:54] <Dioxin> this is the precise point I'm stuck on ;)
[11:55] <ikonia> you need a specific image though, or you are just going to user the installer desktop
[11:55] <ikonia> I made an install, then made an image of it, then booted that image
[11:55] <ikonia> the cd installer wasn't good for that functionality
[11:55] <ikonia> (from memory)
[11:57] <koolhead17> Dioxin: and once your done, make sure to update the wiki in case you find t outdated!! :P
[11:57] <Dioxin> not a specific image per se.
[11:58] <Dioxin> the LTSP appears to use a specific image for the thin-client end, and doesnt seem to point to anything regards a custom one
[11:59] <Dioxin> I'm intending to only access the target system via ssh or some such (diskless and headless ;) )
[11:59] <Dioxin> but I've got it hooked up to a monitor and keyboard for now
[12:01] <Dioxin> would it make sense to boot from a CD, install to NFS on the PXE Server, and then amend the root file system to load itself onto the Ram Disk?
[12:01] <ikonia> that's an interesting idea
[12:01] <ikonia> I'd be interested in how that works out for you
[12:02] <ikonia> could you not just mount the NFS server AS the root file system on th client
[12:02] <ikonia> that is in essense running in rmdisk
[12:02] <Dioxin> but let just say that I unplugged the network, wont the system drop out?
[12:02] <ikonia> yes
[12:03] <ikonia> how much ram does your client have ?
[12:03] <Dioxin> 8 gigs
[12:03] <ikonia> so you should be able to hold an uncompressed file system quite comfortably
[12:03] <ikonia> I think the CD is 2.4 gig when uncompressed
[12:04] <Dioxin> the plan is once I get one running system, to twink to reduce the footprint
[12:04] <Dioxin> tweek*
[12:04] <ikonia> Dioxin: I like the idea of installing it to NFS, tune the install, then make a boot image from that
[12:05] <Dioxin> is it trivial to install to NFS from the Server Boot Disk?
[12:05] <ikonia> should be, it's just going to ask you for target file systems
[12:05] <ikonia> although.....I'm wondering if it won't like it as it's not /dev/something
[12:06] <Dioxin> only one way to find out ;)
[12:06] <ikonia> good attitude
[12:07] <Dioxin> ok I'm going to pop away for a short while to burn a disc and see if I can install to NFS
[12:07] <ikonia> Dioxin: I'll try to dig out some notes
[12:16] <KM0201> whats the command to reconfigure your sources list, so you can choose a different mirror?
[12:18] <Dioxin> ikonia: just creating a Ubuntu Server USB stick
[12:19] <KM0201> .
[12:20] <Dioxin> KM0201: not sure of the command but I think you can access them from /etc/apt/sources.list
[12:20] <KM0201> Dioxin: i know that, but then i have to retype it all, don't want to do that
[12:20] <KM0201> there's a way to reconfigure it (you know how when you were installing, and you chose your repository mirrors)
[12:21] <KM0201> so you can choose another mirror
[12:24] <Dioxin> KM0201: http://repogen.simplylinux.ch/index.php ;) this at least saves the typing ;)
[12:24] <KM0201> i don't care about the typing
[12:24] <KM0201> i just want the command
[12:29] <Dioxin> ikonia iSCSI seems to be the only option
[12:33] <ikonia> Dioxin: in what respect
[12:34] <Dioxin> I've booted the server iso, and gone through the install steps
[12:34] <Dioxin> I'm at the setup disks part
[12:35] <Dioxin> and the only option is iSCSI (unless I should have set some tags earlier)
[12:35] <ikonia> ah, so it hits a problem at the target file system install point
[12:35] <Dioxin> yes
[12:35] <Dioxin> should I execute a shell and somehow mount the NFS?
[12:35] <ikonia> Dioxin: suggestion, do it in a VM, then export/copy the file systems off
[12:36] <ikonia> nah, I don't think that will work
[12:37] <Dioxin> it is possible to mount a VM image to copy from it?
[12:38] <ikonia> not directly, I meant do the install, then copy the file systems off it
[12:39] <Dioxin> I'm not 100% sure how its best to approach that part
[12:39] <Dioxin> install the server edition and just copy the rfs from the running VM?
[12:40] <ikonia> Dioxin: not running, but certainly from a VM
[12:40] <ikonia> Dioxin: eg: build  VM, install, then boot the VM from the iso so the file systems are not in use, then copy the file systems out of the VM into directories/NFS mounts to make a "boot" image for your diskless clients
[12:40] <ikonia> follow ?
[12:40] <Dioxin> think so
[12:44] <smb> Daviey, Is upgrade testing for orchestra installations from O to P something that is on your radar?
[12:45] <Daviey> smb: sort of, depends what the subject is.
[12:45] <smb> Daviey, subject it sucks? ;-P
[12:46] <Daviey> smb: I mean, /what/ are you testing? :)
[12:47] <smb> Daviey, basically standard install on oneiric then upgrading to precise. One thing that seems consistent is that cobbler_web is in rc state.
[12:49] <smb> Then I think whatever squid configuration is set up in O , it did not seem to move to squid3 in P
[12:49] <KM0201> does anyone know the command to reconfigure your source list?... so you can choose a different mirror (rather than retyping your entire source list)
[12:49] <KM0201> i was thinking it was dpkg-reconfigure .... then i draw a blank
[12:50] <twb> KM0201: there's no standard one
[12:50] <twb> You can use mirror:// urls in your sources.list
[12:50] <twb> Or there are older tools like apt-spy, but I don't think they work for ubuntu
[12:50] <KM0201> hmm
[12:50] <KM0201> i was thinking there was something like dpkg-reconfigure... then you could choose a new source.
[12:50] <KM0201> oh well, no big deal
[12:51] <twb> if you use mirror method it gives you a public one based on your geoip
[12:51] <KM0201> ok
[12:51] <Daviey> smb: Yes, the squid3 issue is known
[12:51] <Daviey> adam_g is tackling that
[12:56] <glenn___> lyxnman: is there something that i can do to perhaps speed up the process for the latest puppet bug fixes so it will be available on the official repos? perhaps test packages?
[12:56] <glenn___> lyxnman: or is this something i should just wait for patiently? :)
[12:58] <ubusnoob> hello ubuntu server newbie looking for some help here with postfix!
[13:11] <Dioxin> if I wanted to cp a root filesystem do I need any special modifiers?
[13:12] <Dioxin> I dont wish to screw up the file owners
[13:13] <ubusnoob> How can I configure postfix to act as a pop server in a small lan with windows clients only with no interenet connection? Just a small lan for ECDL courses.
[13:17] <xranby> Dioxin: in short: use cp -a   it will preserve the premissions.
[13:17] <ikonia> Dioxin: arp
[13:17] <ikonia> cp -arp
[13:17] <Dioxin> thanks, next question, how do I write permissions to my NFS :(
[13:18] <ikonia> same as a normal file system
[13:19] <Dioxin> my nfs mount is complaining that its a read only file system
[13:20] <xranby> Dioxin: on your nfs server edit the /etc/expots file
[13:21] <seas> anyone here who got pacemaker+gfs2 working in 12.04? cannot figure out how to because dlm_controld.pcmk and gfs_controld.pcmk got removed and pacemaker doesn't support running alongside cman (yet?)
[13:21] <Dioxin> xranby: cheers... it was set as ro :(
[13:25] <Dioxin> xranby: ok I changed the exports to rw, remounted and restarted nfs but still no permissions
[13:25] <xranby> Dioxin: you are tackling a differnet issue here
[13:26] <xranby> the nfs server simply refuses you to create some permissions
[13:26] <xranby> it would have worked if you did the copy to a harddrive under controll by the user who did the copy
[13:26] <Dioxin> permission denied on the write
[13:26] <Dioxin> let me see if I can remove the hdd file permissinos
[13:27] <ikonia> whoaaa
[13:27] <ikonia> Dioxin: stop stop
[13:27] <ikonia> Dioxin: two seperate issues
[13:27] <ikonia> Dioxin: 1.) fix your NFS permissions
[13:27] <ikonia> Dioxin: 2.) deal with your pxe image creation
[13:27] <Dioxin> drwxr-xr-x 2 root root      4096 2012-03-06 19:35 node
[13:27] <Dioxin> isnt that the issue?
[13:27] <ikonia> Dioxin: that means only root can write to it
[13:28] <Dioxin> exactly, my VM isnt root
[13:28] <ikonia> Dioxin: hang on
[13:28] <ikonia> Dioxin: you've not copied anything to the NFS file systems yet have you ?
[13:28] <ikonia> or have you ?
[13:28] <Dioxin> no I havent
[13:29] <ikonia> Dioxin: what is your current usename on the system
[13:29] <ikonia> ok, so then your VM has nothing to do with this - don't include it in the discussion to confuse it
[13:30] <Dioxin> on the VM I'm user:ubuntu on the nfs server I'm user:dioxin
[13:30] <Dioxin> I would have thought that the nfs mount on the VM doesnt have root permissions on the folder
[13:31] <ikonia> Dioxin: no - forget the vm
[13:32] <ikonia> Dioxin: ok, so you are logged in as "dixon" so your user has no capabilities to write to a file system with permissions owned by "root:root"
[13:32] <ikonia> Dioxin: you need to fix that, so you can copy your vm file systems off your machine and copy them to your nfs server
[13:32] <ikonia> Dioxin: try not to confuse/make the issue more complex, it's a simple permissions issue at this time
[13:33] <Dioxin> but user:dioxin isnt doing the writing, I thought
[13:34] <ikonia> Dioxin: how are you trying to copy the file systems ?
[13:35] <Dioxin> in VM booted from liveCD, mounted nfs, mounted VM HDD, cp -arp * ../node/
[13:39] <lamont> tiaz: barberry/elderberry /1
[13:39] <lamont> bah
[13:39] <lamont> .1
[13:39] <Dioxin> ikonia: ok, I've fixed it, it was the folder permissions + exports rw issue + no_root_squash issue
[13:40] <Dioxin> apologies for the confusion
[13:43] <ikonia> Dioxin: no need for apologies
[13:43] <Dioxin> I should have done this on my other computer... 10/100 is slow!!!
[14:21] <Daviey> smoser: cloud-init used standlone, local-hostname: foo .. doesn't set foo in /etc/hosts.. is that known?
[14:21] <Daviey> (using nocloud)
[14:22] <smoser> you have to tell it to do that.
[14:22] <Daviey> smoser: shouldn't it do it auto-magically ?
[14:22] <smoser> its a mess.
[14:22] <smoser> if you do it automagically, someone gets pissed off.
[14:23] <smoser> if you dont
[14:23] <smoser> someone else gets pissed off
[14:23] <smoser> so it is how it is
[14:23] <smoser> if you want it, tell it to do so in config
[14:23] <Daviey> smoser: ok, is it worth updating the README? :)
[14:24] <Daviey> smoser: But i do want to mention, it's a very graceful way of starting vm's
[14:24] <Daviey> good job. :)
[14:24] <smoser> thanks.
[14:24] <smoser> you need to set
[14:24] <smoser> manage_etc_hosts:
[14:24] <smoser> to true
[14:24] <smoser> or template
[14:24] <Daviey> ah nice, i thought i'd have to echo myself
[14:25] <smoser> or 'localhost'.
[14:25] <smoser> localhost might be the best decision there.
[14:25] <smoser> rbasak, which do you use ?
[14:26]  * rbasak looks
[14:27] <rbasak> smoser, Daviey: http://paste.ubuntu.com/876070/
[14:27] <Daviey> rbasak: do you have a wrapper?
[14:28] <rbasak> Daviey: the wrapper I showed you at the oneiric release sprint you mean?
[14:28] <rbasak> Daviey: yes, I do :)
[14:28] <Daviey> rbasak: a wrapper for starting cloud-img's using cloud-init data?
[14:29] <Daviey> on localhost, with an iso attached as datasource?
[14:29] <rbasak> Daviey: no, not yet. I plan to add support for local stuff, but haven't started that yet
[14:30] <Daviey> ah, ok, cool
[14:30] <smoser> Daviey, i'll take a patch to put 'manage_etc_hosts: localhost' into that README
[14:30] <smoser> if you'd like.
[14:31] <Daviey> smoser: Do i get mentioned in the AUTHOR's file?
[14:31] <Daviey> Do i need to sign a CLA?
[14:34] <smoser> well, you'll have to ask your manager if you need to sign the CLA
[14:34] <smoser> but traditionally... in to get into the AUTHORS file, there is usually a paypal or fermented beverate exchange required.
[14:35] <rbasak> smoser: shouldn't you be checking with _your_ manager? :)
[14:35] <Daviey> smoser: sadly, i don't have a manager, only a director :(
[14:35] <smoser> hm.. i think rbasak is right, i'll ask my manager.
[14:36] <Daviey> cool
[14:36] <smoser> Daviey, do should Dave Walker need to sign the CLA to contribute a documentation fix ?  I'm not sure if his employer, Canonical, has signed the CLA or not.  Could you check with legal ?
[14:36] <smoser> s/do //
[14:40] <Daviey> smoser: wilco!
[14:52] <pabelanger> So, more upstart and openstack :)  Any thoughts about adding mysql as a 'start on' dependency for nova, glance, etc packages.  Obviously this only affect installs where mysql is on the local system.
[15:44] <pabelanger> So using start on (local-filesystems and net-device-up IFACE!=lo) and started mysql for glance's upstart help glance cleaner however I'm using an 'and'
[16:13] <pabelanger> I just added a debdiff for bug 950935 If accepted, I'd like to do the same for nova
[16:54] <reisi> how did one fix a package with broken scripts? x11-common (1:7.6+7ubuntu7.1) seems to refer to "tempnam": "tempnam: No such file or directory"
[16:59] <jorge> somebody could help me to understand how nova-* packages works in oneric 11.10 server? I can see 3 versions with apt-cache show nova-api (for example), but i seems the operate system is not upgrading the packages.
[17:08] <SpamapS> jorge: apt-acache policy nova-api will show you the priority/version/etc
[17:11] <jorge> *** 2011.3+git20111117-0ubuntu1 0      2011.3-0ubuntu6.4 0      2011.3-0ubuntu6 0
[17:11] <jorge> what is the newer one?
[17:12] <jorge> i think 2011.3-0ubuntu6.4 is the newer version for all nova packages, right?
[17:22] <jorge> SpamapS: I can see here my policy is to use nova* packages of *** 2011.3+git20111117-0ubuntu1. However,
[17:23] <jorge> SpamapS: I think 2011.3-0ubuntu6.4 version has some bug fixes and a security update.
[17:24] <jorge> SpamapS: I expected that the system automatically upgrade to this new version, but this didn't happen.
[17:25] <SpamapS> jorge: 2011.3+git is higher than 2011.3
[17:28] <SpamapS> $ dpkg --compare-versions 2011.3+git '>>' 2011.3 && echo 2011.3+git is higher
[17:28] <SpamapS> 2011.3+git is higher
[17:28] <SpamapS> jorge: that *should* have been ~git
[17:34] <nOStahl> hey can I dd a drive from ubuntu server installer?
[17:34] <zul> jorge: hmmm?
[17:37] <nOStahl> anyoe know?
[17:42] <SpamapS> nOStahl: dd might not be available. What exactly are you wanting to do?
[17:43] <nOStahl> zero out the first meg and last meg of my drive
[17:43] <nOStahl> doesn't look like dd is an option
[17:43] <SpamapS> nOStahl: just for fun? ;)
[17:43] <nOStahl> guessi 'll have to boot up a desktop live cd heh
[17:43] <SpamapS> nOStahl: you could boot a live CD and do that.
[17:43] <nOStahl> ya
[17:44] <nOStahl> I'm switching a power edge 2850 from raid to single drive mode
[17:44] <SpamapS> nOStahl: what reason do you have for 0'ing those two sections though?
[17:44] <nOStahl> had hard drive go out
[17:44] <SpamapS> nOStahl: what would the 1st MB/last MB have to do with that?
[17:45] <nOStahl> I'm told meta data is held there from the raid
[17:45] <SpamapS> nOStahl: but what would zero'ing them out do?
[17:45] <nOStahl> get rid of it
[17:46] <nOStahl> I'm just following orders :)
[17:46] <nOStahl> lol
[17:46] <SpamapS> yay, you got rid of it. Why go through trouble to do that?
[17:47] <nOStahl> just want to get a fresh install of ubuntu server running on the box (tried wouldn't boot kept trying to pick up raid config
[17:47] <nOStahl> even though its been disabled in bios with the hardware raid controller etc.
[17:49] <stgraber> jjohansen: apparmor? :)
[17:49] <jdstrand> stgraber: I am preparing the upload
[17:49] <stgraber> jdstrand: rocks!
[17:49] <jdstrand> stgraber: it should be soon
[17:50] <stgraber> cool so I can hopefully upload a new LXC with the mount rules later this afternoon/this evening then
[18:10] <arosales> lynxman: Thanks for the mcollective patch to remove ohai in bug 948437 :-)
[18:11] <lynxman> arosales: np :)
[18:15] <jorge> SpamapS: sorry man! i'm back now. i've tested the commands ... 2011.3+git is higher
[18:16] <jorge> SpamapS: the issu is that i'm having some problems with the packages... now, i know i have to wait some new packages that fixe the problems.... i'm having this problem: https://bugs.launchpad.net/nova/+bug/855660
[18:27] <stgraber> jdstrand: yeah! thanks for the upload
[18:28] <jorge> uvirtbot: ok, thanks! i think it will be available in essex... i'm using diablo packages from ubuntu oneiric repository. my packages are up to date and i'm still having problems. anyway, thanks!
[18:28] <jorge> uvirtbot: ok thanks! i think it will be available in essex... i'm using diablo packages from ubuntu oneiric repository. my packages are up to date and i'm still having problems. anyway, thanks!
[18:28] <jdstrand> stgraber: your welcome. hopefully it'll work well for you
[18:29] <jdstrand> s/your/you're/
[18:31] <SpamapS> jorge: you can force the installation of any version with  apt-get install package=xxxx .. though it will likely "upgrade" you to the other version on your next upgrade unless you use "pinning"
[18:34] <jorge> SpamapS: ok! thank you!
[18:38] <_GoRDoN__> Hi! I'm having some problems with mdadm. Yesterday I switched a failde disk in my raid5 array to a new one. Rebuild startede automatically and after it finished everything worked perfectly. However after I rebooted my machine array stopped working and by that I mean cryptsetup wouldn't accept any passpharases or keyfiles. After couple of hours of googling and panicing =) I noticed that mdadm was using /dev/sdd instead of /dev/sdd1 and if I 
[18:39] <SpamapS> _GoRDoN__: truncated at "and if I"
[18:40] <_GoRDoN__> ...and if I unplugged the new disk, remaining array worked again. So how'll I tell mdadm to use /dev/sdd1
[18:42] <SpamapS> _GoRDoN__: I'd remove /dev/sdd from the array (mdadm --remove /dev/mdX /dev/sdd)
[18:42] <SpamapS> _GoRDoN__: Then zero both superblocks with mdadm --zero-superblock /dev/sdd and mdadm --zero-superblock /dev/sdd1
[18:42] <SpamapS> _GoRDoN__: then add /dev/sdd1 (mdadm --add /dev/mdX /dev/sdd1)
[18:43] <SpamapS> _GoRDoN__: Make sure your data is backed up btw. :)
[18:53] <_GoRDoN__> SpamapS: /dev/sdd disappeared from the array stats in /proc/mdstat  when I unplugged it and after that I added sdd1 to array. Wouldn't rebuild overwrite superblocks or can I zero them after/during rebuild?
[18:54] <patdk-wk> that is just what it's actually using, realtime
[18:54] <patdk-wk> check what the config file says, in /etc/mdadm
[18:55]  * SpamapS does not ever trust /etc/mdadm
[18:55] <patdk-wk> no, but you should update it
[18:55] <patdk-wk> so it's accurate
[18:55] <SpamapS> ...
[18:55] <SpamapS> we should really just not have the stupid thing
[18:55] <SpamapS> an on disk file to define the on disk state of superblocks.. wtf?
[18:56] <_GoRDoN__> There are no devices listed in /etc/mdadm/mdadm.conf
[18:56] <patdk-wk> ok, then that isn't helping to cause the issue
[18:57] <patdk-wk> a rebuild of sdd1 won't overwrite the superblocks of sdd
[18:57] <SpamapS> _GoRDoN__: well if sdd1 has never been part of the array, then you don't need to zero superblocks
[18:57] <SpamapS> _GoRDoN__: but as patdk-wk says.. still zero'ing /dev/sdd's superblock is a good idea to keep it clean
[18:57] <_GoRDoN__> No it's brand new disk
[18:58] <_GoRDoN__> Can I do that while array is rebuilding or should i remove it from array first?
[18:58] <patdk-wk> hmm
[19:01] <SpamapS> _GoRDoN__: I think it might be safe... fun experiment if you're willing to don the bomb-squad suit and light it on fire to see if it burns or not ;)
[19:07] <_GoRDoN__> It seems that rebuild has reserved the disk entirely to itself and neither removing or zeroing wont work
[19:08] <_GoRDoN__> And fdisk says that partitiontable is invalid
[19:17] <patdk-wk> how exactly did you tell mdadm to use that disk?
[19:17] <patdk-wk> last I knew, it wouldn't automatically reuse a disk unless it was marked as a spare
[19:19] <_GoRDoN__> patdk-wk: mdadm /dev/md1 --add /dev/sdd1
[19:29] <_GoRDoN__> Okay... I'll try to remove both sdd and sdd1 from array, repair partitiontable, zero superblocks for both and add sdd1 back to array and see how it goes
[19:58] <stgraber> jjohansen: apparmor tech support? :)
[19:58] <stgraber> jjohansen: so I can now use mount without it getting into an infinite loop, though it doesn't seem to actually work...
[19:59] <stgraber> jjohansen: I started with http://paste.ubuntu.com/876509/ to just allow everything but I'm still get the mount DENIED
[19:59] <stgraber> jjohansen: my first try was with "mount /** -> /usr/lib/lxc/root" which is actually what we want to allow but that didn't work either
[20:00] <stgraber> jdstrand: ^
[20:00] <jjohansen> stgraber: hrmm, it should work, I did test.
[20:00] <stgraber> [36108.427446] type=1400 audit(1331323087.420:170): apparmor="DENIED" operation="mount" parent=32353 profile="/usr/bin/lxc-start" name="/usr/lib/lxc/root/" pid=32375 comm="lxc-start" src_name="/home/stgraber/data/vm/lxc/containers/precise/rootfs/" flags="rw, rbind"
[20:01] <jjohansen> stgraber: alright, I will take another look, I may need to give you a test kernel
[20:01] <stgraber> jjohansen: ok :)
[20:02] <jjohansen> stgraber: is there anything specific you where doing?  You where just using lxc-start and the lxc profile right
[20:03] <stgraber> jjohansen: yep, just lxc-start with the profile I linked above which is the standard lxc profile with mount added
[20:03] <stgraber> I have apparmor 2.7.100-0ubuntu1 and 3.2.0-18-generic (64bit)
[20:04] <jjohansen> stgraber: okay, thanks. I'll start from a clean install, maybe I had something present messing with things
[20:09] <stgraber> jjohansen: I'm testing on my laptop so I'm not excluding I have something causing problems but I'm not running any non-official packages or kernel
[20:11] <jjohansen> stgraber: uh my bet is its the kernel, on the bind path, I am building a new kernel to test
[20:39] <axisys> how do I install ubuntu on a remote system.. i only have ilom access through the SP.. it is physically connected to the switch at the site..
[20:39] <axisys> it is sun fire x4270
[20:40] <axisys> java wbstart to the remote console shows all the devices grayed out.. so cannot map a iso to the cdrom device
[20:40] <guntbert> axisys: what is "ilom access through the SP" ?
[20:41] <axisys> guntbert: network management port
[20:42] <axisys> i can bring the system up through that port.. i can use java ws to remote console into it and could map a iso to the cdrom device if not grayed out
[20:42] <axisys> so other option would be install over the network
[20:42] <axisys> like net install
[20:42] <axisys> planning to install latest LTS
[20:42] <axisys> 64bit
[20:43] <axisys> net install is not so easy compare to install from a virtual cdrom..
[20:44] <guntbert> axisys: why can't you  map the iso? On my vps that is managed via the web GUI
[20:45] <axisys> guntbert: the web gui has an option for that.. but it is grayed out like this
[20:45] <axisys> http://picpaste.com/pics/ilom-t8IvCeF9.1331325919.png
[20:46] <axisys> i tried few versions of java webstart .. same issue
[20:48] <guntbert> axisys: ah, now I know what ilom means :-)  - one idea: try to power on the machine, maybe the redirections are only available when powered on
[20:52] <axisys> guntbert: it is on
[20:53] <axisys> Host Power: On
[20:53] <guntbert> axisys: and when it is off?
[20:53] <axisys> guntbert: have not tried with off
[20:54] <guntbert> axisys: I'm just fishing around :-)    maybe try to ask support
[20:54] <axisys> guntbert: means ubuntu server mailing list?
[20:55] <guntbert> axisys: no, Oracle
[20:55] <axisys> guntbert: oh ok
[20:56] <guntbert> axisys: I assume they will know under which circumstances you can attach a virtual device
[20:57] <axisys> guntbert: i will go ahead and create a ticket.. thanks
[20:58] <guntbert> axisys: don't they have a forum too?
[20:58] <axisys> guntbert: they should.. i thought google would find something in there..
[21:00] <guntbert> axisys: they are talking about some sort of dongle needed
[21:01] <axisys> guntbert: hmm.. that's first time.. i usually make virtual device fine.. this is the first server
[21:01] <guntbert> axisys: I searched for      ilom "attach devices"
[21:02] <axisys> Remote console does not need it
[21:02] <axisys> you were looking at local console
[21:14] <guntbert> axisys: you know definitely more about it than I do :)
[21:14] <axisys> I am just going to do a net install.. it might take a while for oracle to get back
[21:14] <axisys> so any good page on net install that I should follow
[21:15] <axisys> i have a server on same network that is running ubuntu lts
[21:15] <axisys> lucid 64bit
[21:17] <guntbert> axisys: not from me, sorry -- with ubuntu netinstall means obviously "boot from the minimal CD"...
[21:17] <axisys> guntbert: i meants PXE
[21:17] <axisys> meant*
[21:19] <guntbert> axisys: I assumed as much -- did you see igor.chudov.com/projects/PXE-Netinstall-Of-Ubuntu/  or    my.opera.com/ilogico/blog/how-to-netboot-and-netinstall-ubuntu  ?
[21:19] <axisys> so far I found this https://help.ubuntu.com/community/PXEInstallServer
[21:20] <axisys> let me look through all
[21:41] <smoser> roaksoax, i just uploaded cloud-init with maas support.
[21:41] <smoser> so right now, cobbler is the last piece of the maas -> cloud-init puzzle.
[21:41] <smoser> i can look at that on monday.
[21:43] <smoser> it looks like we just need a hole in the preseed files for kickstart variable MAAS_PRESEED
[21:57] <koolhead17> zul, :)
[21:59] <roaksoax> smoser: cool stuff
[22:00] <roaksoax> smoser: monday will distributed stuff from orchestra to cobbler and so on
[22:24] <axisys> i have 8 300GB disks and 2 148GB disks.. should I do a raid10 of 8 big and raid1 of 2 small disks and then apply lvm on top.. or do I raid1 of 2 small and the 8 disks all LVM w/o md ?
[22:24] <axisys> cost is not an issue.. space is not an issue..
[22:32] <SpamapS> axisys: lvm of 10 disks would be almost as bad as RAID 0
[22:33] <SpamapS> axisys: a RAID 10 for the 8 disks will give you excellent performance for almost any workload, assuming you have enough bus to handle 10 disks reading/writing in parallel :)
[22:33]  * SpamapS goes to eat lunch
[23:11] <jjohansen> stgraber: do you have a specific config file you are using or are you just using the default
[23:22] <ubu-stu> Q How do you create a samba share on a PDC so that only one person can update the web site vis their windows login.?
[23:45] <SpamapS> ubu-stu: its highly dependent on how you have your users setup
[23:46] <SpamapS> ubu-stu: if the PDC's users are the "real" users, then just create a share with 'write list = thatuser'