[00:31] <ScottK> nigelb: With lots of pain.
[00:32] <nigelb> ScottK: that seems to be the general opinion :(
[00:33] <nigelb> I tried letting the debian/rules run, but that didn't really untar it.
[00:38] <LowValueTarget> my apt is screwed
[00:38] <DSpair> Oh, BTW...
[00:38] <DSpair> Got my server fixed.
[00:39] <DSpair> The Ubuntu 10.04.1 "rescue broken system" options let me do everything I needed. Thanks Ubuntu!!!
[00:39] <DSpair> Steps I had to take::
[00:39] <DSpair> 1. Had to resize my LVM to make room for a raw partition to hold the /boot volume.
[00:40] <DSpair> 2. Had to format that partition and restore the contents of /boot
[00:40] <DSpair> 3. Booted from the rescue option on the server install disc.
[00:40] <DSpair> 4. Opened a shell on the root FS of the machine.
[00:40] <DSpair> 5. Mounted all filesystems
[00:40] <DSpair> 6. Performed an aptitude dist-upgrade (which failed because of the chroot environment).
[00:41] <DSpair> 7. cd'd to /var/cache/apt/archives and performed a 'dpkg -i --force-all *.deb'
[00:41] <DSpair> 8. Rebooted.
[00:41] <DSpair> 9. Logged in and obtained root
[00:41] <DSpair> 10. Performed another dist-upgrade.
[00:41] <DSpair> 11. Danced a little jig!
[02:05] <databits> I'm having some issues, with some virtual server's with apache.  Everything was working perfectly, before I left town.  Now when I bring up the url in a browser, it is listening the root web directory for the server.
[02:05] <databits> Does anyone know what might be causing this issue ?
[02:17] <databits> figured it out.
[02:30] <WALoeIII> how/why does a package have a name like 2:
[02:30] <WALoeIII> like redis-server
[02:30] <WALoeIII> 2:1.2.0
[02:35] <twb> WALoeIII: that is the "epoch number".
[02:35] <twb> WALoeIII: it is used when upstream's version number decreases, because Debian versions MUST increase monotonically.
[02:35] <twb> WALoeIII: for most purposes you can ignore the epoch number.
[02:36] <WALoeIII> ugh
[02:37] <fuho> join
[02:37] <fuho> Hi
[02:38] <fuho> Is this the right place to have a question about how to install VNC server on my VPS running Ubuntu Server 10.04?
[02:38] <ScottK> nigelb: What's the patch system for the package?  If you've got a patch, you can just add it to the patch system without untarring the interior tarball.  Just check other patches to make sure you've got the right path in the patch..
[02:39] <twb> fuho: yes; install vncserver4 or so
[02:39] <fuho> twb: dO i have to install Gnome or something like that?
[02:39] <twb> fuho: only if you want to run gnome inside your VNC session
[02:40] <fuho> twb: I thought VNC is always graphical, and for terminal we have Putty.
[02:40] <twb> VNC *is* graphical, but you could run, say, xlogo or xterm in it instead of GNOME.
[02:41] <fuho> twb: Tried installing the whole ubuntu-desktop metapackage, took over 4 hours and then when installing VNC it didnt created ".vnc" folder, so i reformatted the whole box, so now i have clean installation again, dont want to do the same mistake again :(
[02:41] <twb> The linuxvnc package, for example, appears to export the text console over VNC.
[02:41] <twb> fuho: I don't know what a ".vnc folder" is.
[02:42] <twb> fuho: why can't you just use the command line over SSH (putty) ?
[02:43] <fuho> twb: It was part of the server, i think it didnt create the directory because i didnt run the vnc....i will show you the link, just a moment. But anyways I would like to avoid installing the whiole ubuntu-desktop again, is there a way to get graphical UI with less resources?
[02:43] <twb> fuho: why do you want a graphical UI?
[02:44] <fuho> twb: I could, its just that i kinda want to have the option too, I never had VPS before and since i think its possible I would like to figure it out
[02:44] <fuho> twb: And to be honest i think it could make things a little bit faster sometimes
[02:44] <twb> We will have to disagree on that point.
[02:45] <fuho> twb: I for exmple dont know how to use vim at all.
[02:45] <fuho> twb: faster for me as a noob, with very basic if any knowledge of terminal
[02:45] <twb> It is much faster in the long term to invest a little time know learning the CLI
[02:46] <twb> If you insist, I can help you install a GUI, but we discourage it.
[02:48] <fuho> twb: Thanks. I should probably know why  do you discourage it. isn't it just another option? And in my eyes options are always better than no options.
[02:49] <Datz> fuho: there are other options to vim
[02:49] <twb> We discourage it because, like I said, learning the CLI will make you more efficient and powerful in the long run.
[02:49] <fuho> twb: Also, I plan to learn CLI (if that means using terminal), really, I just think it would be great t have the UI too
[02:52] <fuho> datz: I believe there are, I think one is called VI,I just found it difficuilt to use. I am sure in the long run it is better solution, but as a long time Windows user it is very diffiuilt to give away everything I am use to.
[02:52] <shauno> gui can be useful, just not so much on a vps, where you tend to be paying for a fairly small slice of ram
[02:53] <twb> fuho: vim is an implementation of vi.
[02:53] <Datz> fuho: I personally use emacs, but there are perhaps better options for you
[02:53] <twb> fuho: you might try "nano", which is more like NOTEPAD.EXE
[02:54] <twb> fuho: it's very limited, but easier for beginners
[02:54] <fuho> twb: I found the tutorial i was following before: http://nuclear-imaging.info/site_content/2010/04/19/vnc-server-setup-for-ubuntu-10-04-lucid-lynx/
[02:54] <twb> Like Datz, I use Emacs.  Not just for editing files, but also for email and IRC and web browsing and a bunch of other things :-P
[02:54] <Datz> fuho: shauno has a good point.
[02:55]  * Datz remembers now that there is an emacs irc client :po
[02:55] <fuho> datz: I trust you guys I just want to try it
[02:55] <twb> Datz: emacs *ships* with two IRC clients, but there are at least three more you can install separately :-)
[02:56] <Datz> humm
[02:56] <fuho> I have 2GB RAM on the server, hope that will be enough
[02:56] <Datz> fuho: that is plenty for a GUI
[02:56] <shauno> that's not half bad for a vps.  I get 360Mb on mine :)
[02:56] <Datz> twb: I installed emacs-nox11 does that ship with IRC too?
[02:57] <twb> Datz: yes, M-x irc (rcirc) and M-x erc (ERC).
[02:57] <shauno> emacs ships with everything but a kernel.  I believe they're working on that tho
[02:57] <fuho> twb:So first I did this : "apt-get install x-window-system-core xserver-xorg gnome-desktop-environment" to install the ubuntu-desktop, i dont think i need the whole DESKTOP package though
[02:57] <twb> fuho: how much *do* you need?
[02:57] <Datz> twb: humm, have to try it sometime. right now I use irssi
[02:58] <twb> Datz: stop by #emacs sometime if you want to talk about it
[02:58] <Datz> twb: sure, thanks
[02:58] <fuho> oh by the way this is not ad, can i recommend VPS here? its really cheap
[02:59] <twb> fuho: I don't care either way; I suppose if you get too annoying an op will yell at you.
[02:59] <fuho> twb: I would like to be able to use synaptics and edit files in some normal text editor with ui
[02:59] <twb> fuho: your local machine is Windows?
[03:00] <fuho> just so you know,cause i think i got a pretty good deal on this its 160GB HDD, 1GB RAM (2GB peak) 600GB bandwith for $13.5/ month
[03:00] <twb> If your local machine was unix, you could "ssh -X example.net synaptics" to just run the synaptics GUI, without needing VNC or gnome.
[03:00] <twb> You can do it on Windows, too, if you can work out how to run an X server on the Windows machine.
[03:01] <fuho> twb: Yes, right now it is, my T23 died, I used to have ubuntu on it, but now I only have W7 desktop box.
[03:02] <fuho> twb: I was really hoping to be able toaccess it from anywhere, therefore some sort of java environmebt
[03:02] <twb> I dunno about that, sorry.
[03:03] <fuho> twb: I dont need that, I just thought it would be cool if possible. So lets just stick with windows and RealVNC client.
[03:03] <twb> FWIW I would recommend you learn aptitude's TUI instead of synaptic
[03:05] <Datz> yea, vnc is slow and consumes too much bandwidth
[03:05] <fuho> twb: What is the advantage of Aptitude to synaptic?
[03:05] <twb> fuho: it runs in putty
[03:06] <fuho> twb: Synaptic doesnt?
[03:06] <twb> An ssh CLI connection uses much less bandwidth than a GUI, regardless of whether it's VNC, RDP or X11.
[03:06] <fuho> twb: I thought that when i do apt-get install something                it is synaptic
[03:06] <twb> fuho: no, that's apt-get.
[03:07] <twb> fuho: synaptic is a GTK2 GUI wrapper around apt.
[03:07] <twb> Aptitude is a different wrapper, and it has a GUI, a TUI and a CLI
[03:07]  * Datz did not know about aptitudes GUI
[03:08] <twb> Datz: it's relatively new; it wasn't around in 8.04
[03:08] <twb> Datz: personally, I've never run it :-)
[03:08] <Datz> ah
[03:08] <Datz> haha, I see
[03:08] <Eric_> ubuntu server is bad!
[03:08] <twb> aptitude-gtk - terminal-based package manager (GUI and terminal interfaces)
[03:08] <fuho> twb: CLI connection is just pure text putty?
[03:08] <twb> fuho: yeah
[03:08] <Datz> talntid: yea, bad to the bone
[03:08] <twb> CLI = command-line interface
[03:08] <talntid> =D
[03:09] <Datz> ;)
[03:09] <twb> talntid: did you want help with something, or are you just venting?
[03:09] <fuho> twb: Oh :) thanks
[03:10] <talntid> i'd love help, but it's above the level of expertise of most of the people in here... so I don't ask.. I'm not here to vent.. just watching :)
[03:10] <fuho> twb: So which one would you suggest X11 (which i think is the one that can do Java magic) or VNC or RDP (i didnt even know you can d that on linux, i thought thats purely windows)
[03:10] <twb> There isn't an RDP *server* for Linux.
[03:11] <fuho> twb: so X1 or VNc? Advantages?
[03:11] <twb> Well, technically there's an RFB (VNC) to RDP bridge, but using it defeats the purpose of using RDP -- namely, that it operates on a higher level than raster damage rectangles.
[03:12] <twb> fuho: X11 (or NX, I suppose) is what you should use for Linux <--> Linux.  VNC is the "lowest common denominator", it works everywhere but it's crap.
[03:12] <twb> NX is a recent X11 protocol compressor, I haven't used it much myself, but it seems to make X11 much more bandwidth-efficient.
[03:13] <fuho> twb: Oh, well unfortunately most of the owlr has Windows so I will go with VNC for now, then when i get a new laptop with linux i will install X11 too
[03:13] <fuho> twb: BNut can you help me install it ?
[03:14] <twb> Sure.
[03:14] <twb> You will need to install vnc4server plus whatever app/desktop you want to run.
[03:14] <fuho> twb: Only if there is no way to install X11 client on windows machine
[03:15] <twb> fuho: there are X servers for Windows, but I don't know if any are any good.  xming is a recent port of Xorg to w32.
[03:15] <twb> (X calls the client side the "server" and the GUI apps the "clients".  Don't worry about why.)
[03:16] <fuho> twb: Just found Xming, reding wiki, not sure if this is the wy to go yet.
[03:18] <fuho> twb: It definitely looks usable and they are still working on it, last release is fron november, screnshots look fine and if you say its less resource hungry i would go that way
[03:19] <fuho> twb: So X11 it is, now how do we start?
[03:19] <twb> fuho: I think you will get VNC up easier than xming
[03:19] <fuho> twb: Do you think you can help me with X11?
[03:19] <twb> fuho: for X11, on the Ubuntu side you just need to ensure that xauth and the app (e.g. synaptic) are installed.
[03:25] <fuho> twb: OMG you were right, it looks like x11 is for connecting to applications not whole desktops,
[03:26] <twb> fuho: X11 is for everything.
[03:26] <fuho> twb: i just jum from one to the other :(
[03:26] <twb> fuho: when you run an app on your unix desktop, it's talking X11 on the local machine.
[03:27] <shauno> vnc's probably a much better idea if you're doing this over the internet.  the only thing X has ever done well, is survive where it shouldn't have
[03:28] <fuho> twb,shauno: Okay, I played with xming a bit, id much rather use VNC now, I am ueśed to it.
[03:28] <fuho> So now I have to pick what UI I want? Gnome or the the other one?
[03:29] <Slugs_> can you start a virtualized hardy session w/o a gui, only conosle?
[03:29] <Slugs_> using kvm
[03:30] <fuho> twb: Also is it possible to not have the Gnome running all the time, to only swith iton when i need it (when I am connecting)?
[03:30] <Slugs_> skitter@vagabond:/ubuntu-kvm/ubuntu-kvm$ virsh start ubuntu error: Failed to start domain ubuntu error: monitor socket did not show up.: Connection refused
[03:31] <fuho> shauno: Thanks, I barely even heard aof Xming today, and now when i installed it I dontthink i want to know more atleast for couple weeks :)
[03:32] <shauno> they're nothing scary .. remote X is how it was designed from day one.  It's just not particularly efficient.
[03:33] <Datz> fuho: you sould be able to do something like "sudo service gdm start/stop"
[03:34] <fuho> datz: letme first see what gdm does
[03:34] <twb> fuho: if you have ssh as well, sure.  Just manually start VNC by sshing in and running "xinit /usr/bin/gnome-session -- /usr/bin/Xvncserver" or something
[03:35] <twb> Datz: that assumes the gdm upstart job is configured to use Xvncserver is Xorg isn't installed -- I'm not sure if that's the case.
[03:35] <Datz> twb: ah, ok.
[03:35] <twb> shauno: X is actually quite efficient if you're using contemporary toolkits (i.e. xlib or xaw).
[03:36] <twb> shauno: it's just REALLY slow if you're using bloaty toolkits/apps that were only designed and tested on local connections, especially gtk/qt and firefox/oo.org
[03:36] <Slugs_> anybody have exp with virsh?
[03:36] <fuho> twb: What should i install first?
[03:37] <twb> For example, for me emacs21 worked quite happily over 256kbps lines, but oo.org 2.0 crawls on a 100mbps lan
[03:38] <fuho> BTW what does logmein use, that seem very fast even on pretty slow connections?
[03:38] <twb> Slugs_: there's a dedicated channel for libvirt, but I can't remember its name.  #virt or #libvirt?
[03:38] <Slugs_> oh ty
[03:38] <twb> Slugs_: you can wait here, too.  Some people know about libvirt here
[03:38] <StrangeCharm> i just set up a new ubuntu install, but when i try to boot into the new system, my machine halts after 'verifying dmi pool data' what's going on? [i'm certain that i'm booting from the right disk]
[03:39] <Slugs_> yeah #ubuntu-virt redirected me here
[03:39] <twb> shauno: of course, NX makes it a lot faster, as should xcb :-)
[03:40] <twb> StrangeCharm: when booting, hold shift or alt to get into grub.  Edit the boot options and change "quiet splash" to "single".  That will boot with more information, so you can see what's happening
[03:40] <fuho> twb, shauno: So is anuone wilingto point me in the right direction, or even push me to the finish?
[03:40] <twb> fuho: I'm better at specific questions
[03:40] <fuho> vanilla ubuntu serve with ssh I am connected to right now
[03:41] <StrangeCharm> twb, i don't think that my bios is actually loading grub. at no point do i get the 'grub loading' line, just halt on/after verifying the dmi pool
[03:41] <fuho> twb: Do I need to install vnc server first?
[03:41] <twb> fuho: you will need to install a VNC server (e.g. vnc4server) before you can do VNC
[03:41] <twb> StrangeCharm: then it's a hardware problem and you need to isolate the faulty component or take it back to your vendor and have them do so
[03:42] <fuho> twb: So i dont need the UI manager (its probably not called UI manager, but I mean Gnome probably)
[03:42] <twb> fuho: you will need to install whatever GUI app or environment you want to run inside VNC
[03:43] <twb> fuho: to begin with, you could just use xterm or xlogo to test that the VNC part is working, and *then* install whatever app/DE you really want
[03:43] <StrangeCharm> twb, i don't think it's a hardware problem. it booted just fine with the pervious os, and the only hardware change was connecting the usb drive with the install media
[03:44] <twb> StrangeCharm: if it isn't getting as far as grub, then surely the bios or hardware is at fault
[03:44] <twb> StrangeCharm: that or you haven't installed grub correctly, I guess.
[03:45] <StrangeCharm> twb, i wasn't exactly involved when installing grub - the install disk did that all on its own. i'm concerned that the install disk did something odd, because i can boot from removable media, and could boot under the previous os
[03:46] <twb> StrangeCharm: I've had lots of problems with d-i installing grub onto the wrong disk (e.g. onto the USB key)
[03:47] <StrangeCharm> twb, d-i? i used the partition editor on the install disk. surely it put grub at the start of the disk with /boot on it?
[03:47] <twb> StrangeCharm: d-i is the installer for Ubuntu Server
[03:47] <StrangeCharm> twb, why is it called d-i?
[03:48] <twb> Because it's the debian-installer program that Ubuntu, uh, borrowed from Debian
[03:48] <twb> cf. ubiquity
[03:48] <fuho> twb: Just want to ask you before I press eneter "sudo apt-get install ubuntu-desktop" I am going to isntall this and then i will install the vnc server, is that correct?
[03:49] <twb> fuho: installing ubuntu-desktop is probably overkill
[03:50] <fuho> twb: https://help.ubuntu.com/community/ServerGUI Looks like I only have two options, X11 or  this.
[03:51] <twb> fuho: what is "this"?
[03:51] <fuho> gwb: ubuntu-desktop
[03:52] <twb> fuho: ubuntu-desktop is a metapackage that pulls in EVERYTHING that a normal Ubuntu desktop has
[03:52] <twb> It'll work, but do you really want to install e.g. rhythmbox?
[03:52] <fuho> twb: nope i just want gnome, so i acan conenct to it using vnc
[03:54] <twb> fuho: try "gnome-core" or "gnome-desktop-environment"
[03:55] <fuho> twb: gnome-desktop-environment
[03:55] <twb> gnome-core is the gnome desktop, but none of the apps.  If I were you, I'd install that and specific apps
[03:55] <fuho> twb: ha, I jsut found it too!
[03:55] <fuho> twb: ok I will do taht
[03:57] <StrangeCharm> twb, so, the fix here is probably to reinstall grub on the right disk?
[03:57] <fuho> twb: Ok, that is 650MB to install. After this finishes do I ahve to REstart the server? This is kinda generic question, do I have to restart the server every time i install something?
[04:01] <twb> fuho: no restart necessary
[04:01] <twb> StrangeCharm: *if* that's the problem
[04:01] <twb> StrangeCharm: your first step should be to diagnose the issue
[04:02] <StrangeCharm> twb, what could i do to rule out issues?
[04:02] <twb> StrangeCharm: boot something else and inspect the MBR and the partitions
[04:02] <twb> StrangeCharm: pull out the USB key and see if that helps
[04:02] <twb> StrangeCharm: check the BIOS boot order
[04:03] <StrangeCharm> twb, options b & c already investigated. i have no idea what i'd be looking for on the mbr & boot partitions.
[04:04] <twb> StrangeCharm: file -s /dev/sda would be a start -- check that they're bootable MBRs
[04:05] <StrangeCharm> twb, bootable mbrs? you mean uncorrupt?
[04:05] <twb> It should say /dev/sda: x86 boot sector; ...
[04:05] <twb> If it doesn't say "boot sector", it's not bootable
[04:06] <StrangeCharm> twb, so, file -s will read the mbr of a disk, and tell me something meaningful about it?
[04:07] <twb> Basically, yes
[04:09] <fuho> twb: Do you have any experience with x11vnc? http://www.karlrunge.com/x11vnc/
[04:10] <twb> fuho: x11vnc exports an existing X session.  It's not what you want.
[04:11] <fuho> twb: oh ok, i will keep looking. people seem to run into lot of problems installing vn4server, so I ama a bit scared.
[04:17] <WALoeIII> can anyone proficient with debhelper/debuild help me diagnose an issue with dh_auto_install
[04:19] <twb> WALoeIII: I can, but -devel might be a better channel
[04:19] <twb> WALoeIII: or #debian-mentors on OFTC
[04:20] <WALoeIII> twb: ok, https://gist.github.com/711240 - it seems when it does the make install it has a permissions issue?
[04:21] <WALoeIII> seems totally weird to me, the 2.0.0rc2 stuff I'm basing this off of wokrs
[04:21] <WALoeIII> works*
[04:21] <fuho> twb: Okay, so gnome-core is installed, now I plan to do this "sudo apt-get install vnc4server xinetd"
[04:21] <twb> I don't know why you'd want xinetd
[04:22] <fuho> twb: Apparently its more secure and can keep better logs.
[04:22] <twb> fuho: keep better logs *than what*?
[04:22] <twb> fuho: why do you think you need *any* inetd implementation?
[04:23] <twb> WALoeIII: cp: cannot create regular file `/usr/bin/redis-server': Permission denied
[04:23] <twb> WALoeIII: upstream is ignoring $DESTDIR
[04:23] <WALoeIII> twb: correct
[04:23] <WALoeIII> uhrrm is that set by dh_something?
[04:23] <twb> WALoeIII: patch upstream's makefile, or override dh_auto_install with a custom invocation.
[04:23] <fuho> twb: io have no clue, i dont even know what it is. Do you think vnc4server is enough?
[04:23] <twb> WALoeIII: upstream is in the wrong to ignore DESTDIR, but it might not be feasible to fix it
[04:24] <WALoeIII> twb: I know where it is now
[04:24] <WALoeIII> makefile
[04:24] <twb> WALoeIII: yep
[04:24] <twb> fuho: yes
[04:24] <WALoeIII> INSTALL_TOP= /usr
[04:24] <icek> Anyone know any cvs server front ends?
[04:24] <WALoeIII> ok thanks, I can do that
[04:24] <icek> I want to be able to minitor like checkouts and committs and such
[04:25] <twb> WALoeIII: try override_dh_auto_install: \n\t dh_auto_install -- INSTALL_TOP=$(CURDIR)/debian/tmp/usr
[04:25] <RoAkSoAx> kirkland: ping?
[04:26] <icek> ??
[04:33] <fuho> twb: vnc4server is installed, from what i read online people tend to edit .vnc/xstarup file. I tried to connecto to the server using VNC Viewer with no luck, do you know whati s that file for?
[04:34] <twb> fuho: no.  I would try "xinit /usr/bin/xterm -- /usr/bin/Xvncserver"
[04:34] <twb> Sorry, make that "xinit /usr/bin/xterm -- /usr/bin/Xvnc4"
[04:36] <fuho> error opening security policy file /etc/X11/xserver/SecurityPolicy Could not init font path element /usr/share/fonts/X11/Speedo/, removing from list! Could not init font path element /usr/share/fonts/X11/Type1/, removing from list! Could not init font path element /usr/share/fonts/X11/CID/, removing from list! Could not init font path element /usr/share/fonts/X11/75dpi/, removing from list! Could not init font path element /usr/sh
[04:38] <twb> Those aren't errors
[04:38] <twb> X is looking for fonts in obsolete places, it is OK that it can't find them
[04:38] <fuho> twb: It still didn't connect
[04:39] <twb> pastebin the full transcript
[04:40] <fuho> twb: http://pastebin.com/tvmQpyEJ
[04:41] <twb> fuho: the error is: xinit:  No such file or directory (errno 2):  no program named "/usr/bin/xterm" in PATH
[04:41] <twb> fuho: you don't have xterm installed
[04:42] <twb> Try "xinit /usr/bin/gnome-session -- /usr/bin/Xvnc4"
[04:43] <buntu> anybody ever installed 10.04 amd64 on ibm blade HS22 ?
[04:43] <twb> !anybody
[04:44] <nigelb> ScottK: That's a relief.  Its using quilt.  I'll need to unpatch to just figure out the directory system.
[04:45] <fuho> twb: now it accepts connection, client asks fro password (not username) i type it says in client"No password configured for VNC Auth"
[04:45] <buntu> my problem on IBM HS22 is fail to boot after finished installation on /dev/sde
[04:46] <fuho> twb: And in terminal : "Tue Nov 23 04:44:09 2010  Connections: accepted: 0.0.0.0::59016  SConnection: Client needs protocol version 3.8  SConnection: Client requests security type VncAuth(2)  Tue Nov 23 04:44:10 2010  SSecurityFactoryStandard: neither Password nor PasswordFile params set  SConnection: AuthFailureException: No password configured for VNC Auth  Connections: closed: 0.0.0.0::59016 (No password configured for VNC Auth
[04:47] <twb> fuho: I don't remember how to do the password bit.  I think you're supposed to use vnc4passwd -- try reading its manpage
[04:47] <fuho> twb: http://pastebin.com/JK4kDtHr
[04:48] <billybigrigger> i keep getting an emails from root@xxxxx (xxxxx being my server) with the Subject: [ -x /usr/lib/php5/maxlifetime ] && [ -d /var/lib/php5 ] && find /var/lib/php5/ -type f -cmin +$(/usr/lib/php5/maxlifetime) -print0 | xargs -n 200 -r -0 rm and the message only contains...Warning: Directive 'register_long_arrays' is deprecated in PHP 5.3 and greater in Unknown on line 0
[04:49] <billybigrigger> i've had this message spamming my inbox for months now hoping that there would be an update that would resolve it...but i'm really sick of seeing it spam my inbox, so how do i get rid of it? as the message isn't really clear as to what file is the culprit
[04:51] <UndiFineD> grep -r -i register_long_arrays /var/www
[04:52] <UndiFineD> or it might have been set in your php.ini ?
[04:53] <fuho> twb: It was "vnc4passwd" i set up password, then ran "xinit /usr/bin/gnome-session -- /usr/bin/Xvnc4" gain and it still wont connect, with the same error: http://pastebin.com/kjzQRFdv
[04:53] <buntu> anybody experience grub problem upon installed on /dev/sde ?
[04:53] <twb> fuho: NFI, sorry
[04:59] <fuho> twb: not even why it shows 0.0.0.0 instead of my ip?
[05:00] <twb> fuho: 0.0.0.0 means "all interfaces"
[05:00] <nigelb> ScottK: err, s/unpatch/untar
[05:38] <zwang> hello
[05:39] <Datz> greetings
[05:44] <twb> LXC question
[05:44] <twb> Suppose my lxc rootfs is on LVM
[05:45] <twb> Can I use lxc.mount.entry /dev/mapper/lxc-barserv /srv/lxc/fooserv ext4 default ?
[05:45] <twb> Can I use "lxc.mount.entry = /dev/mapper/lxc-barserv /srv/lxc/fooserv ext4 defaults", rather -- in /etc/lxc/fooserv.conf, that is.
[05:50]  * twb tries
[05:53] <SpamapS> a blockdevice is a blockdevice
[05:53] <twb> SpamapS: I thought maybe the root filesystem was special as far as lxc.mount.entry was concerned
[05:54] <twb> i.e. I couldn't mount it on-demand during lxc-start
[05:57] <SpamapS> twb: I've only booted two lxc containers ever, but the root fs seems to be the easy part. ;)
[06:33] <twb> SpamapS: doesn't work
[06:33] <twb> http://paste.debian.net/100515/
[06:53] <databits> What is a good SMTP server to run to use from accepting php mail form's ?
[06:57] <twb> postfix
[07:03] <databits> that is what I'm looking into right now ... thanks
[08:27] <latenite> Hi folks, I want to manage my contacts and adresses serversided. So any client on my LAN can access these contacts. Is there a "contactserver-tool" in any form? I am looking for tools. thanks :)
[08:27] <twb> latenite: address book ("yellow pages") data is typically stored in LDAP nowadays.
[08:28] <twb> You could set that up, but I don't think there's a "turn key" solution for it in Ubuntu
[08:39] <latenite> twb, would you know a nice tuorial on ldap and adressbooks? I never got my hands on ldap. I feel like its a big monster...*scared*
[08:39] <twb> latenite: nope
[08:40] <latenite> twb, is ldap hard t learn?
[08:40] <twb> LDAP itself isn't complicated, but it's messy and some of the edge stuff like auth can be confusing
[08:41] <latenite> twb, what do you meen by "edge stuff"? *brand new*?
[09:21] <MTecknology> What would you guys suggest using if I wanted to pump out a quick and easy 'planet'
[09:21] <MTecknology> I know there's PlanetPlanet - but that doesn't seem like an 'easy' solution
[09:22] <MTecknology> I'm looking for something that's easy for users to add onto as well
[09:25] <Callum__> okay, so I am shutting down my web server remotely and then using my remote file server to kick it back up via my custom ipmitool frontend for the first time...
[09:25] <Callum__> let's see if it works
[09:31] <twb> Callum__: do you have a backup plan? :-)
[09:33] <Callum__> twb: well, it worked fine, but the backup plan was just to work with ipmitool manually if the thing didn't work, the program isn't finished yet =P
[09:34] <Callum__> anyway
[09:34] <Callum__> SELinux is giving me troubles with remote CUPS printing (over the Internet)
[09:34] <twb> Callum__: I just ssh example.net lpr <foo.pdf :P
[09:35] <Callum__> actually, its probably been causing me problems with even local printing via PostScript as well because I haven't been able to print all the time locally
[09:35] <Callum__> kernel spams something like this: [  248.167806] type=1503 audit(1290504722.212:60):  operation="open" pid=897 parent=1 profile="/usr/sbin/cupsd" requested_mask="r::" denied_mask="r::" fsuid=0 ouid=0 name="/var/run/samba/gencache.tdb" , when I try to print
[09:37] <Callum__> if SELinux is even running on this thing
[09:39] <Callum__> none of the SELinux utils are installed, so this may be AppArmor being an asshole
[09:39] <Callum__> twb: any ideas?
[09:39] <twb> dunno
[09:39] <twb> That says cupsd can't read /var/run/samba/gencache.tdb
[09:40] <twb> And yes, that's an apparmor message
[09:40] <twb> On ubuntu apprmor is on by default, selinux is not
[09:40] <Callum__> ugh, it looks just like SELinux apparently
[09:40] <Callum__> so do I have to change a rule or something?
[09:40] <twb> Hm, actually, maybe it's a generic LSM message, so it could be either
[09:40] <Callum__> I have no experience with AppArmor OR SELinux
[09:40] <twb> Callum__: only if you want it to work :P
[09:41] <twb> Callum__: look in /etc/apparmor for cupsd stuff
[09:41] <Callum__> /etc/apparmor.d/usr.sbin.cupsd, /etc/apparmor.d/cache/usr.sbin.cupsd and/etc/apparmor.d/abstractions/cups-client...
[09:42] <twb> Callum__: also, talk to #ubuntu-hardened about it
[09:42] <Callum__> what for?
[09:42] <twb> Because they're the ones that know about apparmor
[09:42] <MTecknology> ok.. so I'm going to try to use PlanetPlanet - Any chance I could get some help figuring out how to set it up?
[09:42] <Callum__> ah
[09:44] <Callum__> well, there are no permissions defined for /var/run/samba/gencache.tdb in /etc/apparmor.d/usr.sbin.cupsd
[09:45] <twb> IIUC apparmor is "default deny" for specific apps
[09:46] <Callum__> to be honest I never understood why we needed SELinux or twb
[09:46] <Callum__> ooops twb
[09:46] <twb> haha
[09:46] <Callum__> I meant AppArmor
[09:46] <Callum__> wth
[09:46] <Callum__> I saw your name and wrote it down
[09:46] <Callum__> DAMN, that happens on IRC as well as IRL
[09:46] <Callum__> what is wrong with me
[09:46] <Callum__> anyway, changed it, let's see what happens
[09:47] <MTecknology> OH! Is PlanetPlanet a python app that gets run on a cron and then just outputs static content?
[09:50] <Callum__> heh, I probably need to go down to the station to restart the printer
[09:50] <Callum__> they kind of lock up when this happens
[09:50] <Callum__> damn
[09:50] <Callum__> Let's try my OTHER printer there
[09:51] <Callum__> HP LaserJet 5Si from the mid 1990s =D still prints brilliantly at 600dpi even when its this old, plenty of toner left in it too
[09:51] <Callum__> sure it needs a little bit of time to warm up, but what other large printer doesn't these days apart from huge photocopiers
[10:18] <joschi> MTecknology: yes, it is
[11:07] <Tiibiidii> uhm i have a doubt... about my virtualbox setup... it's working fine, but i'm wondering if i could setup things better
[11:07] <Tiibiidii> i mean: my issue is with the hostname resolution
[11:07] <Tiibiidii> i have this ubuntu server vm...
[11:07] <Tiibiidii> from this i can ping fine the ip address of the host and vice-versa
[11:08] <Tiibiidii> (the network is bridged)
[11:08] <Tiibiidii> but the hostname resolution doesn't work
[11:09] <Tiibiidii> i mean... the local name resolution should be done by the local wifi router
[11:09] <Tiibiidii> i can obviously setup statically the ip address inside the /etc/hosts file
[11:41] <MTecknology> Is there anyone around that knows much about PlanetPlanet? I'm trying to separate the config directory from the template directory.. I have /opt/planet/nginx and /opt/planet/nginx-data. In nginx/config.ini I have template_files = nginx-data/index.html.tmpl  -  planet.py spits this out when I try to run that - planet.htmltmpl.TemplateError: Htmltmpl error: IO error while reading template 'index.html.tmpl': (2) No such file or
[11:56] <soren> MTecknology: I use absolute paths.
[11:56] <soren> MTecknology: Works great.
[11:56] <MTecknology> soren: that must get to be one massive line
[11:56] <soren> 349 characters.
[11:57] <soren> I'm a big boy. I can handle it.
[11:57] <MTecknology> :P
[11:57] <MTecknology> soren: how hard is it to make a nice theme for this when I'm all done? I think that's all I really have left
[11:58] <MTecknology> I won't do it now - it's 05:58 - sleepy time - that'll be a task for after my break
[12:00] <soren> MTecknology: As hard as html and css.
[12:00] <MTecknology> soren: I'm screwed..
[12:00] <soren> MTecknology: Me too. I stole mine.
[12:00] <twb> soren: HTML *is* hard.  Look how many people end up using tables for alignment rather than doing it properly
[12:01] <soren> twb: People fail at the simplest tasks. it's common.
[12:02] <soren> twb: I'm reasonsably good with html, but I still can't make anything that doesn't look like crap. Frustrating, really.
[12:02] <soren> I have the technical skills, but completely lack the aesthetic skills.
[12:02] <twb> Bah
[12:02] <MTecknology> soren: where did you steal yours from?
[12:03] <MTecknology> I know what this needs to look like in the end - and I'm not excited for it..
[12:03] <twb> The content provider is supposed to provide *content*, not style.
[12:03] <soren> MTecknology: Planet Ubuntu, I believe. It's ok, it was for planet.ubuntu-dk.org
[12:03] <twb> That's why I use w3m, since it doesn't implement CSS, I get a consistent style across all sites :-P
[12:03] <MTecknology> oh
[12:04] <MTecknology> soren: even an absolute path doesn't work...
[12:04] <soren> MTecknology: What's the error now?
[12:05] <MTecknology> http://dpaste.com/278785/
[12:05] <soren> MTecknology: index.html.tmpl doesn't look like an absolute path to me.
[12:06] <MTecknology> template_files = /opt/planet/nginx-data/index.html.tmpl /opt/planet/nginx-data/atom.xml.tmpl /opt/planet/nginx-data/rss20.xml.tmpl /opt/planet/nginx-data/rss10.xml.tmpl /opt/planet/nginx-data/opml.xml.tmpl /opt/planet/nginx-data/foafroll.xml.tmpl
[12:06] <soren> That's fine. You must not be using that configuration file.
[12:07] <MTecknology> http://dpaste.com/278786/
[12:09] <MTecknology> :(
[12:09] <soren> MTecknology: grep template_files  nginx/config.ini
[12:09] <MTecknology> soren: only the one line shows up..
[12:10] <soren> -v?
[12:10] <MTecknology> .....
[12:10] <MTecknology> rm -r nginx-data/*tmplc nginx-data/cache
[12:10] <MTecknology> that made it work..
[12:10] <soren> Awesome.
[12:11] <MTecknology> well... learned something new :P
[12:11] <MTecknology> thanks :)
[12:12] <MTecknology> Now I just need to add feeds and make it look pretty
[12:12] <MTecknology> easy peasy
[12:13] <MTecknology> and the fact that it generates static content - amazing
[12:28] <soren> Daviey: Ok, just so we're clear.. We can all commit stuff to the nova packaging branch now, but we review things before commit. Is that accurate?
[12:28] <Daviey> soren: If that is good for you?
[12:28] <soren> Daviey: That sounds perfect.
[12:28] <Daviey> soren: \o/
[12:29] <soren> Blimey, i have a lot of nova branches.
[12:29] <Daviey> lol
[12:29] <soren> 52 of them.
[12:30]  * soren cleans house
[13:37] <shauno> seem to be having problems routing from eth0 to ppp0; ufw is disabled and ip_forward is set in sysctl - what else am I missing?
[13:38] <soren> shauno: What are you trying to do?
[13:38] <soren> shauno: ...and what makes you say it doesn't work?
[13:39] <shauno> trying to pass traffic from a device on eth0 to a modem.  everything the device tries to reach times out
[13:40] <shauno> (it's an embedded device that hits a http server on a regular basis, but unfortunately doesn't have any settings for a http proxy)
[13:43] <soren> shauno: You (very likely) need masquerading set up.
[14:01] <shauno> soren: masq turned out to be just the google-fodder I needed, ty
[14:05] <soren> shauno: Sure
[14:07] <shauno> luckily I just need to capture a few minutes worth of traffic from it, so it doesn't have to be tidy - just functional
[14:21] <zul> soren: why is there two urls in the watch file?
[14:21] <soren> zul: One has snapshots, the other has releases.
[14:22] <soren> Click 'em :)
[14:22] <zul> soren: gotcha
[14:22] <zul> soren: so you should be able to do a uscan and it will download the tarballs for you?
[14:24] <soren> zul: Yes. that's what "debian/rules get-orig-source" does.
[14:24] <zul> sorry i never used uscan before
[14:24] <soren> zul: "bzr bd -S" also does it for you.
[14:25] <zul> ok...duh...need more caffine :)
[14:32] <soren> zul, Daviey: We need better defaults for networking.
[14:32] <zul> soren: like?
[14:33] <soren> zul: Something that doesn't suck.
[14:33] <soren> zul: What's the default in Eucalyptus?
[14:33] <Daviey> soren: I agree :)
[14:33] <soren> MANAGED-NOVLAN?
[14:33] <zul> soren: yeah :)
[14:33] <Daviey> soren: asking for avaliable IP addresses :)
[14:33] <zul> soren: having never touched eucalyptus i dont know :)
[14:33] <soren> Daviey: Oh. Which network mode?
[14:34] <Daviey> range, notation or comma separated :)
[14:34] <Daviey> soren: lemme confirm
[14:34] <soren> Daviey: Ta.
[14:35] <Daviey> soren: Interesting, http://pb.daviey.com/3vQ8/raw/ :)
[14:36] <Daviey> soren: but yes, MANAGED-NOVLAN
[14:36] <soren> Daviey: Ok.
[14:37] <soren> We don't have that :)
[14:37] <soren> It's on my TODO, though.
[14:37] <hggdh> JamesPage: I found the issue with my couchdb -- an interesting side effect
[14:37] <Daviey> soren: hmm
[14:38] <Daviey> soren: what does nova call the default?
[14:38] <soren> Daviey: VlanManager.
[14:38] <JamesPage> hggdh: yep - I found a few issues as well.....
[14:38] <Daviey> soren: How does VlanManager differ from Managed-novlan ?
[14:38] <soren> Daviey: It uses VLAN's :)
[14:39] <Daviey> :P
[14:39] <soren> Daviey: Other than that, it's mostly the same, IIRC.
[14:39] <soren> Daviey: ..but the choice of -NOVLAN over just MANAGED was intentional.
[14:39] <Daviey> then using that makes sense IMO.
[14:40] <soren> Let me think about it. The goal is simply to have something that works as well as possible out of the box with as few questions asked as possible.
[14:40] <Daviey> soren: Well... asking what IP addresses should be used (to the user), is pratical IMO.
[14:42] <soren> I'm unsure how valuable it really is. Sure, it should be debconfable, but if we can provide some defaults that give you a working cloud setup without asking any questions in 98% of the cases, that's a major win.
[14:43] <Daviey> totally
[14:43] <soren> We have a benefit over Eucalyptus, though.
[14:43] <soren> Eucalyptus didn't work well on a single box.
[14:43] <Daviey> soren: Ok... That would mean we need to expect a dhcpd to be avaliable on the network... ?
[14:43] <soren> Daviey: No.
[14:44] <soren> Nova works just fine on a single box, so the impact of choosing the VLANed option isn't big.
[14:44] <Daviey> soren: Ideally, we work well on one box - but equally work well on (n)
[14:45] <soren> Daviey: Sure. It's just that when you're going to n boxes, you need to do things differently anyway.
[14:45] <Daviey> soren: Are there doc's on the network settings nova currently supports?
[14:49] <[diablo]> guys is the default dns server in 10.10 server BIND or PowerDNS please? When choosing "DNS Server" in the installation process
[14:50]  * [diablo] hopes its PDNS
[14:51] <Daviey> [diablo]: Bind
[14:51] <[diablo]> doh
[14:51] <[diablo]> ah well :)
[14:52] <RoyK> [diablo]: apt-get install pdns-server .....
[14:52] <[diablo]> nod
[14:52] <[diablo]> damn I soooo love ubuntu server over RHEL
[14:53] <RoyK> hm... powerdns looks promising
[14:53] <RoyK> but then, bind works too :þ
[14:53] <Daviey> powerdns has a bounty on it, doesn't it?
[14:53] <Daviey> if you find a security bug
[14:54] <RoyK> dunno
[14:54] <RoyK> couldn't find one on first google
[14:54] <[diablo]> I like pdns cos I backend it to openldap
[14:57] <Daviey> No, i was thinking of djbdns.
[14:57] <Daviey> (tinydns)
[15:02] <soren> Daviey: Good question.
[15:02] <soren> Daviey: /me checks
[15:10] <soren> Daviey: The stuff on http://nova.openstack.org/nova.concepts.html looks accurate enough to me.
[15:10] <soren> Daviey: At a glance, at least.
[15:11] <Daviey> soren: thanks!
[15:17] <HackeMate> hello
[15:17] <HackeMate> is there a way to restart the dhcp3 server and modify the dhcpd.conf without being root or executing sudo?
[15:20] <pmatulis> HackeMate: no
[15:20] <HackeMate> a part of give 777 permissions
[15:20] <HackeMate> what about give the user a certain group?
[15:21] <HackeMate> i dont know what group is required for dhcpd
[15:23] <pmatulis> HackeMate: trying to circumvent standard security measures is not the way to go
[15:23] <HackeMate> that's why I find a best way
[15:23] <pmatulis> HackeMate: state why you are trying to do that
[15:23] <HackeMate> I have created a cgi that need restart the server
[15:24] <HackeMate> and I wont give 777
[15:24] <HackeMate> the cgi is a web-based app to add mac addresses to the dhcpd.conf
[15:25] <pmatulis> HackeMate: so allow the user that the cgi is run under to issue the commands you want without a password
[15:26] <pmatulis> HackeMate: do this by editing /etc/sudoers (ideally with visudo) and using NOPASSWD (man sudoers)
[15:27] <HackeMate> but then I will let the www-data to run everything sudo can execute
[15:28] <pmatulis> HackeMate: no.  you specify the commands he can issue
[15:28] <HackeMate> oh, that's exactly what i want
[15:28] <HackeMate> so i modify in sudoers what www-data can execute without password
[15:28] <HackeMate> perfect
[15:29] <highvoltage> stgraber: ^^^ now /that's/ completely insecure :)
[15:48] <HackeMate> another strange question, I get timeout in ssh when idle
[15:48] <HackeMate> I have configured the TCPKeepAlive and ServerAliveInternal but still die
[16:16] <soren> zul, Daviey: Ok, do you want to review https://code.launchpad.net/~openstack-ubuntu-packagers/ubuntu/natty/nova/ubuntu/ now?
[16:16] <zul> soren: sure
[16:17] <zul> soren: what about --FAKE_subdomain?
[16:17] <zul> is it still needed?
[16:18] <soren> zul: The docs say it is.
[16:18] <soren> zul: But they're lying.
[16:18] <zul> soren: k
[16:18] <soren> zul: Feel free to nuke it.
[16:19] <zul> soren: ack
[16:20] <zul> soren: looks good to me...im guessing daviey will look at it after the meeting and ill upload it
[16:21] <Daviey> two eyes is enough IMO.
[16:21] <soren> Daviey: I have two eyes.
[16:21] <Daviey> err, 4 eyes - perhaps :)
[16:21] <soren> Daviey: Yeah, that I don't have :)
[16:21] <soren> zul, Daviey: It will use r412. That seems fine to me.
[16:21] <zul> soren: ack
[16:22] <zul> soren: ill start the new changelog after i upload it?
[16:22] <soren> zul: So to build it, just check out the packaging branch and do "bzr bd -S".
[16:22] <zul> soren: right
[16:22] <zul> soren: already did it a couple of times ;)
[16:22] <soren> zul: Alright, just making sure :)
[16:23] <soren> Using "bzr bd" just may provide different results than dpkg-buildpackage directly. "bzr bd" is reproducable and consistent, so it's preferred.
[16:23] <soren> I know you know, I'm just elaborating in case someone is following along at home :)
[16:24] <MTecknology> soren: so.. 10:24 - I'm considering a nap... but the css and template editing it coming along :)
[16:25] <soren> MTecknology: masochist.
[16:25] <MTecknology> soren: You know how often I hear that?...
[16:25] <MTecknology> I'm starting to think it might be true. :(
[16:25] <soren> tmi
[16:25] <soren> :)
[16:26] <MTecknology> soren: wanna see it?
[16:26] <soren> Oh, the planet?
[16:26] <soren> phew. My mind was elsewhere.
[16:26] <soren> MTecknology: Sure.
[16:27] <MTecknology> soren: You'll have to add it to your hosts file - waiting on someone to decide if we're going to actually do it and update dns - 69.168.53.33 planet.nginx.org
[16:27] <MTecknology> woah... were you thinking something you shouldn't be thinking?
[16:28] <soren> MTecknology: WEll, you just revealed that people call you a masochist a lot.
[16:28] <soren> MTecknology: http://planet.nginx.org/ looks decent. I coudln't have done it better myself.
[16:28] <soren> And with that, I bid you all a good day.
[16:28] <soren> I must eat.
[16:28] <MTecknology> I should nap... or do homework
[16:28] <MTecknology> soren: thanks - I'll ttyl
[16:29] <yann2> what is the preferred disk format for windows VMs under KVM - raw, qcow2?
[16:29] <soren> /dev/null :)
[16:29] <MTecknology> soren: hey.. our ip's are incredibly close.. where do you live?
[16:29] <soren> Denmark.
[16:30] <MTecknology> oh.. nowhere near then
[16:31] <reggie_> hey folks I am having a permissions issue with synaptic can anyone help?
[16:32] <yann2> soren, if only...
[16:33] <zul> Daviey: can you review the packaging branch?
[16:36] <reggie_> can anyone help me figure out my permissions issue with synaptic
[16:36] <ScottK> reggie_: synaptic is a desktop package.  You should ask in #ubuntu.
[16:37] <zul> how is it different running servers on vm any different from running servers on bare metal other than hardware
[16:37] <reggie_> lol well I installed gui on top of serverso my first thing was to try and play with synaptic but for some odd reason it will not accept my password
[16:38] <reggie_> ubuntu tells me to come here since this is server
[16:38] <ScottK> Once you've installed the GUI, it's not.
[16:38] <reggie_> I have no issues installing from the command line but trying to use synaptic is a no go
[16:39] <ScottK> So the server solution would be don't use synaptic.
[16:39] <reggie_> ScottK, if you dont want to help I can understand but I would still like to find out the root of this problem
[16:40] <ScottK> reggie_: Fine, but it's off topic for this channel.
[16:41] <reggie_> I understand
[16:41] <reggie_> I will try back in ubuntu then
[16:41] <ScottK> Good luck.
[16:46] <reggie_> ikonia?
[16:46] <reggie_> i am here ikonia
[16:46] <ikonia> hello
[16:46] <ikonia> do you want to explain the issue for me a bit please ?
[16:47] <reggie_> hey how are you thank you for helping me out
[16:47] <reggie_> yes sir
[16:47] <reggie_> first i'm a win admin tryingt o incorporate ubuntu server into my network
[16:47] <ikonia> ok
[16:47] <reggie_> so i am fairly new with all of this but i know a little bit of the basics
[16:48] <reggie_> so being a win admin my first gut instinct was to install the gui even though it is frowned upon
[16:48] <reggie_> so i did that and went to try out synaptic and it is giving me a no go with my password even though the password is correct
[16:49] <ikonia> reggie_: ok, so lets step back and run a few tests
[16:49] <reggie_> so we looked at my permissions with the groups command and it does appear that I have admin rights
[16:49] <ikonia> reggie_: firstly can you run "sudo apt-get update" ?
[16:49] <reggie_> yes sir
[16:49] <reggie_> i can
[16:49] <reggie_> i can install with no issues from command line
[16:49] <ikonia> reggie_: ok, can you do gksudo synaptic
[16:49] <ikonia> "gksudo synaptic" sorry
[16:50] <reggie_> one sec let me try
[16:50] <reggie_> yea it seems to come up when i run that command
[16:50] <reggie_> wow
[16:51] <reggie_> so what the heck is the issue from accessing from the gui directly
[16:51] <ikonia> reggie_: ok, so hang on, wait 20 minutes, then ping me and we'll run some more tests
[16:51] <reggie_> ohh ok thank you
[16:51] <reggie_> will do
[16:52] <hallyn> SpamapS: wanna chat upstart for a second?
[16:52] <hallyn> i think ivoks might join in too
[16:53] <hallyn> SpamapS: you hae a bp on making upstart more server-capable right?
[16:58] <zul> soren Daviey: upload done
[16:58] <zul> soren Daviey: i started a new release in the bzr branch as well
[17:00] <SpamapS> hallyn: yes!
[17:00] <SpamapS> hallyn: are you trying to upstart-ify something?
[17:01] <hallyn> sorry, need three ins now
[17:01] <hallyn> three minutes now
[17:06] <hallyn> SpamapS: ok.  so, no, not trying to upstart-ify anything,
[17:06] <hallyn> SpamapS: rather, we're wondering about the upstart behavior when stopping a service
[17:06] <hallyn> apparently, it doesn't just kill the service, but also all its children?
[17:07] <hallyn> which of course can prevent a service from doing an orderely shutdown
[17:09] <SpamapS> hallyn: upstart has no notion of "children"
[17:09] <SpamapS> hallyn: it tracks a process ID as the "job"
[17:10] <SpamapS> hallyn: and sends TERM then KILL to that pid
[17:10] <hallyn> ivoks: ^
[17:10] <ivoks> yeah
[17:10] <ivoks> upstart is an as***
[17:10] <SpamapS> hallyn: if the process dies and it has children, they get assigned to init (upstart) as zombies and handled
[17:10] <hallyn> SpamapS: i wanted to see if you'd run into that, for starters.  all right, lemme pull down the code
[17:10] <hallyn> bc i think you are wrong :)
[17:10] <SpamapS> ivoks: this is no different than an init script
[17:11] <ivoks> init script sends TERM to parent
[17:11] <ivoks> and waits
[17:11] <ivoks> upstart does massacre
[17:11] <ivoks> anyway, i worked around that... so, i'm not that angry any more
[17:12] <SpamapS> ivoks: massacre in what way?
[17:12] <ivoks> it kills children before it kills parent
[17:12] <hallyn> ivoks: i wasn' ttyring to soothe your anger, but rather wonderng whether there is something platform should do for usptart on server
[17:12] <\sh> ivoks: should we change to systemd? ;-)
[17:12] <ivoks> i have a specific parent
[17:13] <SpamapS> ivoks: that would be crazy
[17:13] <ivoks> that is terminating children for a minute
[17:13] <ivoks> but it can't finish cause someone sends term to kids
[17:13] <SpamapS> ivoks: what I think you're hitting is the kill timeout
[17:13] <ivoks> it's not
[17:13] <ivoks> it's TERM, -15
[17:13] <smoser> anyone able to help me. i'm a dolt. https://launchpad.net/~awstools-dev/+archive/awstools/+build/2060214
[17:13] <ivoks> i get children writting that they received -15
[17:13] <SpamapS> ivoks: was the *parent* killed though?
[17:14] <SpamapS> ivoks: see if it sends SIGKILL to the parent, then the children get assigned to init and it kills them off.
[17:14] <ivoks> SpamapS: i'm in the meeting right now, let me get back when that finish
[17:14] <SpamapS> ivoks: try raising kill timeout
[17:14] <ivoks> parent doesn't exit until all kids terminate
[17:14] <SpamapS> would bet thats the issue.
[17:14] <\sh> smoser: chmod 755 ec2-activate-license when it's there?
[17:14] <ivoks> this isn't apache or something simple
[17:15] <ivoks> these kids need to be terminate in a specific order
[17:15] <ivoks> terminated
[17:15] <hallyn> ivoks: thanks, let us know when you're out of the mtg, ttyl
[17:15] <ivoks> those kids also fork, so... it's mess :)
[17:16] <SpamapS> I can't find in upstart's code where it sends SIGTERM to anything except the main job pid
[17:16] <smoser> \sh, yes. thank you. i was tihnking it was a different error.
[17:16] <smoser> http://launchpadlibrarian.net/59504186/buildlog_ubuntu-maverick-i386.ec2-api-tools_1.3.57419-0ubuntu2%7Emaverick2_FAILEDTOBUILD.txt.gz is what i intended to point at
[17:17] <smoser> but never mind. i have to run i'll sort it out somtime.
[17:18] <\sh> smoser: Package java6-runtime-headless is a virtual package provided by: <- no candidates
[17:18] <smoser> right. but why ?
[17:18] <\sh> smoser: looks like no package provides java6-runtime-headless
[17:19] <smoser> hm..
[17:19] <\sh> smoser: default-jre-headless should provide that package (at least on maverick)
[17:19] <smoser> http://packages.ubuntu.com/maverick/java6-runtime-headless
[17:19] <SpamapS> hallyn: have to run out for a bit, but you should try asking keybuk as well.
[17:19] <reggie_> hey ikonia
[17:20] <hallyn> SpamapS: thx, ttyl
[17:20] <SpamapS> ivoks: I'd also be interested in what daemon.log says that upstart did.. whether it says it sent TERM/KILL to the main process and such
[17:20]  * SpamapS will bbl
[17:21] <\sh> smoser: yes..but the buildd does think it's not provided anyhow..
[17:43] <progre55> hi guys. I've set up exim4 on a server, but some emails I send from it are returned with error "550-Verification failed for <me@mysite.com>\n550-Unrouteable address\n550 Sender verify failed". Any suggestions, please?
[17:44] <blackxored> hello guys, I can't remember the name of this app, so I ask, a LAMP application which was intended for streaming audio/video, along with a pretty management interface, pops up some suggestions???
[17:45] <blackxored> i'm getting some scaped html i shouldn't
[17:45] <blackxored> ups sorry
[17:46] <RoyK> try apt-cache search or freshmeat or even google
[17:46] <Scunizi> I'm trying to scp from /var/www on one machine to /var/www on another machine and am prompted for the password to the remote machine. After entering it all files/directories respond with "Permission Denied" when trying to copy. Any assistance appreciated.
[17:48] <b0gatyr> salutations
[17:49] <ivoks> SpamapS: i'm back
[17:49]  * pmatulis wonders why his freshly installed maverick server shows a load of 0.90
[17:54] <ivoks> SpamapS: it sent TERM, but to all the processes
[17:56] <progre55> Scunizi: your user doesnt have permissions for the /var/www directory
[17:56] <progre55> and I'm guessing, on both servers
[17:56] <Scunizi> progre55: how do I add the user to www-data via cli?
[17:57] <Scunizi> progre55: would it be ... #useradd -g <user> www-data ?
[17:58] <ivoks> adduser user group
[17:58] <Scunizi> or #useradd -g www-data <user>
[17:58] <progre55> Scunizi: umm.. useradd
[17:58] <Scunizi> no -g ?
[17:58] <ivoks> adduser user group
[17:58] <Scunizi> k
[17:58] <ivoks> notice *adduser*, not useradd
[17:58] <Scunizi> ivoks: does it take effect immediately?
[17:59] <ivoks> user needs to log out
[17:59] <Scunizi> ok.. thanks.
[17:59] <ivoks> if you are adding user to www-data, you are probably doing something wrong
[17:59] <progre55> Scunizi: btw, why would you add a user to www-data just to scp?
[18:00] <progre55> Scunizi: just scp it to some other location, and then ssh into the server and mv it into /var/www
[18:00] <Scunizi> progre55: I was just trying to avoid that
[18:00] <progre55> that's a bad idea
[18:01] <Scunizi> progre55: why
[18:01] <progre55> for security reasons
[18:01] <Scunizi> progre55: so you don't want any other users to be listed in the www-data group other than apache?
[18:02] <progre55> www-data can run some critical processes, that you wouldnt want your ordinary user to run
[18:02] <progre55> or have access to
[18:03] <Scunizi> progre55: ok.. that makes sense.. then how do you remove a user from a group?
[18:03] <progre55> anyways, can anyone help out with exim4, please? )
[18:03] <progre55> Scunizi: hmm.. sec
[18:04] <hggdh> zul: I am in doubt about samba bugs 678498,659854,393012 -- are the first two really duplicates of the last one?
[18:05] <zul> first one and last one are probably duplicates
[18:06] <hggdh> zul: and, if they are, is it not a good idea to open a task for samba(Ubuntu) on 393012?
[18:06] <Scunizi> progre55: I could directly edit /etc/group and remove the user from there on the line that reads www-data:x:33:<user>
[18:06] <zul> hggdh:probably :)
[18:06] <ivoks> SpamapS: http://pastebin.com/n709p8pZ
[18:06] <hggdh> zul: doing it
[18:07] <zul> hggdh: merci
[18:07] <hggdh> zul: mon plaisir
[18:08] <progre55> Scunizi: usermod -G theNeededGroup username
[18:08] <Scunizi> progre55: thanks
[18:08] <progre55> wait
[18:08] <Scunizi> k
[18:09] <progre55> Scunizi: the needed group is the group that the user actually belongs to, not the want you want to remove
[18:09] <zul> hggdh: have you tried to run the test rig on natty since the last time we did it?
[18:09] <progre55> after this comman, the user will only belong to the group you specified here
[18:09] <hggdh> zul: no, I was waiting for the new Natty kernel. I will try it today
[18:09] <progre55> you have have a list of comma-separated groups
[18:09] <progre55> Scunizi: ^^
[18:10] <zul> hggdh: cool...
[18:10] <Scunizi> progre55: ah.. so that command re-associates a user to one group or several groups in a comma-separated list?
[18:10] <progre55> yep
[18:11] <Scunizi> progre55: ok.. thanks
[18:11] <progre55> you can also : "id -nG username" to find out which groups the user belongs to
[18:11] <Scunizi> progre55: yea.. I was just thinking that I didn't know all the groups that the primary user on the server belonged to..
[18:16] <Scunizi> On a side note.. when logged into a headless server via ssh.. how do you log the user out?  log off the ssh session?
[18:18] <progre55> Scunizi: "exit"?
[18:18] <hggdh> Daviey: any news from the euca bug?
[18:19] <Scunizi> progre55: that's what I though.. just shutdown the ssh session.
[18:19] <progre55> but what's a headless server?
[18:19] <Pici> A server without a monitor.
[18:20] <progre55> ah )
[18:20] <progre55> makes sense =)
[18:21] <progre55> so, I've configured exim4 on my headless server, and sending email works fine, except I get error 550 for some email addresses. Any ideas, please?
[18:21] <progre55> "550-Verification failed for <me@mysite.com>\n550-Unrouteable address\n550 Sender verify failed"
[18:22] <hggdh> zul: natty still does not recognise the NetXtreme II card...
[18:23] <zul> hggdh: crap
[18:23] <hggdh> aye...
[18:23] <zul> hggdh: there is a bug number for this isnt it?
[18:23] <hggdh> zul: 6 machines rendered useless :-(
[18:24] <hggdh> zul: yes, there is, just a sec
[18:26] <zul>  https://bugs.launchpad.net/ubuntu/+source/linux/+bug/676245
[18:27] <hggdh> aye
[18:28] <Pici> progre55: It could be a few things: invalid/missing reverse dns, spf record missing
[18:32] <progre55> Pici: where do I check the reverse-dns settings?
[18:33] <Pici> progre55: Wherever you have your dns setup.
[18:34] <progre55> Pici: somewhere in resolv.conf?
[18:35] <Pici> progre55: no. You have a domain name right?
[18:35] <progre55> yep
[18:35] <progre55> Pici: saplo.com
[18:37] <Pici> progre55: The reverse dns entry for 79.125.15.2 doesn't point there though.
[18:37] <Pici> So wherever you update your dns records, you should have an option to set a reverse dns entry.
[18:37] <progre55> but it's an amazon server, and I'm using the default amazon dns
[18:38] <oru_work> greetings how to tell which php version i have installed. php --version returns some weired stuff
[18:39] <progre55> oru_work: php -version
[18:42] <Pici> progre55: From a little googling it looks like there is a way to setup reverse dns within Amazon's AWS, but since I don't have one or have access to one, I don't know where that setting would be.
[19:03] <progre55> Pici: oh thanks, I'll try googling then )
[19:03] <progre55> thanks for all the support
[19:12] <WALoeIII> Pici: don't try to send email from AWS, its not worth it
[19:12] <WALoeIII> use some relay
[19:12] <WALoeIII> authsmtp.net
[19:12] <WALoeIII> sendgrid
[19:12] <Pici> WALoeIII: Don't tell me, tell progre55 (who just quit)
[19:12] <WALoeIII> oh :\
[19:13] <WALoeIII> well, he'll figure it out the hard way :()
[19:28] <oru_work> can someone have a look at this please http://pastebin.com/702fYJ3Q
[19:36] <orudie> php -version returns the following complaints/errors and i'm not sure how to fix them m http://pastebin.com/702fYJ3Q
[19:40] <navanjr> help.... "mvn (Apache Maven >= 2.2.0) is not in the path" what have i missed?
[19:40] <navanjr> im attempting to install zenoss from source
[19:47] <\sh> whoosa mcollective + puppet == production rollout in less then 15 minutes...world record (rollout of 25 SOA services on live servers)
[19:50] <Daviey> \o/
[19:51] <navanjr> is this room a good place to find general server support?
[19:52] <navanjr> im new to ubuntu and having, what i think is, a path or version issue installing Zenoss
[19:54] <Daviey> \sh: How was your mcollective setup experience?
[19:56] <Daviey> navanjr: it is a pretty good place, but you sometimes find more active support in #ubuntu
[19:56] <navanjr> thanks!
[19:59] <\sh> Daviey: not good...we are working on it...it's sometimes not straight forward as it should
[20:00] <Daviey> \sh: That was my thought when i looked into it a while ago
[20:00] <Daviey> (I didn't finish setting it up)
[20:04] <dbowlby> I have a ubuntu 10.10 server that I use for a firewall.  It currently only uses about 3gb of disk space.  I'd like to move this from a hard disk to a usb key.  What steps would I have to perform to copy my installation to usb key and make sure it boots
[20:05] <\sh> Daviey: but the idea behind mcollective is awesome...I'm thinking about implementing something similar like that in python + rabbitmq + python-carrot / python-celery
[20:07] <dbowlby> \sh, cluster talk?
[20:07] <\sh> dbowlby: link? :)
[20:08] <dbowlby> \sh, using mq to pass stuff onto the cluster is an interesting notion
[20:08] <Daviey> dbowlby: rsync the files from the server to the usb pendrive, run grub-install on pendrive, and change /etc/fstab to point to new disks
[20:08] <Daviey> i think that is all
[20:09] <dbowlby> Daviey, cool, I'll give it a shot
[20:09] <Daviey> \sh: Oh aye... that is interesting... keep me posted if you get your hands dirty with code.
[20:09] <dbowlby> I already have a java mq sender and consumer
[20:10] <Daviey> \sh: I would love puppet to natively discover the puppetmaster :)
[20:10] <\sh> Daviey: well, it will be a part of my (DC)² project (launchpad.net/dc2)
[20:10] <Daviey> \sh: I am aware of your project :)
[20:11] <dbowlby> would you require the same software on each box, or have components located on a nfs share (depends on the app I assume)
[20:18] <dbowlby> \sh, Daviey, have you guys played around with perceus?
[20:18] <Daviey> common NFS defeats the purpose IMO
[20:18] <Daviey> dbowlby: no
[20:19] <dbowlby> Daviey, well you have to confirm that apps are loaded, and seems kinda silly to waste repetitive space for executables
[20:20] <dbowlby> Daviey, server specific space I'd understand
[20:21] <dbowlby> Not like the NFS couldn't be mirrored
[20:24] <\sh> dbowlby: nope...
[20:25] <dbowlby> are you guys looking at this from a cloud perspective or from a clustered perspective
[20:26] <\sh> dbowlby: I'm looking from a bare metal datacenter perspective :)
[20:26] <binBASH> hi sh ;)
[20:26] <\sh> binBASH: hey :)
[20:27] <binBASH> your tool sounds intressting ;)
[20:27] <dbowlby> \sh: what kind of things are you looking to manage?  Are you looking to plug in a server and have it autoconfigured once you enter the mac address?
[20:27] <\sh> dbowlby: that already works ;)
[20:28] <dbowlby> \sh: yeah I guess with dhcp in the mix you don't need to do even the mac step
[20:28] <dbowlby> \sh: so what are you looking to manage then, just event calls across the platform?
[20:31] <\sh> dbowlby: well, what we do is FAI for OS Deployment (which means, MACs maintained via mysql/couchdb enabled tftp server), auto inventarisation works already, and deployment too...now we are working on integrating puppetd single runs inside the FAI deployment, after reboot, we should have a working production system with apps etc. fully fledged...now we are looking at rolling app deployment with puppet and mcollective (eventually not mcollective but
[20:31] <\sh> the idea behind it)
[20:32] <\sh> dbowlby: when everything works here, we can increase the focus to VMWare ESX (which already works with PXE enabled VMs) and UEC (XEN/KVM etc.
[20:36] <\sh> dbowlby: amazon style cloud servers are different
[20:36] <dbowlby> \sh: nice
[20:36] <toast018> afternoon everyone!
[20:36] <orudie> is there a way to back one version in php from php 5.3.x to 5.2.x ?
[20:36] <dbowlby> \sh: yeah, it is different
[20:37] <dbowlby> \sh: I like the idea of making it work UEC :)
[20:37] <\sh> dbowlby: the FAI maintainer is working to integrate gpxe into FAI but I doubt, that amazon or rackspace will support bootps over internet ;)
[20:37] <dbowlby> \sh: awwww, why not ;)
[20:37] <dbowlby> \sh: I could put up a pptp :)
[20:37] <dbowlby> \sh: and hope it doesn't catch on fire
[20:38] <\sh> dbowlby: well...time is here my enemy...;) and right now, I'm the only one here @office who is working on the topic "bastard automation from hell"
[20:38] <toast018> how can I temporarily unblock .exe files in my mail server? I use ubuntu 8.04 with sendmail, spam assassian, and clamav
[20:38] <toast018> is it in spam assassain?
[20:39] <\sh> dbowlby: eventually I'll find the time in 2011 to attend another UDS and present the DC² project :)
[20:39] <jmgalloway> anyone here know how to use hsql?
[20:40] <soren> zul: In the future, I'd appreciate it if a release was done like so: "dch --release && debcommit --release && bzr bd -S && dput", and then, if it's accepted, "bzr push".
[20:40] <dbowlby> \sh: definitely an interesting thing to play with, you've got me thinking :)
[20:44] <toast018> anyone? :)
[20:45] <\sh> dbowlby: the real problem is sell automation to your management, 'cause in the beginning it costs time and money, but in the end it saves you money and time and (human) resources
[20:46] <dbowlby> \sh: in my world I have no problem selling that.  I have a particular use case that it would be VERY useful.
[20:47] <\sh> dbowlby: actually the use case is already starting at no. of servers >= 2 ;)
[20:48] <dbowlby> \sh: yeah, I was thinking selfishly :)
[20:51] <zul> soren: ack
[20:51] <orudie> is there a way to back one version in php from php 5.3.x to 5.2.x ?
[20:52] <orudie> anyone ?
[20:52] <zul> no
[20:52] <dbowlby> toast018: from what I've found from googling, "SpamAssassin does not block, it only scores according to configured rules. "
[20:52] <toast018> yea I found that as well just now. Seems that in version 2 you could choose to block it.
[20:53] <toast018> thats no good. :( lol
[20:54] <toast018> my surpervisor has an email with an exe in it coming from a legit company and needs to file to update a program.
[20:55] <toast018> o well just found out I got a bigger problem...
[20:56] <toast018> I changed my eth ports from eth0 to eth1 because it was going bad and even though everything works (email, client page, etc) I cant putty into the system any more it times out. Also the external IP times out as well.
[21:00] <pindemon> somebody here using zarafa+exim?
[21:02] <zul> SpamapS: there is a mysql merge with your name on it
[21:03] <dbowlby> toast018: sounds like you're toast ;)  Just have them rename the exe to a .hateemail
[21:04] <dbowlby> toast018: you could buy some time with management if you told them e-mail is so 80s and everyone sends their files via SMS now.  They'll spend hours trying to figure it out.
[21:05] <toast018> dbowlby: That was my suggestion. I hope they do that instead of forcing me to change all the rules to allow one email. lol. I wished that would work
[21:21] <toast018> I have adjusted the firewall to allow ssh but still cant connect on port 22 while using eth1
[21:21] <toast018> it works fine on eth0
[21:22] <RoyK> pastebin ufw status
[21:22] <RoyK> and also dmesg output
[21:24] <toast018> the ufw status is showing that all traffic from my subnet is allowed to port 22
[21:25] <RoyK> toast018: what's your ip?
[21:25] <RoyK> I might try to do an ssh connect from here if you like
[21:26] <RoyK> ufw iptables rules won't usually deal with which interface the traffic comes from
[21:29] <toast018> 12.52.251.72
[21:29] <toast018> I think its a firewall issue. It was not enabled. when I enabled it port 22 opened but all others closed.
[21:32] <RoyK> no contact with 12.52.251.72:22 from here
[21:32] <toast018> I blocked it from outside my subnet. Try port 25
[21:33] <toast018> I just looked on mxtoolbox.com and it shows it open but always like a second look... lol
[21:33] <RoyK> doesn't make ssh debugging easier
[21:33] <toast018> I got it working on my network now. sorry...
[21:33] <toast018> thanks for the help... :)
[21:33] <RoyK> I'd use fail2ban or denyhosts for blocking bots
[21:33] <RoyK> and just leave it open elsewhere
[21:34] <RoyK> ufw is usually down, if you enable it, it'll block everything not specifically allowed
[21:36] <toast018> Ill look into fail2ban and denyhosts. I am still learning as I go here.
[21:36] <toast018> total newb...
[21:36] <RoyK> nothing bad in being a newbie
[21:37] <RoyK> I was a linux newbie myself 15 years ago....
[21:38] <toast018> im forcing myself to not install any gui at all on the servers ... Im hoping that will help me in the long run
[21:38] <RoyK> it will
[21:38] <RoyK> you can be certain of that
[21:39] <toast018> even setup a box at home with no gui so I can log into it and mess around. if I would start keeping better notes I would be doing alot better... lol
[21:40]  * RoyK is teaching his boss vi these days
[21:40] <toast018> nice
[21:40] <toast018> what is the diffference between vi and vim?
[21:40] <\sh> RoyK: wasn't emacs for the bosses and VIs for the others? ;)
[21:40] <RoyK> vim is "improved"
[21:40] <RoyK> \sh: boss isn't really a bossy style
[21:41] <RoyK> which is good
[21:41] <RoyK> toast018: most distros use vim these days - for good reason
[21:41] <toast018> Thats what I use on our systems. took a min to get use to and still learning. lol
[21:42] <RoyK> toast018: try 'set -o vi' in bash
[21:43] <toast018> what will that do?
/whatiwroteyesterday
[21:43] <RoyK> it'll turn bash into vi
[21:44] <toast018> what would the benefits be?
[21:44] <guntbert> RoyK: and how to revert that step?
[21:44] <RoyK> well, if you like vi syntax, it's greeat
[21:45] <RoyK> guntbert: set -o emacs
[21:45] <RoyK> but then, it will only affect the current shell, so you can just quit the shell
[21:46] <guntbert> RoyK: thats understood :-) - but I always ask for/tell the way out :-)
[21:47]  * RoyK really wants this issue with his controllers to get fixed so he can get 2x110TB online
[21:58] <Egonis> I have a basic ufw configuration to masquerade traffic from eth1 to ppp0 -- however I have a /28 subnet from my ISP, which I am using on br0 successfully from a shorewall configuration previously. How do I add a route to allow traffic from the /28 subnet to pass through the bridge to ppp0?
[22:00] <Egonis> i.e., should I be using route add, or a direct arp command? I used to use 'proxyarp' in shorewall, which did the job nicely, but wanted to try to achieve this through ufw and alittle hard-coding
[22:13] <Egonis> I'm going to rephrase my question -- sorry: I have a /28 subnet from my ISP which I use for a few servers that are attached to eth1, eth1 is also serving 10.0.0.0 for LAN PC's. How do I add a route for my subnet to route through my ppp0 interface?
[22:13] <Egonis> I am using UFW
[22:45] <rimp> hello
[22:46] <rimp> is there anyone here who can help me
[22:48] <Pici> We won't know until you ask a question.
[23:08] <SpamapS> zul: I'll get started on the mysql merge tonight. :)
[23:10] <SpamapS> ivoks: ack, just got your pastebin message (sorry been away from IRC for a bit) ... I don't see any upstart messages there. upstart should log what its doing in daemon.log
[23:11] <ivoks> SpamapS: that's all i have in daemon.log
[23:12] <ivoks> SpamapS: but you can see that all forks got TERM signal
[23:14] <ivoks> SpamapS: corosync is the parent that should stop crmd, lrmd, cib, pengine...
[23:14] <ivoks> SpamapS: but, those just die before it gets the chance to terminate them
[23:15] <ivoks> SpamapS: this is with 'expect fork'
[23:15] <SpamapS> ivoks: yes, but what I don't see is if anything got KILL first.
[23:15] <ivoks> they didn't get KILL
[23:15] <ivoks> they got TERM
[23:15] <SpamapS> ivoks: did corosync get KILL ?
[23:15] <ivoks> no
[23:15] <SpamapS> hm ok
[23:16] <ivoks> SpamapS: on line 119 is example how it should look like
[23:16] <SpamapS> I spent a little of the time digging around in upstart's code and I still haven't seen where anything except the single, forked main process is sent any signals
[23:17] <ivoks> SpamapS: well, lines 0 -> 117 are result of 'stop corosync, shilw lines 121 -> 409 are result of pkill -TERM corosync
[23:17] <ivoks> shilw?
[23:17] <ivoks> it must be getting late :)
[23:17] <ivoks> ^ while
[23:18] <SpamapS> ivoks: Indeed, it does seem that they all got sent TERM's .. can you pastebin the upstart job too?
[23:18] <ivoks> sure
[23:19] <ivoks> SpamapS: http://pastebin.com/sb4By88x
[23:19] <ivoks> SpamapS: i've tried without expect fork, 'service' and all i could find
[23:25] <SpamapS> ivoks: did you try 'expect fork 1' ?
[23:25] <SpamapS> I think thats the default..
[23:25] <SpamapS> but not sure
[23:26] <SpamapS> ivoks: actually, 'expect daemon' might be more appropriate w/ corosync
[23:27] <ivoks> SpamapS: i can try
[23:29] <ivoks> SpamapS: same results
[23:29] <ivoks> SpamapS: this is lucid
[23:33] <SpamapS> ivoks: interesting.. corosync might not daemonize in the usual way..
[23:34] <SpamapS> ivoks: I believe if you take out 'expect fork' it may actually work properly
[23:34] <ivoks> i think i've tried that
[23:34] <ivoks> but i can try it again
[23:34] <SpamapS> ivoks: the problem is corosync seems to just close its stdin/stdout/stderr, drop the controlling terminal, and then declare itself the leader of its process group, without ever forking
[23:34] <ivoks> SpamapS: hm.... just noticed:
[23:35] <ivoks> corosync start/running, process 1194
[23:35] <ivoks>  1195 ?        Ssl    0:00 /usr/sbin/corosync
[23:35] <SpamapS> ivoks: ok maybe strace deceived me
[23:37] <SpamapS> ivoks: corosync actually calls clone, I wonder if upstart can't handle clone
[23:38]  * SpamapS is preparing a stack of questions for keybuk on Thursday. :)
[23:38] <ivoks> :)
[23:38] <ivoks> that's his last day?
[23:39]  * ivoks needs some sleep
[23:39] <ivoks> SpamapS: without expect fork it doesn't work (cause, i guess, there's no pid 1194)
[23:39] <SpamapS> ivoks: not sure, but I am going to be chatting with him about upstart.
[23:40] <SpamapS> ivoks: right, it may need to be 'expect clone'
[23:40] <ivoks> :)
[23:40] <ivoks> ok, i really have to go
[23:40] <ivoks> SpamapS: thanks you
[23:41] <ivoks> thank
[23:41] <SpamapS> http://paste.ubuntu.com/535702/
[23:41] <SpamapS> thats the startup sequence..
[23:42] <ivoks> yeah, it clones
[23:43]  * SpamapS will be especially happy when upstart supports chroots
[23:43] <ivoks> SpamapS: good night
[23:43] <ivoks> :)
[23:46] <SpamapS> ivoks: I feel a new upstart bug report coming. :)
[23:47] <SpamapS> ivoks: there's a new "proc adapter" that is coming, that will probably be far more reliable than the ptrace method used to implement "expect fork"
[23:47] <SpamapS> ivoks: but for now, I'd say if a process clones instead of forks to daemonize, it cannot be managed via upstart in the regular way
[23:47] <SpamapS> ivoks: HOWEVER, you *can* manage it with tasks
[23:48] <SpamapS> ivoks: which will at least let you start it, and pkill it, based on upstart events
[23:55] <jiboumans> smoser: what's the right way to set a hosts hostname via cloud-init? http://ubuntu-smoser.blogspot.com/2010/03/introducing-cloud-inits-cloud-config.html implies there's currently no 'proper' way
[23:55] <SpamapS> jiboumans: o/ !
[23:55] <jiboumans> SpamapS: o/
[23:55] <SpamapS> jiboumans: couldn't you just shove the hostname into /etc/hostname and run 'hostname' ?
[23:56] <jiboumans> spamaps: i can via runcmd: but i happen to know cloud-init is already setting it to the value of the ec2 metadata service
[23:56] <jiboumans> i want it to use a different value though (a hostname from my domain not, ec2s)
[23:57] <jiboumans> runcmd: - ["cat my.host > /etc/hostname"] - ["hostname"] would probably do it
[23:57] <jiboumans> but seems... ugly :)
[23:59] <gl1d3r> Hi