[00:20] <lifeless> hallyn: I have no opinion :P
[00:21] <lifeless> hallyn: #ubuntu-devel; feel free to ping poolie, jam, jelmer, james_w, barry as particular udd champions
[01:11] <hallyn> lifeless: actually, assuming you don't see that problem with oneiric containers (i don't), I guess we'd have to get that bug SRUd for it to help you out
[01:20] <hallyn> oh yeah.  that lxcguest comes from ppa anywa!  (for lucid)
[01:26] <lifeless> hallyn: :)
[01:35] <DanaG> hmm, trying to run byobu on serial console, just for the heck of it.
[01:35] <DanaG> It just silently hangs.
[01:36] <twb> byobu /dev/ttyS0 115200
[01:36] <DanaG> No, I mean, the server is the target.
[01:37] <twb> Can't remember if I've done that before
[01:37] <twb> I don't see why it wouldn't work
[01:37] <DanaG> I mean, I have getty listening on the tty.
[01:38] <DanaG> Oh, and aptitude also hangs.
[01:39] <DanaG> Looks like it's anything ncurses.
[01:40] <DanaG> Not that I need serial console very often, when I have SSH.
[01:40] <lifeless> hallyn: hey, so inside the container what is poweroff *expected* to do today? nothing? poweroff successfully, but slowly? something else?
[01:41] <hallyn> lifeless: it tells init to kill all tasks, and when utmp reflects poweroff, the lxc monitor (parent of its init) kills init and cleans up the container
[01:41] <DanaG> https://bugs.launchpad.net/ubuntu/+source/linux/+bug/329012?comments=all
[01:41] <hallyn> what we want is for a way for init to just tell the kernel 'i'm shutting down" and the kernel knows to either shut down the hardware, or kill init
[01:41] <DanaG> Should I un-expire that?
[01:43] <lifeless> hallyn: ok, so if its not doing that and is just hanging, thats a bug right ?
[01:43] <lifeless> hallyn: [which the workaround of zerging /run is dealing with]
[01:45] <twb> DanaG: I have *definitely* run curses apps, including aptitude, over serial
[01:45] <twb> DanaG: so you have fucked up something -- perhaps the baud rate or parity bits are most likely
[01:49] <hallyn> lifeless: yes
[01:49] <lifeless> hallyn: should I file a separate bug about that, or is the current 'cannot restart' sufficient ?
[01:49] <TheEvilPhoenix> is there a guide for setting up a git repository system on ubuntu server somewhere?
[01:50] <hallyn> lifeless: I'm doing some changes to your lxc-start-aufs script, think i'll be pushing tomorrow
[01:50] <lifeless> hallyn: cool
[01:50] <hallyn> lifeless: if there is a different cause, pls file a new bug
[01:50] <hallyn> though i may ignore it until next week (lxc sprint)
[01:50] <lifeless> hallyn: I think there are two bugs (A: sudo poweroff -n doing nothing inside the container, and B: lxc-stop -n container breaking things)
[01:51] <lifeless> I haven't explicitly filed A at the moment.
[01:51] <twb> lifeless: the latter does a hard stop; it doesn't telinit 0 or anything
[01:51] <twb> Of course it breaks things
[01:51] <lifeless> twb: sure, which is the whole power management thing
[01:52] <twb> lifeless: hm, you do it via acpid signalling?
[01:52] <lifeless> twb: no, but I believe that is the plan
[01:52] <twb> I'm doing it by sending a signal to upstart
[01:52] <hallyn> lifeless: do you know of a generic way to know whether dhcp info will be in /etc/dhcp or /etc/dhcp3?
[01:52] <lifeless> twb: I'm just using lxc as provided, running it into the group, and then whinging to serge :)
[01:52] <hallyn> lucid keeps them under dhcp3, oneiric under dhcp...
[01:53] <lifeless> s/group/ground/
[01:53] <hallyn> but wondering how to be smarter about know where to check
[01:53] <twb> http://paste.debian.net/125045/
[01:53] <lifeless> hallyn: ideally we'd get it from libvirt, but libvirt doesn't recognise containers as domains yet.
[01:53] <twb> pkill -INT lxc-start   # Sends a ctrl-alt-del to containers' upstarts.
[01:53] <hallyn> eh i'll just do the other if the one doesn't exist
[01:54] <twb> Then, IIRC, I have upstart patched to make ctrl-alt-del.conf enter runlevel 0, not 6
[01:54] <qman__> so, my ntp stopped working
[01:54] <lifeless> (libvirt would know the ethernet address -> ip mapping from its dhcp server, and we should probably have the ethernet address in the external config
[01:54] <qman__> at some point, with no changes made by me
[01:55] <lifeless> hallyn: yeah, direct lookup works for me :>
[01:55] <Martyn> evening
[01:55] <qman__> and my clients say 'no server suitable for synchronization found' when I try to sync to it
[01:55] <twb> qman__: doing a long step or a short step?
[01:55] <qman__> not sure
[01:56] <twb> (-g)
[01:56] <qman__> the time difference is only a few minutes
[01:56] <twb> Dunno then
[01:56] <qman__> -g isn't a valid option to ntpdate
[01:56] <twb> tcpdump
[01:56] <twb> qman__: ntp isn't ntpdate.  ntpdate is obsoleted by current ntpd
[01:57] <qman__> well, I was trying to sync clients (ntpdate) to the server (ntpd)
[01:57] <qman__> but ntpd clients are also not syncing
[01:57] <twb> qman__: that's not how it works
[01:57] <twb> ntpdate sets the clock ONCE, and then it drifts out of sync forever.
[01:58] <twb> ntpd contains both client and server parts; you're supposed to install it on every host.
[01:58] <qman__> I did, and it was working, for a couple years
[01:58] <qman__> but it has suddenly stopped
[01:58] <twb> Historically you'd have to use BOTH ntpdate (once, at boot) to get the clocks set up, and then run ntpd to keep them in sync.  Nowadays you can just use ntpd -q -g or so for the initial large step.
[01:58] <qman__> guessing at the server side since all my clients are all over the place
[01:59] <lifeless> hallyn: I'll file a separate one for A.
[02:04] <lifeless> hallyn: I think the -n on ssh in that script is a mistake
[02:04] <lifeless> hallyn: as it will prevent interactive use if someone wants that
[02:04] <lifeless> hallyn: you might like to drop it as you edit
[02:07] <qman__> stopping ntp, running ntpd -q -g, then starting ntp again seems to have fixed the problem
[02:07] <qman__> not sure what happened
[02:08] <twb> qman__: your clocks are presumably more out-of-sync than you realized
[02:08] <DanaG> twb: it was working a little while ago, and then stopped working when I tried to launch a new instance.
[02:08] <qman__> that was on the server, now all the clients are able to sync
[02:08] <DanaG> Or rather, a session started on local tty (uvesafb), then attached via serial, worked.
[02:09] <qman__> the server's clock was actually accurate within a minute
[02:09] <qman__> it was some of the clients that were off by as much as five or ten minutes
[02:10] <twb> Isn't uvesafb that really horrible pre-KMS one from the gentoo guy?
[02:11] <DanaG> Yeah, but it's on a server with an ASPEED graphics chip.
[02:11] <DanaG> I really wish ATI/AMD would make an IPMI ES1000.
[02:11] <DanaG> Or such.
[02:12] <DanaG> Heck, ASPEED doesn't even offer EDID.
[02:12] <twb> qman__: dunno, man
[02:12] <twb> qman__: ntpd is supposed to restart automagically whenever the network interfaces change, maybe that hook isn't there
[02:13] <twb> And you can't use the normal vesafb driver?
[02:13] <DanaG> uvesafb is more flexible.
[02:13] <DanaG> So I can do fbset.
[02:13] <twb> OK, I didn't think anyone would bother to rice a server that much
[02:13] <twb> But fair enough
[02:13] <DanaG> Well, it's useful in the rare case I want to connect using my netbook.
[02:14] <DanaG> I have it go to 1024 by default.
[02:14] <qman__> I only have one system I actually connect to the local console with
[02:14] <qman__> the rest I just SSH
[02:14] <DanaG> And unfortunately, HP didn't give us the serial-over-lan feature the ASPEED chip supports.
[02:14] <qman__> and the only reason I use that one's local console is IRC
[02:14] <DanaG> Yeah, ssh most of the time for me, too.
[02:15] <DanaG> With my laptop, I often connect to its serial-over-lan port to grab kernel stacktraces.
[02:15] <DanaG> But the server shouldn't be panicing. =P
[02:15] <DanaG> http://paste.ubuntu.com/658358/
[02:15] <DanaG> Oh, and lemme' see if it works via ssh.
[02:15] <twb> DanaG: for oopses, you could use netconsole
[02:16] <twb> DanaG: that basically just sends printks to a specified IP as UDP packets, and you can netcat them out
[02:16] <DanaG> Once, I had a panic on my netbook, before it even initialized intel KMS.
[02:16] <DanaG> Okay, so starting byobu via ssh, detaching, and reattaching via serial works.
[02:16] <DanaG> Well, aside from drawing glitches.
[02:17] <twb> I'd be interested to know if this is screen's fault, or byobu's
[02:17] <DanaG> ah, glitches were due to putty window being smaller than the console.
[02:17] <kirkland> me too
[02:17] <DanaG> Aptitude is doing the same thing.
[02:17] <DanaG> Aptitude within the reattached byobu, works.
[02:18] <twb> I wish tmux would magically do everything I use in screen, so I could switch to it
[02:18] <twb> ICBF doing all that C coding myself
[02:19] <DanaG> oh yeah, so I managed to work around HP's SMBIOS error, and get ipmi_si to load with a manually specified address.
[02:20] <DanaG> Now, what useful things can I do with it, in-band?
[02:24] <DanaG> hmm, ipmievd doesn't see the ecc event I inject.
[02:25] <hallyn> lifeless: yeah, tough call.  Interative may get confusing.
[02:26] <lifeless> hallyn: indeed ;)
[02:26] <lifeless> found out postgresql-8.4 was also nerfed by the lxc-stop thing
[02:26] <lifeless> now I have to figure out why sshd in the container isn't detecting the end-of-process at the far end
[02:27] <hallyn> lifeless: so after you do 'poweroff', if you do 'lxc-ps --name <container>', does it show postgresql still running?
[02:27] <DanaG> Yup, no log of events I inject.
[02:28] <lifeless> hallyn: let me fire it up
[02:28] <DanaG> okay, nothing on ipmievd.
[02:29] <lifeless> hallyn: is this relevant ? :
[02:29] <lifeless>  sudo lxc-start -n lucid-test-lp3
[02:29] <lifeless> lxc-start: No such file or directory - inotify event with no name (mask 32768)
[02:29] <lifeless>  * Setting up resolvconf...
[02:30] <lifeless> hallyn: sudo lxc-ps -n lucid-test-lp3
[02:30] <lifeless> CONTAINER    PID TTY          TIME CMD
[02:30] <lifeless>             7633 pts/6    00:00:00 sudo
[02:30] <lifeless>             7634 pts/6    00:00:00 lxc-ps
[02:30] <lifeless>             7635 pts/6    00:00:00 ps
[02:30] <hallyn> i think you have to do --name
[02:30] <lifeless>  sudo lxc-ps --name lucid-test-lp3
[02:30] <lifeless> CONTAINER    PID TTY          TIME CMD
[02:30] <lifeless> lucid-test-lp3  6878 ?        00:00:00 init
[02:30] <lifeless> lucid-test-lp3  7400 ?        00:00:00 sshd
[02:30] <hallyn> we shoudl really make -n work for consistency
[02:30] <hallyn> ok
[02:30] <lifeless> I'll file a bug on -n
[02:30] <hallyn> thanks - it's annoying
[02:31] <hallyn> i don't know what the inotify thing means
[02:31] <lifeless> bug 820720
[02:32] <hallyn> thx i'll add that to the pile
[02:33] <hallyn> drat, no aufs in ec2 (and i just lost my other box for a bit)
[02:33] <lifeless> doh!
[02:33] <lifeless> this poweroff thing is baseline lxc - not the aufs script
[02:33] <hallyn> yeah
[02:34] <hallyn> but i wanted to test the aufs script some more before shipping :)
[02:34] <lifeless> of course :)
[02:34] <hallyn> i'll start working on the -n thing meanwhile :)
[02:36] <DanaG> okay, got ipmievd to work.
[02:36] <DanaG> so anyway, my pastebin was strace of byobu.
[02:36] <DanaG> And it says it wrote a message to stdout, but actually didn't.
[02:37] <DanaG> And when I did as it said, and chmodded /var/run/screen/ to 777, it told me to chmod it to 775.
[02:37] <DanaG> if I sudo strace -F fgconsole, it hangs here:
[02:37] <DanaG> open("/proc/self/fd/0", O_RDWR
[02:41] <DanaG> and aptitude does this:
[02:41] <DanaG> readlink("/proc/self/fd/0", "/dev/ttyS5", 4095) = 10                    stat("/dev/ttyS5", {st_mode=S_IFCHR|0600, st_rdev=makedev(4, 69), ...}) = 0                        open("/dev/ttyS5", O_WRONLY
[02:46] <lifeless> hallyn: I've just switched to ssh  -q -n -T
[02:46] <lifeless> hallyn: stops a hang on the ssh process if its output is a little racy, and squelches the 'added ... to hosts file'
[02:50] <lifeless> bah, the -T isn't sufficient
[02:51] <DanaG> hmm, off to walk the dogs.
[03:35] <twb> Ugh.  xdm's postinst assumes it's *running* at configure time (as at lucid):
[03:35] <twb> (transmute)root@trimserver:/# dpkg-reconfigure xdm System start/stop links for /etc/init.d/xdm already exist.  cat: /var/run/xdm.pid: No such file or directory
[03:58] <hallyn> lifeless: actually i think i'm going to drop the 'ssh' bit in there and use lxc-monitor instead
[03:58] <hallyn> trick is i need to tweak lxc-monitor to be willing to exit when the container stops :)
[04:29] <jdevel> i'm setting up a mail server and I wanted to know some personall favorite spam filters if anyone would share.
[04:40] <twb> jdevel: postgrey
[04:40] <twb> If you are reading the payload to detect spam, you've lost
[04:41] <twb> But we do also run crm114 for a couple of abysmally stupid users, for whom handfuls of spam a day get past non-payload techniques.
[04:42] <twb> jdevel: checking SPF records, and adding one to your own domain, is also obviously a Good Thing
[04:42] <jdevel> twb, I appreciate your feedback
[04:42] <jdevel> I was reading that spamassassin is a resource hog
[04:42] <twb> it is
[04:43] <jdevel> postgrey isn't that just a white/grey/black list type of filter?
[04:44] <jdevel> seems a lot of people use seperate machines for filtering spam
[04:44] <DanaG> weird... I just got a spew of the banlist.
[04:44] <twb> jdevel: https://secure.wikimedia.org/wikipedia/en/wiki/Greylisting
[04:45] <jdevel> i'll check it out
[04:45] <twb> jdevel: it drops messages from peers that do not adhere to the SMTP specifications.
[04:45] <jdevel> also, for virus detection.. integrated into postfix I'm guessing clamav?
[04:45] <twb> Unless you have windows users, you don't need virus detection
[04:46] <twb> http://linuxmafia.com/~rick/faq/index.php?page=virus
[04:46] <jdevel> yes could be a number of system types
[04:49] <twb> https://secure.wikimedia.org/wikipedia/en/wiki/SPF_record if you missed that comment.
[04:54] <ScottK> jdevel: The standard Ubuntu server setup is amavisd-new + clamav + spamassassin - see the server guide in /topic for information on setting it up.
[04:55] <ScottK> postgrey is a good addon if you want greylisting.
[04:55] <jdevel> I honestly think I may just forgo the spam filtering and virus scanning
[04:56] <jdevel> If I need to deploy this in a more production like environment with considerable traffic I'll approach those avenues
[04:57] <twb> more traffic = more load
[04:57] <jdevel> I'll just roll with postfix, dovecot, mysql, virtual
[04:57] <twb> payload scanning doesn't scale, which is why it *isn't* useful for high-traffic environments.
[04:58] <jdevel> which is why I'll check out postgrey when I need to get into spam filtering
[04:59] <twb> jdevel: and SPF
[05:03] <DanaG> Say, is there some package I can install to have my system e-mail me if major stuff happens in syslog?
[05:03] <hallyn> lifeless: ok, i've got lxc-start-ephemeral mostly working (for oneiric-on-oneiric containers).  Only hitch is dhclient is failing with some error about libc.  aufs-related, I woudl assume.
[05:03] <DanaG> Major stuff would be, say, ECC errors.
[05:06] <hallyn> Daviey: (going by most likely to be helpful timezone :)  Any chance I could get a push to lxc with http://people.canonical.com/~serge/lxc_0.7.4.2-3ubuntu6.debdiff   ?
[05:06] <lifeless> hallyn: interesting
[05:06] <hallyn> lifeless: you didn't see that at all?
[05:06] <lifeless> hallyn: nope
[05:07] <lifeless> hallyn: but for ephemeral I've only tried lucid on (natty, oneiric)
[05:07] <hallyn> dhclient: error while loading shared libraries: libc.so.6: cannot open shared object file: Error 116
[05:07] <ScottK> hallyn: It should probably wait until after Alpha 3 is out since it's in Main.
[05:08] <DanaG> ah: https://help.ubuntu.com/community/Smartmontools
[05:08] <hallyn> ScottK: drat, good point, thanks
[05:08] <jdevel> twb, would you recommend using fail2ban on postfix/dovecot?
[05:09] <twb> I don't know about postfix/dovecot, but I would not use fail2ban for SSH
[05:09] <twb> http://cyber.com.au/~twb/doc/iptab
[05:10] <hallyn> lifeless: go figure!  an ephemeral lucid container on oneiric can dhclient jsut fine!
[05:11] <lifeless> hallyn: \o/
[05:11] <DanaG> Say, how can I make my server shut down upon, say, the 5th time of pressing the power button?
[05:11] <jdevel> twb, thanks
[05:12] <twb> DanaG: what's wrong with doing it on the first press?
[05:12] <DanaG> It's too easy to hit accidentally, when I'm shoving the server around on the floor.
[05:13] <DanaG> Actually, second time would be better than fifth time.
[05:13] <twb> DanaG: buy a molly guard.
[05:13] <DanaG> hmm, I could tweak /etc/acpi/shutdown.sh to write a count to /tmp.
[05:14] <DanaG> Some file in /tmp.
[05:14] <twb> https://secure.wikimedia.org/wikipedia/en/wiki/Molly_guard
[05:14] <DanaG> Har, my 8GB Patriot drive is: Kingston Technology Company Inc. 2 GB USB stick
[05:15] <DanaG> Oh yeah, and my server has a jumper (that I attached a button to), to inject an NMI into the system.
[05:16] <DanaG> All I get in dmesg is "Uhhuh.  NMI received for unknown reason 2c.  Stumbling, but attempting to continue."
[05:16] <jmazaredo> can i see the ftp uploads of files, creation of folder in my ftp log or just the login/fail/ip only?
[05:16] <hallyn> lifeless: the lxcguest for lucid is waiting to build in ubuntu-virt ppa (which is where the ubuntu template grabs it from).  That ppa also will have the proposed lxc package with lxc-start-ephemeral.
[05:16] <lifeless> hallyn: very cool
[05:16] <hallyn> I did just also see the 'lucid-won't-shutdown' - I don't always see it though.  unsure what's the problem
[05:16] <lifeless> hallyn: I will give it a spin and get you some feedback
[05:17] <hallyn> lifeless: could be awhile before it gets built :)  but thx.  Oh, pls do look at the usage help (i changed it)
[05:17] <hallyn> good night
[05:18] <lifeless> ciao
[05:54] <jdevel> anyone feel free to chime in... I have ufw used to setup the chains for iptables etc.. ports I want open etc.. the usual FW stuff.  I also added fail2ban to the server.  I'm still a bit new to iptables and chains etc.. will the fail2ban chain rules interfere with the ufw chain rules?
[05:54] <jdevel> or can they co-exist together without issues
[05:55] <jdevel> I imagine it comes down to which is stacked on the top/bottom regarding enforcing rules
[05:59] <ScottK> Since they both modify the iptables config, you'd have to look at the results and evaluate if they were llikely to interfere.
[05:59] <ScottK> I don't use either, so no idea.
[07:28] <nandemonai> Hey guys. I don't suppose anyone knows if there is a working headless ps3 media server?
[07:29] <nandemonai> Ooo hang on, looks like ps3mediaserver can be run gui-less.
[07:30] <twb> nandemonai: if it needs X, there's Xvfb to make it headless
[07:31] <nandemonai> Hmm also a good idea. Thanks twb
[07:32] <nandemonai> I assume (haven't used it before) that I'd have to use ssh x forwarding to use it if going down the xvfb route?
[07:34] <twb> No, with xvfb the GUI is never displayed anywhere
[07:34] <twb> The client just thinks it is
[07:36] <nandemonai> Ahh ok.
[07:36] <nandemonai> Hmm thanks again, I'll look into it.
[07:38] <twb> If you actually need to interact with the GUI, you could of course use something like ssh -X or X0rfbserver
[07:38] <twb> The latter exports the X session over VNC, and it probably most appropriate.  Note that unlike Windows/OS X VNC servers, it exports a *new* session, and isn't tied to what's on the local display (if anything).
[09:02] <jamey-uk> I tried setting up a bridge as listed here https://help.ubuntu.com/11.04/serverguide/C/network-configuration.html but then I can't access the internet any more. What am I doing wrong here?
[09:04] <twb> jamey-uk: pastebin your network/interfaces
[09:04] <Daviey> hallyn: Just had a quick look... the /var/run entries should be /run, no?
[09:05] <twb> jamey-uk: here's a working one: http://paste.debian.net/125063/
[09:05] <Daviey> hallyn: and 'binary-install/lxc::" /could/ be handled by a .install file?  dh_fixperms should chmod it automagically, i'd have thought
[09:06] <jamey-uk> twb: http://pastebin.com/USUPKYrZ yeah I've immediately noticed that I spelt 'broadcast' wrong for br0, and my network parameter is 192.168.1.0 whereas I'm assuming it should be 192.168.1.1 (or should they both end in 0?)
[09:06] <twb> jamey-uk: you shouldn't be raising both br0 and eth0
[09:07] <jamey-uk> twb: oh, I clearly misunderstood it then… what should I be doing?
[09:07] <twb> jamey-uk: if eth0 is participating in the bridge, you just raise the bridge
[09:07] <twb> jamey-uk: see how I have no stanza at all for dmz or managed (which are my eth0 and eth1)?
[09:14] <jamey-uk> twb:	sorry, haven't edited interfaces much before, is raising when you put in the 'auto eth0' line?
[09:17] <twb> Basically, yes
[09:18] <jamey-uk> so something more like the example here? https://help.ubuntu.com/community/KVM/Networking#Creating a network bridge on the host
[09:18] <twb> If you don't have that the associated iface foo stanza doesn't matter
[09:18] <twb> Yes
[09:18] <twb> Not sure why they turn STP off, but whatever
[09:19] <jamey-uk> twb: so basically, br0 becomes what my eth0 was, and then a bridge will exist?
[09:19] <twb> Ye
[09:19] <twb> Yes
[09:20] <jamey-uk> Thanks, I'll give it a go :)
[09:48] <andyw> o/
[09:59] <Loocy> there are many apps and script but can any one recommend a good php or something stable userfriendly script for webbased 1.voice 2. video chat ? 3. i need some face book like bloging script. similer or which integrates with facebook accounts and its blogs.?
[10:04] <adac> hi guys. I've set up denyhosts and added a host of mine in the whitelist (hosts.allow: All: vhost.example.com) but denyhosts does not recognize the hosts DNS. It only works with IP... any ideas whats might wrong?
[10:05] <twb> netfilter doesn't support hostnames.
[10:06] <twb> iptables-restore and ipset will resolve them, but only once, at load time.
[10:10] <adac> twb, but it works on my other linux server and on my centos host.
[10:13] <jamey-uk> twb: so that seems to have worked, my original connection works and the bridge seems active. I've following the instructions at https://help.ubuntu.com/community/KVM/Networking#Converting an existing guest but the guest doesn't pick up an address. Do I have to follow the next steps re. DHCP?
[10:20] <twb> Sorry, I'm going home in a minute and I can't be bothered helping
[10:21] <jamey-uk> Okay, thanks anyway
[10:21] <jamey-uk> Could anyone else help me with getting my KVM bridged network working?
[10:24] <Loocy> there are many apps and script but can any one recommend a good php or something stable userfriendly script similer to face book like bloging script. similer or which integrates with facebook accounts and its blogs.?
[11:29] <jamespage> Daviey: iscsi target tests for i386 and amd64 completed OK for oneiric - I've updated the test tracker
[11:30] <Daviey> jamespage: you rock star.
[11:30] <_ruben> which iscsi target would that be? just curious
[11:31] <jamespage> _ruben: well I just tested against the iscsitarget package - http://iscsitarget.sourceforge.net/
[11:32] <_ruben> ah ietd
[11:33] <_ruben> (using scst.sf.net myself)
[12:52] <marshall> hey ubuntu
[13:09] <smb> smoser, utlemming FYI, bug 791850 bisected. I posted to xen-devel as I would not be completely at the bottom of "why". But in doubt/emergency we seem to have the option of reverting one patch for PV on HVM to be able to boot. But I am sure there should be a better solution.
[13:26] <RoAkSoAx> adam_g: for some reason i though you were writing one
[13:33] <KM0201> how can i add a current raid (/dev/md0) to fstab?.. if i manually mount it (sudo mount /dev/md0 /mt/mtpoint) it mounts just fine... when i try to add its uuid to fstab, i always get an error and have to "Skip or Manually recover".. if I delete the UUID i add to fstab, my machine boots normal, and I can manually mount the disk no problem...   http://pastebin.com/1yRNDGpy           if there's other info you need, just ask.
[13:36] <hallyn> Daviey: first off, as ScottK pointed out, we're in ffe, so pls dont' push anyway :)  I'll ask again on monday
[13:37] <hallyn> Daviey: secondly, my moral objections to /run aside, I think it's unnecesary and too early to switch the entries.  (It would make more delta between natty packages etc)
[13:39] <hallyn> Daviey: lastly, for the .install file, it would be nice.  I was thinking I couldn't do that one with .install, but maybe it's only renaming that doesn't work?
[13:39]  * hallyn looks at the manpage
[13:39] <hallyn> Daviey: yeah, .install should work.  I'll give that a shot.
[13:40] <Daviey> hallyn: Yeah, planend to queue it up.
[13:44] <hallyn> Daviey: ?
[13:45] <KM0201> how can i add a current raid (/dev/md0) to fstab?.. if i manually mount it (sudo mount /dev/md0 /mt/mtpoint) it mounts just fine... when i try to add its uuid to fstab, i always get an error and have to "Skip or Manually recover".. if I delete the UUID i add to fstab, my machine boots normal, and I can manually mount the disk no problem...   http://pastebin.com/1yRNDGpy           if there's other info you need, just ask.
[13:48] <hallyn> Daviey: not sure what you mean by 'queue it up'.  AIUI the FFE is voluntary, so if you dput it'll go through.
[14:18] <hallyn> Daviey: updated http://people.canonical.com/~serge/lxc_0.7.4.2-3ubuntu6.debdiff .  Will let it sit till monday now.
[14:38] <jdevel> hey twb
[14:38] <jdevel> you around here?
[14:41] <marshall> I've got an LDAP client that seems to cache ldap groups. When I restart nscd, it doesn't seem to recognize ldap groups.
[14:41] <hallyn> SpamapS: regarding bug 820675 (and all the others :)  I assume those are something you're going to track as you implement the upstart runlevel 1 nuttiness?  :)
[15:14] <zul> jamespage: ping how is the octopussy stuff coming?
[15:15] <jamespage> zul: Daviey is reviewing ATM pre-upload to NEW
[15:15] <jamespage> its OK and working
[15:15] <zul> jamespage: k the rsyslog ensemble stuff is almost done server done...need to write the client
[15:16] <Daviey> zul: whats that you are doing?
[15:16] <zul> writing ensemble formulas for rsyslog
[15:17] <Daviey> zul: What WI is that part of?
[15:17] <zul> the syslog one
[15:17] <zul> i probably should add that
[15:18] <Daviey> zul: the firehose?
[15:19] <zul> yeah or it could be in the orchestra one
[15:19] <Daviey> zul: erm, adam is already working on that
[15:20] <zul> erm...i didnt know that
[15:20] <Daviey> I might be wrong...
[15:20] <zul> i thought he was doing the nova ensemble formulas
[15:20] <lynxman> zul: hey :)
[15:20] <lynxman> zul: have you seen this one before? http://pastebin.ubuntu.com/658695/
[15:20] <Daviey> zul: He wrote puppet ones, but was then looking to also write ensemble ones
[15:21] <zul> lynxman: no have you tried nova-manage help?
[15:21] <zul> Daviey: hmmm....
[15:22] <lynxman> zul: yes, same result no matter which command I feed
[15:22] <zul> lynxman: open up a bug in launchpad for the nova project then
[15:22] <lynxman> zul: hah, it was the lxc line in nova.conf
[15:23] <lynxman> zul: which one is used now to enable lxc?
[15:23] <zul> --libvirt_type=lxc
[15:23] <lynxman> exactly what I was using, hmm...
[15:24] <lynxman> and now it works, I'll be damned
[15:24] <lynxman> zul: thanks
[15:54] <zul> smoser: have you seen this?  ssh -v ubuntu@pegasus.trellisnet.co.uk -p 8773
[15:54] <zul> damn it
[15:54] <zul> smoser:  https://bugs.launchpad.net/nova/+bug/820962
[15:55] <TREllis> zul: LOL
[16:04] <adam_g> zul: i had done the puppet modules for rsyslog and was considering redoing them in ensemble, cause it would be great to have along openstack stuff
[16:04] <adam_g> Daviey: ^
[16:04] <adam_g> i haven't started them yet, and if you have something ill use that instead, or help work on it with you if you'd like
[16:04] <Daviey> adam_g: Ah, i thought you had started them.
[16:05] <adam_g> Daviey: not yet, i did start a collectd formula, tho.. which'll compliment openstack similarly
[16:06] <CharlieSu> Hi. I'm on Ubuntu 10.04.1 LTS and when I install the CouchDB package and then try to run /etc/init.d/couchdb stop I get 'cd: 88: can't cd to /root' and the service never stops.  Anyone know what this may be?
[16:06] <Daviey> adam_g: What is the sitrep of the openstack formula's?  Are they good for next week?
[16:06] <Daviey> multi-node?
[16:06] <Daviey> CharlieSu: that looks like a bug :)
[16:07] <CharlieSu> Daviey: yeah I know it.  Do you know if there is a way to install a package like CouchDB and have it not automatically start after it is installed?
[16:07] <zul> adam_g: we need to work on the openstack/orchestra integration at the same time i think
[16:08] <zul> adam_g: i already have an ensemble formula for nagios at least
[16:08] <adam_g> Daviey: i want to modify them to make use of the new formula config stuff.. but yea, http://no-carrier.net/openstack.svg
[16:09] <adam_g> zul: cool, we can do that next week for sure
[16:10] <zul> adam_g: we can start on monday even ;)
[16:12] <adam_g> zul: is that rsyslog formula branch pushed anywhere?
[16:12] <zul> adam_g: not yet...i got pulled into something else
[16:12] <zul> adam_g: ill push it this afternoon (just about to go for lunch)
[16:13] <adam_g> k
[16:13] <Daviey> adam_g: diagram looks good!
[16:14] <adam_g> Daviey: there's still a couple of things i need to work out with scaling the swift storage up and down, but i hope to address those next week.
[16:14] <Daviey> adam_g / RoAkSoAx: Do i understand correctly that the formula needs to only reside on the consumer/user's machine.. not on the cobbler server?
[16:14] <Daviey> adam_g: sounds good.
[16:15] <adam_g> Daviey: using orchestra? i dont know. on ec2 its pushed to a S3 bucket for agent consumption. i think they're using webdav for that, running alongside cobbler?
[16:16] <Daviey> yeah.. webdav is the data store.
[16:16] <Daviey> I'm trying to make sure we haven't missed anything in the use case.
[16:17] <adam_g> is deploying via orchestra ready to go? is it available on any test cluster yet?
[16:17] <lynxman> Daviey: hey, I need sponsoring for ruby-bundler, it needs to go to universe
[16:18] <Daviey> lynxman: it's on my todo.
[16:18] <lynxman> Daviey: thanks a big bunch sir :)
[16:18] <SpamapS> Daviey: the formulas do only need to live on your local machine but are copied to the webdav service on deploy
[16:19] <Daviey> SpamapS: Great!  thanks.
[16:29]  * Daviey screams a little.
[16:29]  * SpamapS tries to toss M&M's into Daviey's gaping mouth
[16:30] <Daviey> CharlieSu: different packages do it in different ways, sadly i don't know off the top of my head on that one.  Please do raise a bug regarding the issue you encountered tho
[16:30]  * TheEvilPhoenix yawns and throws balls of paint at Daivey and SpamapS
[16:30] <Daviey> SpamapS: that would be wonderful.
[16:30] <Daviey> TheEvilPhoenix: That isn't very nice :(
[16:30] <TheEvilPhoenix> *cough* The ***Evil*** Phoenix *cough*
[16:39] <lynxman> TheEvilPhoenix: so... chaotic neutral or lawful evil? :)
[17:33] <robbiew> kim0: ping
[17:54] <MoooookiE> Hello, I changed the default ubuntu mysql config. But now if i want to start mysql with "service mysql start" i just geht this errors: http://privatepaste.com/4d276bd1a3 here is my config: http://privatepaste.com/ba2bdd68f6
[17:59] <aurigus> what happens if you simply run mysql?
[17:59] <aurigus> im too slow :/
[18:25] <ewj> I'm trying to upgrade an Ubuntu 8.04 server to new OpenSSH/OpenSSL/Apache versions for security reasons, but it looks like the 8.04 repos don't have them - is it possible to install newer versions of just those packages without doing a full distro upgrade?
[18:26] <astrostl> "maybe"
[18:26] <astrostl> you can find out by adding the repos for a higher version, and modeling it in aptitude
[18:26] <astrostl> e.g. add the lucid repo, and then aptitude -sy install openssh
[18:26] <patdk-wk> ewj, your going about it wrong :)
[18:27] <patdk-wk> the whole point of using an LTS release, is there is no security concerns
[18:27] <ewj> I'm sort of new to this, so that's def. possible, patdk-wk :)
[18:27] <astrostl> shouldn't 8 LTS have SECURITY patches though?
[18:27] <astrostl> i run into bugs constantly on lucid, but security stuff always seems to be backported
[18:27] <ewj> yes, I think it does, but I'm not seeing recent versions of, for instance, the OpenSSL libs showing up through aptitude
[18:28] <patdk-wk> you won't see recent versions
[18:28] <patdk-wk> cause the security issues are BACKPORTED
[18:28] <patdk-wk> you got audited? and they said you have old insecure crap?
[18:28] <ewj> shouldn't they be showing up as some sort of update, though? for instance, 'openssl version' tells me 'OpenSSL 0.9.8g 19 Oct 2007'
[18:29] <astrostl> look at the actual package changes on the repo web sites
[18:29] <ewj> yup, basically, an audit
[18:29] <astrostl> something can stay at version g and still have, say, security features from h rolled into it
[18:29] <patdk-wk> the audit should link you to the CSE's
[18:29] <patdk-wk> and you just cross reference them wit hthe ubuntu cse's that show it fixed
[18:30] <davros> lol isnt ubuntu supposed to be easy :/
[18:30] <ewj> http://packages.ubuntu.com/hardy/openssl says 0.9.8g-4ubuntu3.13 should be the most recent - is there a way I can verify that's the version we're running?
[18:30] <mdeslaur> ewj: your auditor is doing it wrong. Most linux distros backport security fixes to old versions. See https://wiki.ubuntu.com/SecurityTeam/FAQ#Versions
[18:30] <patdk-wk> mdeslaur, all auditors do it wrong :)
[18:30] <mdeslaur> patdk-wk: yeah, most of them :P
[18:31]  * patdk-wk swears auditors are people that are unable to maintain things anymore
[18:32] <mdeslaur> hehe
[18:35] <MoooookiE> Hello, I have a problem with mysql. Someone tell me that the problem could come from apparmor. I have change this file: /etc/apparmor.d/usr.sbin.mysqld but i have no init.d script to restart apparmor. i'm not sure if i have apparmor
[18:35] <mdeslaur> MoooookiE: what version of Ubuntu are you running?
[18:35] <mdeslaur> MoooookiE: did you have apparmor denied messages in dmesg?
[18:36] <MoooookiE> DISTRIB_DESCRIPTION="Ubuntu 10.04.2 LTS"
[18:36] <MoooookiE> no I don't have denied messages
[18:37] <mdeslaur> MoooookiE: if you don't see denied messages in dmesg, your problem isn't coming from apparmor
[18:37] <MoooookiE> mdeslaur: ok, thanks :)
[18:37] <MoooookiE> mdeslaur: have you an idea what my mysql problem could be? i just geht this errors: http://privatepaste.com/4d276bd1a3 here is my config: http://privatepaste.com/ba2bdd68f6
[18:37] <mdeslaur> MoooookiE: see here for more info about debugging apparmor: http://wiki.apparmor.net/index.php/AppArmor_Failures
[18:39] <mdeslaur> MoooookiE: oh, hmm...you did change a bunch of default file locations
[18:39] <mdeslaur> MoooookiE: in that case, maybe you are being blocked by apparmor
[18:39] <MoooookiE> mdeslaur: yes but I don't have denied messages :)
[18:39] <mdeslaur> MoooookiE: are you sure you don't have denied messages in /var/log/kern.log?
[18:39] <mdeslaur> MoooookiE: does aa-status show mysql as being confined?
[18:39] <MoooookiE> mdeslaur: yes I'm sure
[18:40] <MoooookiE> mdeslaur: I also don't have aa* tools
[18:40] <mdeslaur> MoooookiE: do you have the "auditd" package installed?
[18:40] <DanaG> argh, can't use zfs and forked-daapd.
[18:40] <DanaG> libavl1/libzavl1 conflict.
[18:40] <mdeslaur> MoooookiE: you don't have the aa tools? oh, that's pretty odd
[18:41] <MoooookiE> mdeslaur: I don't need apparmor . Just thougt it could be default on ubuntu.
[18:41] <mdeslaur> MoooookiE: it definitely is by default, which is why it's odd you don't have it installed
[18:41] <patdk-wk> not if you do min-install
[18:41] <mdeslaur> well, if you don't have the tools, and you don't see denied messages, it may not be apparmor
[18:42] <mdeslaur> patdk-wk: seriously? I'll have to fix that...
[18:42] <MoooookiE> mdeslaur: do you have an idea what it could be instead of apparmor?
[18:42] <patdk-wk> pretty sure, never installed on any of my installs
[18:43] <mdeslaur> MoooookiE: sorry, no
[18:43] <MoooookiE> someone else here who can help me with ubuntu+mysql?
[18:44] <KM0201> is there something i can install on my server, so i can use a web browser to view details of the server (i read about SWAT, but it seems unsupported now).. i just wann asee details like disk use, processes, etc.
[18:45] <sbeattie> MoooookiE: the contents of /var/log/mysql/server1.err or some of the other /var/log/mysql* may be of interest.
[18:46] <hggdh> Daviey: there?
[18:47] <ewj> if 'apt-cache show openssl' is showing me two versions of the package (one of which is "fixed" and the other not), can I be confident that the fixed one is the one currently in use?
[18:47] <mdeslaur> ewj: dpkg -l | grep openssl will show you what's installed
[18:48] <MoooookiE> sbeattie: all empty
[18:48] <MoooookiE> sbeattie: and /var/log/mysql/server1.err is not created
[18:48] <ewj> mdeslaur: ah, excellent, thank you, that's what I was looking for!
[18:48] <MoooookiE> sbeattie: there are just this entrys in /var/log/syslog
[18:48] <sbeattie> ewj: 'apt-cache policy openssl' will also show you which version is installed
[18:49] <ewj> sbeattie: also good to know, thanks :)
[18:50] <renatocan_> Hello! I need a web interface to change smbpasswd passwords. Any suggestions ?
[18:53] <patdk-wk> use command line instead?
[19:00] <KM0201> patdk-lap: any idea on like a general WebUI to view performance oof the server>?
[19:00] <KM0201> woops
[19:00] <KM0201> patdk-wk: see above.
[19:00]  * patdk-wk just uses munin
[19:00] <genii-around> renatocan_: ebox has a samba module
[19:00] <KM0201> patdk-wk: munin?.. is tha tin the repos.
[19:01] <patdk-wk> yep
[19:01] <patdk-wk> or cacti, or ...
[19:01] <patdk-wk> I would say munin is the most painless to install, cause it mostly autoconfigures itself
[19:01] <KM0201> munin it is!..lol
[19:02] <renatocan_> thanks. I'll take a look at ebox.
[19:04] <philipballew> if I want to have my comp run rsync -azv  /home /myusername  /path/to/externalharddrive once a week how would i do that with a script and a cron job?
[19:04] <patdk-wk> why use a script
[19:04] <patdk-wk> just shove that in cron
[19:05] <JoeGazz84> Hello. Anyone know of any good panels I can use on ubuntu webserver 10.04? EHCP messes up nginx files and webmin is ugly and it doesnt work very well...
[19:05] <JoeGazz84> Any others?
[19:05] <JoeGazz84> free of course
[19:05] <patdk-wk> JoeGazz84, I always enjoyed vi
[19:06] <philipballew> i just make a text file with that command and set the file with root privileges? patdk-wk
[19:06] <patdk-wk> philipballew, no
[19:06] <patdk-wk> you use crontab -e, probably after you use man crontab
[19:06] <JoeGazz84> patdk-lap: vi? is that the full name?
[19:06] <philipballew> gonna need to look into this
[19:07] <patdk-wk> well, man 5 contrab
[19:07] <patdk-wk> JoeGazz84, it's a joke, none of them are really any good, cause if you make any changes from what it wants for a default, it gets screwed all up
[19:07] <patdk-wk> so it's best to just do it manually
[19:08] <JoeGazz84> Oh, but the thing is, there are a multitute of us who are clueless to all the shell commands... Are there any other panels, IDC if the y are necessarily /good/, we just need a decent one
[19:09] <patdk-wk> you don't need any shellcommands, it's just editing text files
[19:09] <JoeGazz84> wait wait, what?
[19:10] <patdk-wk> that is all the panels do, edit text file, restart program
[19:10] <Daviey> hggdh: hey.. i got your email.. good stuff!
[19:10] <hggdh> Daviey: thanks. I am now looking into the udeb integrations
[19:10] <Daviey> hggdh: do you want to raise a merge proposal against my branch?
[19:11] <hggdh> Daviey: will do, I was not aware we could request mergers to ./+junk
[19:11] <Daviey> hggdh: Yeah, i am kinda concerned that the d-i integration *might* need to be within the app.
[19:11] <guntbert> renatocan_:  be warned, ebox does its own thing, it doesn't work with the system's config files
[19:11] <Daviey> cjwatson suggested referecing against netcfg.
[19:11] <hggdh> Daviey: yes, I was wondering... libxmlrpc... are they available on install?
[19:12] <hggdh> Daviey: yes, I added it this way to ./debian/control (and depending on network-integration)
[19:12] <hggdh> cjwatson's work on Eucalyptus udeb is not going to waste ;-)
[19:13] <adam_g> zul: did glance-api.conf + glance-registry.conf get moved to a single glance.conf in the last week or so?
[19:14] <zul> dont think so
[19:14] <zul> why?
[19:15] <adam_g> installing from trunk for the first time since last week, and it looks like thats whats goin on. ugh
[19:16] <Daviey> hggdh: libxml can be, that isn't a concern
[19:16] <hggdh> Daviey: ack
[19:17] <Daviey> adam_g: AFAIK they are still seperate, there is a NEW one called glance-scrubber.conf which utlemming is working on atm
[19:17] <adam_g> hmm
[19:17] <Daviey> adam_g: you might be having pain if that is missing?
[19:20] <adam_g> not having any pain yet.. just wondering what the hell has happened in the last 4 days :)
[19:21] <adam_g> the ensemble formulas need to be updated accordingly.. next week we should discuss how this is going to maintained moving forward. is there an agenda up anywhere?
[19:23] <hggdh> Daviey: I really do not know how to propose a merge to ./+junk/cobbler-enable. Is it possible?
[19:29] <Daviey> hggdh: should be possible.. :/
[19:29] <Daviey> hggdh: push it up to your +junk/$something
[19:30] <hggdh> Daviey: ack
[19:35] <hggdh> Daviey: lp:~hggdh2/+junk/cobbler-enroll2
[19:41] <Daviey> hggdh: pah, you are quite right - you can't merge proposal +junk branches
[19:41] <Daviey> that sucks.
[19:42] <Daviey> hggdh: I'll merge locally
[19:42] <Daviey> thanks
[19:45] <Daviey> hggdh: linking against debian-installer, but doesn't use it?
[19:51] <hggdh_> Daviey: I forgot to add an include <debian-installer.h>
[19:52] <hggdh_> I am *guessing* d-i intercepts/wraps some common calls, like printf
[19:53] <hggdh> Daviey: I am going to build the packages now, and see what happens
[19:53] <Daviey> hggdh: nfi
[19:53] <hggdh> heh?
[19:54] <Daviey> hggdh: I don't know of a better of testing udebs.. the way i did it with euca was to netinstall, and have an addional archive with my udeb
[19:54] <Daviey> a real PITA to setup
[19:54] <Daviey> there must be a simplier way
[19:54] <hggdh> I agree, something simpler must exist. I hope...
[19:56] <Daviey> hggdh: You also have to fude the Contents.gz IIRC to include udebs.
[19:56] <Daviey> fudge*
[19:56] <Daviey> really not very awesome.
[19:57] <hggdh> indeed :-(
[19:59] <Daviey> hggdh: I'm tempted to have two cobbler-enrolls.. one for general consumption, and one for installer
[19:59] <Daviey> ideally, code sharing.
[19:59] <hggdh> Daviey: does not sounds like a bad idea
[20:00] <hggdh> Daviey: let me see if I can build the momster, at least
[20:00] <Daviey> hggdh: good cookies
[20:01] <hggdh> :-)
[20:07] <Daviey> RoAkSoAx: How much do you have on your plate atm?
[20:15] <hggdh> Daviey: OK, beginner's question, chicken&egg: how can I bzr bd if there is no .tar.gz, and quilt is complaining about that?
[20:16] <maxb> *quilt* is complaining about that?
[20:18] <Daviey> hggdh: bzr bd --split
[20:19] <Daviey> hggdh: there isn't yet an upstream tarball, so splitting by /debian and assuming everything above it is upstream is safe.
[20:19] <Daviey> hggdh: --split does that for you
[20:20] <hggdh> Daviey: ah, my fault, I also had --native there
[20:20] <hggdh> works like a champ (mind you *not* a chimp)
[20:23]  * zul thinks its time to roll a new cobbler tarball
[20:27] <Daviey> zul: best wait a few hours to dput, just incase.
[20:31] <RoAkSoAx> Daviey: right now im working on ensemble/orchestra, need to review/postpone some workitems related to cluster/powernap need to do a few testdrive fixes, and need to start with the arm/cobbler stuf
[20:31] <RoAkSoAx> Daviey: what do you hve in mind?
[20:33] <kirkland> RoAkSoAx sounds *very* busy :-)
[20:33] <RoAkSoAx> kirkland: it is :) :(
[20:34] <HotPornYo> Check it out: http://bit.ly/HotPornYo
[20:36] <kirkland> moderator???
[20:36] <RoAkSoAx> !ops
[20:36] <RoAkSoAx> heh i think we need new moderatos
[20:36] <HotPornYo> lol
[20:36] <HotPornYo> it's a joke guys
[20:37] <oCean> HotPornYo: not funny, not acceptable
[20:38] <HotPornYo> I'm tracking to see how many people actually click it as part of a field study.  It just takes you to a page that says "Thanks for your vote"
[20:38] <Daviey> RoAkSoAx: I wondered if you wanted to setup a fresh oneiric machine in the lab for usage next week?
[20:38] <Daviey> (as a cobbler server)
[20:38] <r4__> I have a working master LDAP server and a slave server...is there a way i can force an update on the slave?
[20:38] <HotPornYo> ...and how fast I get booted.
[20:39] <tkeith> How is the private cloud software for Ubuntu Server? Does it all work well? Is it production-ready?
[20:39] <RoAkSoAx> Daviey: I guess I could do that tomorrow... remind me please :)
[20:39] <Daviey> RoAkSoAx: sure thing! Great news
[20:48] <hggdh> Daviey: conditional compilation could do the trick, but I guess we have to move out of the simple ./rules you have
[20:49] <cemerick> Just getting started in ubuntu 11.04 — I'm attempting to set PGDATA in /etc/postgresql/8.4/main/environment (as recommended in /etc/init.d/postgresql), but that does not seem to impact the value provided to postmaster via -D.  Any thoughts?
[20:50] <Daviey> hggdh: Hmm
[20:50] <Daviey> hggdh: I wondered if the udeb version could not just make use of the functions in the primary one?
[20:50] <Daviey> link against it?
[20:51] <hggdh> IDK. I was basing on the eucalyptus work, but there the udeb is simply for, ah, udeb
[20:51]  * RoAkSoAx goes on the hunt for pandboard's accessories
[20:52] <hggdh> but it explicitly links against debian-installer, and has this #include <d-i.h>
[20:55] <Daviey> hggdh: Yeah, it might just be easier to duplicate the code for this cycle.  We can't spend too much time on it atm.
[20:56] <Daviey> working > ideal
[20:56] <hggdh> agree
[20:57] <hggdh> and (just looked) has a few di_* calls in there, doing some sort of magic that I guess only Colin can decipher
[20:59] <Daviey> Yeah, the docs are not good for drive by contributors
[20:59] <hggdh> aye
[21:00] <Daviey> hggdh: BTW, did you see the last debconf template question?
[21:00] <Daviey> the shutdown note?
[21:00]  * hggdh considers secluding self in a small shell on a frigid northen beach
[21:00] <hggdh> Daviey: let me get the details
[21:00] <Daviey> hggdh: It needs to halt following executing this code..
[21:01] <hggdh> Daviey: what, trhe install?
[21:01] <Daviey> hggdh: yah
[21:02] <Daviey> hggdh: We are abusing d-i to execute this code, purely to get it registered with cobbler.
[21:02] <Daviey> cobbler then decides how/if to install the node.
[21:02] <hggdh> this is fun
[21:02] <Daviey> hggdh: sarcaism isn't good :)
[21:03] <hggdh> ah, sorry
[21:03]  * hggdh runs in circles, screaming
[21:04] <hggdh> Daviey: OK. let's get back to beginning. We are considering two packages, one run-time, one d-i
[21:05] <hggdh> on the d-i, we need to have -- somewhat magically -- to have the templates & questions & results integrated with the call to cobbler-enroll, and *then* d-i must stop
[21:05] <Daviey> hggdh: yeah, the standalone is purely a want.. the udeb is a need.
[21:05] <Daviey> hggdh: yeah, calling halt.
[21:06] <Daviey> Could do with colins help here.. it might be possible to skip directly to halt in the installer
[21:06] <Daviey> i honestly don't know
[21:06] <hggdh> yes
[21:06] <Daviey> if not, we need to implement that
[21:06] <hggdh> there are at least two "halts"
[21:07] <hggdh> one -- d-i stops churning, but the machine is still alive (something else running?)
[21:07] <hggdh> the other is a real, honest-to-god, halt. I guess you mean the first one
[21:08] <Daviey> hggdh: no, i mean a power off, hard halt.
[21:08] <hggdh> oh, that should be easier
[21:08] <Daviey> hggdh: This work is purely for hardware discovery.
[21:08] <hggdh> ok
[21:08] <hggdh> huh
[21:09] <hggdh> so someone powers on a new machine, boots an ISO (or preseed), this code gets run, and the machines poffs
[21:10] <hggdh> or PXE-started?
[21:10] <Daviey> hggdh: well the beauty of it being on the iso, means it can be PXE started.
[21:11] <Daviey> The real use case is PX starting, but on the iso for smaller shops and for example
[21:11] <Daviey> The larger outfits would never fill in the fields we are adding, that would all come down the pipe from a preseed
[21:11] <hggdh> yes, makes sense
[21:13] <hggdh> OK. We are back to the udeb
[21:13] <zul> has anyone tried the ipmi stuff out in cobbler yet?
[21:13] <hggdh> will look for mor examples, but I would say the code right now is missing the integration completely
[21:15] <hggdh> ah, cool, starting to understand how it works
[21:16] <Daviey> zul: I don't have any 'spare' IPMI enabled servers.
[21:16] <Daviey> hggdh: do you know if the lab machines have IPMI?
[21:16] <Daviey> ah crap, they are HP so iLo.
[21:18] <kirkland> RoAkSoAx: ivoks: this makes me sad: http://lwn.net/Articles/452949/
[21:18] <kirkland> "(The free service for Fedora and Ubuntu appears to still be functioning, for now - but who builds a high-availability system on those distributions?). "
[21:21] <mdeslaur> kirkland: so much for impartial journalism :P
[21:24] <hggdh> Daviey: the server lab is HP, the QA is Dell
[21:26] <Daviey> kirkland: *sigh*
[21:26] <kirkland> mdeslaur: par for the course for LWN
[21:27] <Daviey> hggdh: Interesting, might try to get some testing in the QA lab then :)..  ISTR Dell's IPMI standard was different to the vendor i used to use, using the same standard :o
[21:28] <hggdh> heh
[21:29] <hggdh> Daviey: depending on how much/long you need it, we can arrange something
[21:30] <Daviey> hggdh: 2 machines, min 2 hours.. I'd quite like the various out of band stuff to be checked with as much as we can pre-release
[21:46] <lau> is there any issues with launchpad browsing code ? it returns 503...
[21:48] <StevenR> lau: link?
[21:55] <DanaG> Say, why would bonjour/avahi not work over ppp0 (ipsec+l2tp)?
[22:02] <lau> StevenR: http://bazaar.launchpad.net/~launchpad-pqm/launchpad/stable/files
[22:04] <_johnny> anyone know if there's a photoshop helpchan on freenode?
[22:05] <StevenR> lau: yeah, broken here too. File a support request maybe? https://answers.launchpad.net/launchpad
[22:06] <Jare> _johnny: dunno, but what's the issue?
[22:07] <Martyn> what's the expected behavior after installing postgresql on Oneric?
[22:07] <Martyn> Should it automatically start the daemon?
[22:07] <_johnny> i've created a vector (path/whatever you call it) cutout which should be a pattern. so what i want to do is use it to create a cirlce, as in, it should repeat it all the way around. i'm 99% sure i've seen it done before, but i'm not very good at google for my situation ;)
[22:08] <_johnny> (i'm not a professional)
[22:10] <RoAkSoAx> kirkland regarless.of.thst comment you would be surprised thst theres quite a.few ppl building ha in LTS
[22:11] <RoAkSoAx> kirkland but the upcoming LTS will be better as past.releases had very updated and fully working ha services/infrastructure
[22:13] <Martyn> okay, something +is+ wrong with the Oneric postgres package
[22:13] <Martyn> it's not starting after install and it didn't start after I rebooted
[22:14] <Martyn> it does run if I manually service posgresql start
[22:14] <_johnny> Jare: http://www.max-realms.com/modules/newbb/dl_attachment.php?attachid=1293562213&post_id=10775 (it's a download as .zip, containing a pdf) :) in case you ever need something similar
[22:18]  * RoAkSoAx cant believe he has spent 100bucks on cables for one pandaboars
[22:19] <Martyn> what the heck did you buy?
[22:23] <adam_g> zul: ping
[22:26] <RoAkSoAx> Martyn: mini usb to usb, hdmi to dvi, and usb to serial
[22:26] <Martyn> Why didn't you buy them from monoprice.com?
[22:26] <RoAkSoAx> Martyn: can't seriously believe cables there are so damn expensive
[22:27] <RoAkSoAx> Martyn: cause wanted to work on it this week but giving  a second though I thingk I'd just return them and buy online
[22:27] <Martyn> I have 6 pandaboards here, pretty much the same setup, and -all- the cables combined didn't come up to $100.
[22:27] <Martyn> Yeah, that would probably be a good idea.
[22:27] <Martyn> That's just a waste of money
[22:27] <kirkland> RoAkSoAx: word
[22:27] <kirkland> RoAkSoAx: throw up a comment to my blog post to that affect ;-)
[22:27] <frumin> hello
[22:28] <RoAkSoAx> Martyn: indeed!!
[22:28] <RoAkSoAx> kirkland: will do... looking at it now
[22:28] <frumin> does anyone know the best way to upgrade from ubuntu server 6 to the latest version?
[22:30] <RoAkSoAx> Martyn: do you recommend using the recommended digikey power supply or just the mini usb to usb?
[22:31] <Martyn> use the right power supply
[22:31] <Martyn> USB tends to brown out
[22:31] <Martyn> it just can't supply the power the panda really wants
[22:31] <Martyn> I actually have mine hooked to a more robust bench power supply now...
[22:31] <RoAkSoAx> Martyn: ok cool, thanks for the advice
[22:48] <RoAkSoAx> Martyn: 4gb of storage is enought right?
[22:48] <RoAkSoAx> 4gb sd cards
[22:51] <RoAkSoAx> Martyn: or will micro sd with adapter work too?
[23:33] <Martyn> more than enough