[00:06] <Daviey> hallyn, rolling own... hmmpf.. can they provide a diff?
[00:07] <Daviey> bye bye hallyn
[00:07] <hallyn> Daviey: dunno
[00:07] <Daviey> ah, just caught you
[00:07] <Daviey> something for tomorrow :)
[00:07] <hallyn> Daviey: i was hoping to tackle the firstboot ones tomorrow, but i'm gonna need bootstrap help from smoser or someone
[00:08] <hallyn> well, maybe i should just try to actually solve the lucid and maverick open-vm-tools bugs now
[00:08] <hallyn> in an SRU-friendly way
[00:08] <hallyn> for natty there would have been no way (so good thing we got it in), but hopefully lucid's kernel hasn't changed too much
[00:10] <hallyn> Daviey: all right, i'm shutting down the client, good night :)
[00:13] <Daviey> o/
[01:31] <mr_orange> Hey, I am installing ubuntu server 10.10 and I am wanting it to  be a mail server as well what option should I choose for the mail configuration?
[01:35] <mr_orange> Hey, I am installing ubuntu server 10.10 and I am wanting it to  be a mail server as well what option should I choose for the mail configuration? can anyone help me
[01:53] <Tukanfan> if you have a backup system where a client backs up to a central server, what do you then do if the client gets compromised?
[01:56] <twb> Tukanfan: push or pull?
[01:56] <Tukanfan> push, because clients is often behind NAT
[01:56] <twb> Using what, rsync over ssh?
[01:56] <Tukanfan> yep
[01:56] <twb> Learn about rrsync
[01:56] <Tukanfan> will google that
[01:57] <twb> It's in /usr/share/doc/rsync/scripts/rrsync.gz
[01:57] <Tukanfan> okay
[01:57] <twb> For example, I allow managers to upload to /var/www as root with command="/usr/local/bin/rrsync /var/www/",no-agent-forwarding,no-port-forwarding ssh-rsa [their hash] [their email address]
[02:03] <Tukanfan> so you execute that command on the server or client side?
[02:07] <Tukanfan> nvm, seems like i found out looking in the script source
[02:08] <Tukanfan> but would it be possible to use it with duplicity too?
[02:27] <twb> What does adding a google-site-verification TXT record to my domain actually *do*?
[02:27] <twb> $manager is asking me
[02:27] <mr_orange> im trying to install a gui for my server but it tells me that it is unable to locate the package. Can someone tell me how I can fix this?
[02:29] <patdk-nb> twb, I think google just uses that to verify you are the owner
[02:29] <patdk-nb> cause only the owner can change dns
[02:29] <twb> So they're reinvented DNSSEC?
[02:30] <twb> Also, presumably that hash is *not* a shared secret?
[02:31] <patdk-nb> http://googlewebmastercentral.blogspot.com/2010/03/dns-verification-ftw.html
[02:31] <patdk-nb> I doubt it
[02:33] <twb> Presumably this is for that "analytics" stuff?
[02:33] <twb> Where you put js into your HTML to make it spy on GUI browser users
[02:50] <twb> Oh, apparently this is to allow us to remove content from google's cache
[02:50] <twb> FSVO content = "whoops, that's accessible externally?"
[02:55] <twb> OK, I get it now.  I talked to $coworker about it
[03:22] <iceman34> anyone know of a a good irc widget
[03:28] <twb> emacs
[03:28] <nealmcb> :)
[03:41] <patdk-nb> hmm, launchpad seems to be back to it's normal forever slowness again :(
[06:25] <twb> So the kernel is OOM-killing apache2 and mysqld.
[06:25] <twb> But upstart restarts them... I'm not sure that is the Right Thing in these circumstances.
[07:16] <Roxyhart0> hi there, sombody know why dhcp3-server doesn't start at boot time? is there any bug?
[07:28] <cmdbbq> hello, i am setting up a server from the commandline (there will be no GUI) over ssh and I have some questions about adding a user. I ran adduser myusername and set a password with passwd then logged in as that user, but it complained of no home directory. so i logged back in as root and created a home directory /home/myuser and set myuser as the owner with chown. now when I log in i am presented with just a $, preceded by no text a
[07:30] <Roxyhart0> ther is someconfiguration that you indicate which kind of shell or c shell the user is going to use. I think it is correct that you are seen
[07:33] <twb> did you use "adduser" or "useradd"?
[07:34] <cmdbbq> ah yes, my mistake, useradd
[07:34] <twb> Do not use useradd, for this reason
[07:34] <twb> useradd is a low-level command; adduser is a high-level command that creates /home and everything else it should do.
[07:35] <cmdbbq> ok, should i delete the user/directory and use adduser to recreate or will this create conflicts?
[07:35] <twb> Recommend you deluser fred, delete the home dir, and then adduser fred.
[07:35] <twb> If it was me I might clean it up by hand, but it's not worth talking you through it
[07:35] <cmdbbq> ok cool thank you very much :)
[07:36] <twb> The reason your shell was $ was because 1) you didn't get the dotfiles from /etc/skel, and 2) your shell was probably /bin/sh not /bin/bash.  Using adduser will fix both.
[07:37] <cmdbbq> twb: everything works as expected now, I appreciate the education
[07:38] <twb> No problem
[07:47] <Roxyhart0> i just installed dhcp-server, however when i boot it doesn't start automatically. Someone know how to fix it to start on boot?
[08:25] <Roxyhart08> hi there, i installed dhcp-server but it can't start on boot time, somebody know how to do it? if there any bug why it is not starting automatically
[09:19] <jamespage> Daviey: morning
[09:20] <jamespage> Daviey: I have a fix from upstream for bug 759943
[09:53] <Daviey> jamespage: yeah, i saw the comment on the bug report.. Have you integrated it, and some mild testing?
[09:53] <jamespage> Daviey: yes - it passed my simple 'Hello World' wsgi test case for both python 3.1 and 3.2 versions
[09:54] <jamespage> Daviey: want me to propose it?
[09:54] <jamespage> Daviey: and feel free to give brutal feedback on my conditional patch application in debian/rules :-)
[09:56] <Daviey> heh
[09:57] <Daviey> jamespage, whilst i look at that, have you seen bug 745946 in any of the recent jenkins tests?
[09:59] <jamespage> So the last round of testing did not throw up any tests with the same symptoms - that was for beta-2
[09:59] <jamespage> Daviey: those two instances where seen during beta-1 I think
[10:01] <Daviey> jamespage, If smoser agrees, might mark it fix released later.. It could have been infrastructure breakage rather than platform.
[10:02] <Daviey> jamespage, What makes you think the patch might be incompatible with mod_python?
[10:03] <jamespage> Daviey: thats what Graham from upstream indicated; description on patch is based on his comments
[10:03] <jamespage> something todo with threading and mod_python not following the rules :-)
[10:05] <Daviey> ah
[10:19] <jamespage> Daviey: Does it look OK to you?
[10:20] <Daviey> jamespage, well - i have a feeling that you can't call it source format 3.0 (quilt) with that.
[10:20] <Daviey> just trying to check.
[10:20] <jamespage> Daviey: ack - thanks for your help :-)
[10:21] <Daviey> jamespage, it seems functional!  And nicely done... just trying to check somthing
[10:21] <Daviey> jamespage, I've not had to have condtional patches before since the new world order of 3.0 (quilt).
[10:22] <Daviey> the problem is, dpkg-source -x, doesn't give you what will necessarily build.  It's a convention that with 3.0 (quilt) that is the case.
[10:22] <Daviey> jamespage, Half balancing 'getting it done' and the other half wondering if the patch should contain the condition.
[10:25] <Daviey> jamespage, hmm, has the patch author already done that?
[10:26] <Daviey> jamespage, usage of #if PY_MAJOR_VERSION == 3 && PY_MINOR_VERSION >= 2 ?
[10:26] <Daviey> ie, if python = 3.2:
[10:26] <jamespage> Daviey: well there is some conditional code for 3.2
[10:27] <jamespage> Daviey: I would be more concerned where there is no condition for 3.2 on code - I think that is where the problems might lie.
[10:29] <Daviey> jamespage, yeah... i just wondered if that was /safe/ code that didn't need conditioning
[10:29] <Daviey> jamespage, http://pb.daviey.com/Im3W/ - probably easier to view!
[10:30] <Daviey> lines starting, 18? DaveWal - is the new patch
[10:31] <Daviey> jamespage, I think you are correct that it doesn't filter.
[10:32] <Rickardo1> Is there any way I can restore the routing tables on my ubuntu server?.. https://gist.github.com/1d52cb79114fcb782a79   I can ping my lan but not internet
[10:32] <jamespage> Daviey: no completely :-(
[10:32] <jamespage> /no/not/
[10:33] <jamespage> Daviey: I was more  concerned that the upstream developer was saying 'I don't think this will work with python 2.7'
[10:33] <jamespage> hence the conditional application for Python 3.2 only
[10:48] <pethkaqeni> may someone help me to install an attansic driver!
[10:48] <pethkaqeni> i have a ubuntu server 9.04 installed
[10:50] <pethkaqeni> may someone help me to install an attansic driver!
[10:50] <pethkaqeni> i have a ubuntu server 9.04 installed
[10:51] <pethkaqeni> HELP plz
[10:53] <pethkaqeni> camooon no one here knows !!!
[10:55] <raphink> I have no idea what attansic is
[10:56] <pethkaqeni> thanks for reply
[10:56] <pethkaqeni> is a network card
[10:56] <pethkaqeni> and the driver nedet is attansic
[11:07] <pethkaqeni> raphink: any idea ???
[13:02] <tyreza> hello
[13:06] <tyreza> is there any one there ?
[13:07] <tyreza> how to check if the server hold the load well or not ?
[13:23] <zul> morning
[13:24] <tyreza> hi
[13:24] <tyreza> how to check if the server hold the load well or not ?
[13:59] <jMCg> tyreza: what do you mean?
[13:59] <tyreza> i mean what i mean
[13:59] <tyreza> just only need to check the server capacity
[14:00] <tyreza> how to check my server is good or not ?
[14:00] <tyreza> perfomance indicater
[14:00] <jMCg> Good for what?
[14:09] <tyreza> need to check my system perfomence
[14:10] <tyreza> i don't know how to explain you in other way
[14:12] <Dr_Jekyll> tyreza: perhaps you mean "top"
[14:13] <Dr_Jekyll> it shows the current system load and the running processes
[14:13] <RoAkSoAx> morning all
[14:16] <tyreza> no not really
[14:16] <tyreza> i host a apache server
[14:18] <tyreza> what i have to do, to understand my actual server configuration is good or need to increase something ?
[14:18] <patdk-nb> tyreza, there are so many variables for you to check
[14:18] <patdk-nb> cpu performance, network performance, disk performance, responce time, ....
[14:20] <tyreza> correct
[14:21] <tyreza> and i can't run a command every few sec
[14:21] <tyreza> to check my all those performance
[14:22] <tyreza> my question is simply is there any tool able to check every ' few variable ' and to notify
[14:22] <tyreza> me
[14:22] <tyreza> when there is an over usage
[14:23] <semiosis> tyreza: nagios
[14:24] <tyreza> correct answer
[14:24] <tyreza> will check that one
[14:24] <tyreza> thanks
[14:36] <b0gatyr> total unrelated but August 4, 1997 marks the day skynet goes operational or is it today?
[14:39] <smoser> jamespage, i think that we should dupe bug 760725 to bug 745946
[14:39] <smoser> Daviey, you need to get over your issues with having open bugs.
[14:41] <jamespage> smoser: they do look very similar
[14:47] <smoser> did you change the test suite to grab 'initctl list' ?
[14:47] <smoser> and 'ps -a' output?
[14:48] <smoser> one other feature request for that... log the timestamp of when you do the collection some where. as it is right now we can't really tell if you just grabbed something in the middle of boot, or if it was 15 minutes later (where the system should have certainly booted).
[14:59] <jamespage> smoser - ack I'll add the timestamp to the list as well
[15:01] <Daviey> smoser: uh?
[15:02] <smoser> Daviey, you seem to just want bugs to be closed.
[15:03] <smoser> i would much rather have the bug open, so that when i find bugs, i can search for other occurences of the same issue
[15:03] <smoser> and say "oh look, this happened before"
[15:07] <Daviey> smoser: Call me old school... but i guess i like a bug to reflect it's current status.
[15:07] <smoser> which is open
[15:07] <Daviey> Ie, if a Fix has been Released, i like to call it "Fix Released"
[15:08] <smoser> if a bug occurs twice in 100 cases, and then doesn't happen to occur in the next 100, i dont think that calling it "fixed" is appropriate
[15:08] <smoser> unless you know you did something that should have had that affect
[15:08] <Daviey> smoser: It does sound like you need a better mail client :)
[15:09] <Daviey> smoser: Hmm.. i see your point... but often bugs are fixed accidently, through someone touching something else.
[15:33] <garymc> Hi guys anyone know anything about ubuntu-ltsp?
[15:42] <raphink> garymc, ask your question, you'll see if someone can answer
[15:43] <garymc> ok m y LTSP thin clients keep losing connection to the website on the LTSP server. The website is PHP based and it is for taking people details over the phone.
[15:43] <garymc> So firefox stops working on all thin clients
[15:44] <garymc> but If I log in from a remote machine like my Macbook its stilll working fine, which makes me think something on the server which lets the thin clients pxe boot is not working properly
[15:45] <garymc> 5 mins later all thin clients are using firefox fine again
[15:45] <garymc> its becoming a hassle when someone takes a call and it has stopped working
[15:46] <garymc> so any help would be appreciated
[15:48] <patdk-nb> garymc, dhcp renewal issue?
[15:49] <garymc> i dont know
[15:49] <garymc> how would i tell?
[15:49] <patdk-nb> well, I dunno either :)
[15:49] <patdk-nb> logs
[15:49] <garymc> im a total newb and i set this up over a year ago
[15:55] <Daviey> .
[16:07] <therobot> Hello, I just want to install one package (redis) from a ppa repository, I have this (https://gist.github.com/02b0b01b8215ff5b2e27) apt preferences file, but apt keeps chosing candidates from this ppa for the other packages, what I am doing wrong?
[16:09] <patdk-nb> heh, generally it's not that easy to select a single package from a repository
[16:09] <patdk-nb> you could just download the .deb you want, and install it using dpkg -i
[16:09] <therobot> then I miss updates
[16:09] <patdk-nb> yep
[16:58] <Slyboots> Mm..
[16:58] <Slyboots> Starting to wonder if I have too much crap installed on my server
[17:22] <TheEvilPhoenix> Slyboots:  why would you wonder about that, out of curiosity?
[17:22] <Slyboots> Seem to be running out of disk space primarlly
[17:24] <Slyboots> Plus its just generaly slow to boot
[17:24] <mynameistux> I'm currently using debian server, but I want to use ubuntu server. Is it possibly for me to install the new server OS without loosing my LVM
[17:25] <patdk-nb> installing it to where?
[17:25] <mynameistux> I have the / partition on sda1, and the LVM uses sda2 and sdb
[17:25] <patdk-nb> if you just install ubuntu ontop of sda1, then no
[17:25] <mynameistux> so I just want to install a new server OS on sda1 (which is not part of the LVM) and then recover the old LVM
[17:25] <patdk-nb> lvm is selfcontained
[17:25] <mynameistux> I don't have enough HDD's to backup my data
[17:25] <patdk-nb> all you have to do is activate the lvm and mount it
[17:26] <mynameistux> ok
[17:26] <mynameistux> can I just say, the support I got in the last couple of seconds was tens of thoasands of times better than all the support I got through the debian channel
[17:26] <mynameistux> your a legend patdk-nb =)
[17:27] <mynameistux> *you're
[17:27] <mynameistux> grammar is good
[17:35] <Wolfsherz> kennt sich jemand mit den bildschirmen von eizo aus?
[17:35] <Wolfsherz> sorry, wrong channel
[18:09] <TheEvilPhoenix> how can i get the system to stop yelling at me about a non-public GPG key in a PPA package?  because when I run 'aptitude update' it triggers an alert about NO_PUBKEY
[18:11] <fosterdv> TheEvilPhoenix: I don't know how to fix that personally, but I found this, http://ubuntuforums.org/showthread.php?p=1653773
[18:17] <TheEvilPhoenix> yep that fixed it
[18:17] <TheEvilPhoenix> fosterdv:  thanks
[18:17] <fosterdv> Anytime :D
[18:25] <TheEvilPhoenix> fosterdv:  i should probably write a bash script for that kind of thing... but i suck at scripting, so meh
[18:25] <TheEvilPhoenix> s/should/could/
[18:26] <fosterdv> TheEvilPhoenix: Lol, you and me both
[18:26] <fosterdv> I've been trying to figure out bash scripting for a while, and then just moved to a new project.
[18:26] <TheEvilPhoenix> fosterdv:  if only i knew how to get input from the command line... for example, ./getppgkey.sh <key here>
[18:28] <alaing> hi does anyone know if/how I go about configing my ubuntu 10.04 server edtion so I can use classic ASP
[18:28] <alaing> I'm currently running an apache web server
[18:29] <TheEvilPhoenix> there's apache plugins for various scripting languages
[18:29] <TheEvilPhoenix> do you mean the ASP .net scripting language?
[18:30] <alaing> no not .net just classic ASP
[18:30] <alaing> I know its really old but its for a university course I'm doing
[18:30] <alaing> and thought it would be good to setup my server and do a bit of practicing before it starts
[18:31] <TheEvilPhoenix> hmm i dont see a plugin/module/etc. for the non-.net version of ASP :/
[18:32] <remix_tj> asp can be executed only on iis
[18:32] <TheEvilPhoenix> ^
[18:32] <alaing> I dont really know to much about asp/asp.net all I know is classic ASP is really old.
[18:32] <fosterdv> http://www.mono-project.com/ASP.NET
[18:32] <fosterdv> ASP Hosting with Apache:
[18:32] <fosterdv> http://www.mono-project.com/ASP.NET
[18:32] <remix_tj> and as far as i know microsoft hope that as far as possible asp to be died
[18:34] <alaing> so my best bet would be to use IIS.....grr microsoft
[18:36] <fosterdv> alaing: Will this help? http://www.howtogeek.com/howto/ubuntu/run-aspnet-applications-on-ubuntu-for-developers/
[18:36] <alaing> fosterdv: no I dont think it will because its .net
[18:37] <alaing> though I may just setup .net anyway so thanks for the links
[18:37] <fosterdv> Yeah, sorry..
[18:39] <fosterdv> Anywhere I looked, that said anything about .net/asp, all pointed back to that first link I posted. Good luck.
[18:40] <fosterdv> http://redditech.wordpress.com/2009/03/01/quick-and-dirty-aspnet-on-linux-ubuntu-mono-and-monodevelop/
[18:42] <alaing> thank you.....just did a google and there seems to be quiet a few ASP web hosting options so perhaps I go with that in the mean time.
[18:43] <fosterdv> Cool, best of luck
[18:44] <alaing> thank you i just wish the university course covered asp.net rather than ASP but its also based on the server-side  of application development which the principles shoudl be the same
[18:45] <fosterdv> That's the truth
[18:56] <ivoks> SpamapS: ping
[18:56] <ivoks> SpamapS: https://bugs.launchpad.net/bugs/661453
[18:57] <TeTeT> I added a bridged device to a kvm guest, now it fails to start: http://pastebin.ubuntu.com/596157/
[18:59] <SpamapS> ivoks: sup?
[19:00] <TeTeT> been solved :)
[19:07] <ivoks> SpamapS: sorry, phone :/
[19:07] <ivoks> SpamapS: problem is that installing dovecot-imapd|pop3d overwrites dovecot.conf
[19:07] <ivoks> SpamapS: ucf doesn't solve that
[19:08] <ivoks> SpamapS: and proposed patch would introduce new problems for mail-stack-delivery package
[19:10] <SpamapS> ivoks: oh dovecot-imapd is the one that screws it up? hrm
[19:10] <ivoks> yes, with their .postinst
[19:10] <ivoks> proposed patch is a good way to solve this
[19:10] <SpamapS> ivoks: the description is really poor (I can say that because I wrote it myself. ;)  ... if you want to make it more clear what the problem is we can possibly address it better.
[19:11] <ivoks> just make sure to rename 01-mail-stack-delivery.conf to something with bigger number than the one that would define protocols
[19:11] <SpamapS> I'm not fixing it at this point
[19:11] <SpamapS> the problem is poorly defined in the description.. I wrote that a while back and forgot the details
[19:12] <ivoks> well, it's not:
[19:12] <ivoks> I suspect the maintainer scripts are modifying the file. If this is true, then doevcot.conf should be removed from the package and maintained by the maintainer scripts directly.
[19:12] <SpamapS> if you want to change the description to have a TEST CASE: I will re-open it gladly.
[19:13] <ivoks> sigh... this patch actualy overcomplicates things
[19:15] <ivoks> but, anyway, since nothing will be changed now, i guess this can be fixed in natty+1
[19:17] <SpamapS> ivoks: just please either open a new report or change that description and re-open it.
[19:17] <SpamapS> ivoks: otherwise it probably won't be fixed because the problem is not totally obviously a bug to most users.
[19:18] <ivoks> hm.. wow
[19:18] <TheEvilPhoenix> fosterdv:  i wrote a bash script that adds gpg keys :P
[19:18] <ivoks> i can't reopen it
[19:19] <ivoks> my ubuntu privileges are degrading :)
[19:20] <cloakable> heh
[19:48] <sjm> I've got a problem with my network.  Anyone know why I could reach anywhere locally on the network, but not through the firewall except through a proxy?  There are no outbound rules on the gateway (and this was working only a few days ago) (Ubuntu 10.04)
[19:49] <RoyK> sjm: wrong default gateway?
[19:49] <RoyK> pastebin netstat -rn
[19:49] <sjm> I've checked that with route -n, also /etc/resolv.conf is the same as a working computer
[19:51] <RoyK> can you ping the gateway?
[19:51] <sjm> http://pastebin.com/b7w4uXpb
[19:51] <RoyK> what does a traceroute tell you?
[19:52] <sjm> yes, pinging the gateway or other things internally to the network is fine.
[19:52] <sjm> traceroute (and pings) stop at the gateway for this computer.
[19:52] <RoyK> does the gateway NAT the connections?
[19:52] <sjm> I can ping externally on a laptop next to this box.
[19:53] <sjm> gateway Is set to nat.
[19:53] <RoyK> dunno then, sorry
[19:53] <sjm> I'm stumped too.
[20:19] <hallyn> jbernard: I suppose that (1) create_default_cgroups.sh needs to be bash (not sh), and (2) its post-stop script should umount /sys/fs/cgroup?
[20:21] <hallyn> jbernard: and finally, (3) cgconfig.conf has to have a mount entry or the gconfigparser bails
[20:24] <hallyn> (pushing a branch based on yours)
[20:25] <jbernard> i agree with 1 and 2
[20:25] <jbernard> is gconfigparser bailing now?
[20:26] <hallyn> it was for me, with empty cgconfig.conf
[20:26] <hallyn> so i just uncommented the last block (actually then added two more), and it worked
[20:26] <hallyn> lp:~serge-hallyn/ubuntu/natty/libcgroup/upstart-jb
[20:27] <jbernard> ahh, that parser make me hate my life
[20:27] <hallyn> I'm not attached to any of that, don't care how it's actually done
[20:27] <hallyn> yeah
[20:27] <hallyn> i know
[20:27] <hallyn> i KNOW
[20:28] <hallyn> the whoel concept of a cgroup-bin doesn't really work for me, because you arent free to change (without reboot) how some cgroups are bound!
[20:28] <hallyn> (or has that been improved upon in the last year?)
[20:29] <jbernard> i belive you are correct, that's what I'm seeing as well
[20:29] <jbernard> i say if it works to your satisfaction, then lets try to get it into natty. It's certainly loads better than the current version
[20:29] <hallyn> so with those changes the package is good for me
[20:29] <hallyn> jbernard: yup, works for me, let's do it
[20:30] <hallyn> that is, the package works for me, let's try to get it into natty :)
[20:30] <jbernard> awesome
[20:30] <hallyn> maybe we should ask for more guinea pigs:
[20:30] <hallyn> hi all, does anyone here use libcgroup?
[20:34] <hallyn> jbernard: is there an open bug for the security issue?
[20:35] <jbernard> bts or lp?
[20:35] <hallyn> jbernard: or, did we decide you can just dput bc you're in universe?
[20:35] <hallyn> lp
[20:35] <hallyn> to tie the bzr tree to
[20:35] <jbernard> i have not seen an lp one
[20:35] <hallyn> well, let me know if you need anything from me
[20:36] <rizzuh_laptop> I have a Lucid VM that is configured only with root, sudo is installed but apparently there's go group to it. How do I configure it and allow a user to use sudo, so I can disable root?
[20:36] <jbernard> hallyn: i will need someone to upload, i don't have that ability
[20:37] <hallyn> jbernard: oh, hey, i may actually have those
[20:38] <jbernard> ;)
[20:38] <hallyn> jbernard: though it wouldn't hurt to have someone else look at it (who wasn't part of writing it)
[20:38] <hallyn> in fact,
[20:38] <hallyn> maybe SpamapS would be a good one -
[20:38] <hallyn> given the upstartification
[20:38] <jbernard> i do agree
[20:39] <hallyn> SpamapS: ^ would you mind?
[20:39] <hallyn> SpamapS: i'll do the upload if you prefer, but your feedback would be appreciated nevertheless
[20:39] <hallyn> bbl
[20:44] <ChmEarl> rizzuh_laptop,  http://paste.ubuntu.com/596226/
[20:44] <rizzuh_laptop> ChmEarl, thanks
[20:45] <ChmEarl> rizzuh_laptop, np
[20:45] <rizzuh_laptop> ChmEarl, run that as root?
[20:45] <ChmEarl> rizzuh_laptop,  thats from a script to setup a lucid xen VM - yeah root
[20:50] <hggdh> TeTeT: ping
[20:50] <TeTeT> hggdh: hi, how can I help?
[20:51] <hggdh> TeTeT: this is about the bug you folks opened -- 765275 -- the interface capture shows errors and RSTs flowing
[20:51] <hggdh> TeTeT: errors == ICMP dest unreachable
[20:52] <TeTeT> hggdh: very confusing - how can there be a dest unreachable, the redirect on the clc worked just fine. We sort of thought another cloud install would interfere with ours
[20:52] <TeTeT> hggdh: as the labs equipment has been used for different clouds already
[20:53] <TeTeT> hggdh: to be honest we gave up on this for now and just went with plain kvm - but if you have a good idea I can give it a test
[20:54] <hggdh> TeTeT: I am not sure yet. But -- for example -- there is a TCP/IP open against 169.254.169.254:80, and this is immediately replied with a RST
[20:54] <hggdh> TeTeT: packet 1367-1368 in the capture
[20:54] <TeTeT> hggdh: yeah, it looked to us like some other host would reply?
[20:55] <hggdh> it might, even (if there is another one with the same IP, which would be bad in a different way)
[20:55] <rriggin> anyone know why my domain name is automatically getting added at the end of anything I request? For example nslookup www.google.com actually tries www.google.com.domain-name.com
[20:55] <TeTeT> hggdh: wouldn't the instance be having the front-end as router and thus reaching the right metadata service
[20:55] <hggdh> but the point is something/someone sent a RST on the line
[20:56] <hggdh> yes
[20:56] <TeTeT> rriggin: take a look at /etc/resolv.conf
[20:56] <TeTeT> hggdh: we used 'sudo tcpdump port 80' and nothing wsa seen from the front-end
[20:57] <TeTeT> hggdh: same as with 8773
[20:57] <hggdh> joy, joy, there we go again... now on tornado watch, probably soon to tornado warning :-(
[20:57] <TeTeT> hggdh: good luck!
[20:57] <rriggin> resolv.conf shows an entry for "nameserver 127.0.0.1" and and enty below it for "search domain.com"
[21:00] <hggdh> TeTeT: so it sounds like another machine is answering
[21:02] <TeTeT> hggdh: yep, though we couldn't figure out where it came from :( and I find it weird that a remote host answers when I thought all the traffic would go through the CC by default
[21:06] <rriggin> dhclient.conf does have a supersede domain-name "domain.com" in it. could that cause the issue?
[21:09] <SpamapS> hallyn: was at lunch, whats up?
[21:09] <rriggin> TeTeT:resolv.conf shows an entry for "nameserver 127.0.0.1" and and enty below it for "search domain.com
[21:09] <rriggin> TeTeT:dhclient.conf does have a supersede domain-name "domain.com" in it. could that cause the issue?
[21:09] <TeTeT> rriggin: weird, if it appends domain-name.com
[21:12] <rriggin> TeTeT: really throwing me for a loop because I have the config files setup the same way they are on the current server that is working properly.
[21:13] <TeTeT> rriggin: as it points to your localhost for DNS, what server is running there? Probably misconfigured bind?
[21:13] <rriggin> TeTeT: dnsmasq
[21:13] <TeTeT> rriggin: what's the output of dig +short www.google.com @127.0.0.1 from the server
[21:16] <rriggin> TeTeT: all dig requests come back with only 127.0.0.1
[21:16] <TeTeT> rriggin: guess you need to check your /etc/dnsmasq.conf then and see what's happening. Unfortunately I know next to nothing of its proper configuration
[21:17] <rriggin> TeTeT: thanks though.
[21:26] <SpamapS> jbernard: hallyn seems to think you have something for me to review/sponsor ...
[21:27] <ChrisBuchholz> Hey guys. Is there a preferred way to install django on my ubuntu server=
[21:27] <ChrisBuchholz> ?*
[21:32] <SpamapS> ChrisBuchholz: you'll get security updates and such with 'sudo apt-get install python-django' .. you will still need to configure your webserver to access it tho.
[21:33] <|TurBo|> hey guys
[21:34] <koolhead17> hi
[21:34] <|TurBo|> im wondering what web interface would u use on ubuntu server 10.10 64bit for web administration
[21:34] <|TurBo|> ?
[21:34] <|TurBo|> i see their is some choices and would want the best for my box
[21:35] <|TurBo|> ?
[21:35] <|TurBo|> would love some input on this
[21:36] <|TurBo|> since it seems that webmin isent the best choice?
[21:36] <guntbert> !webmin
[21:36] <|TurBo|> yeah that's what i was thinking to
[21:36] <guntbert> |TurBo|: to my knowledge there is none
[21:36] <patdk-nb> !ebox
[21:36] <koolhead17> security relate
[21:36] <|TurBo|> hmm
[21:36] <|TurBo|> thanks alot
[21:37] <|TurBo|> ill checkit out
[21:37] <guntbert> ebox is horrible
[21:37] <patdk-nb> it's the *other* option :)
[21:37] <|TurBo|> lol
[21:37] <koolhead17> guntbert: what is the point of using such apps
[21:37] <guntbert> |TurBo|: ebox is in no way a replacement for webmin
[21:37] <|TurBo|> hmm
[21:37] <patdk-nb> personally I think all are horrible
[21:37] <|TurBo|> is freepbx any good ?
[21:37] <koolhead17> patdk-nb: true
[21:38] <SpamapS> I was never ecstatic about webmin..
[21:38] <guntbert> koolhead17: for the parttime admin they provide an easier way around the system (even a full time admin might not know every corner...)
[21:38] <SpamapS> but it was good enough I could put it on a box that I couldn't support remotely and have the night operator restart services over the phone..
[21:38] <|TurBo|> but the issue is that i cba to run gui on server as it is a waiste of cycles, but i want to be abel to admin it from another client without having to use to much of its reso
[21:39] <|TurBo|> is it possibel to do this in a good way ?
[21:39] <SpamapS> its a lot easier to tell somebody "click DNS, then 'restart'" than "ssh.. no EHS-EHS-AYCH box12"
[21:39] <koolhead17> SpamapS: true
[21:39] <TeTeT> hggdh: found another prob in UEC 11.04: http://pastebin.ubuntu.com/596257/
[21:40] <koolhead17> hey TeTeT
[21:40] <TeTeT> hi koolhead17
[21:40]  * koolhead17 is happy with the natty beta2
[21:40]  * patdk-nb wonders why you would need to restart dns :)
[21:40] <|TurBo|> lol
[21:41] <koolhead17> patdk-nb: haha
[21:43] <koolhead17> i would never recommend any one to use web interface for administration though :)
[21:44] <jbernard> SpamapS: hallyn has pushed a libcgroup branch: lp:~serge-hallyn/ubuntu/natty/libcgroup/upstart-jb
[21:44] <jbernard> SpamapS: you're feedback would be most appreciated
[21:46] <semiosis> just encountered a problem with the recently updated language-selector-common package, specifically in ec2 but it may affect other platforms as well
[21:47] <semiosis> a fresh maverick install, updated & upgraded to the latest packages, leaves dpkg in a bad state, the error is...
[21:47] <semiosis> Errors were encountered while processing: language-selector-common E: Sub-process /usr/bin/dpkg returned an error code (1)
[21:47] <SpamapS> jbernard: reading.
[21:47] <guntbert> semiosis: please !pastebin the complete output
[21:48] <SpamapS> jbernard: upstartifying it 9 days before release seems a bit.. risky. ;)
[21:48] <jbernard> SpamapS: yes, _but_ it fixes two security bugs
[21:48] <SpamapS> jbernard: are the bugs related to the boot ordering?
[21:48] <semiosis> guntbert: the complete output of what?  apt-get upgrade?
[21:49] <jbernard> SpamapS: the are not,
[21:50] <jbernard> SpamapS: buffer overflows, both of them i believe
[21:50] <SpamapS>         [ ! grep "^cgroup" /proc/mounts &> /dev/null ] && { stop; exit 0; }
[21:50] <guntbert> semiosis: the command from which you showed us one line ^^^
[21:50] <SpamapS> double negatives are bad in english *and* shell script. :)
[21:50] <guntbert> semiosis: but please use a pastebin!
[21:50] <jbernard> SpamapS: no problem, i can fix that
[21:50] <SpamapS> grep -q is your friend
[21:50] <SpamapS> all pre-start's are run with set -e
[21:51] <SpamapS> so I usually like to use     test || { stop; exit 0;}  ...
[21:51] <SpamapS> jbernard: keep in mind I'm not filtering my review at the moment.. you might be ok with it "as is"
[21:53] <SpamapS> But yeah, I'm surprised the ! &&'s haven't bit you .. I'd think  [ ! grep ... ] && foo would mean if it does find what its looking for it would exit 1
[21:54] <jbernard> SpamapS: no problem, ill clean that up
[21:54] <jbernard> SpamapS: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=615987
[21:54] <SpamapS> jbernard: I'm testing right now if that actually is true
[21:56] <kirkland> \o/
[21:56] <kirkland> rabbitmq-stomp is now built in natty!
[21:56] <jbernard> kirkland: well done!
[21:57] <semiosis> guntbert: http://pastebin.com/TURKEzAh
[21:58] <SpamapS> kirkland: what was broken?
[21:58] <kirkland> SpamapS: build dep
[21:59] <kirkland> SpamapS: needed to symlink an erlang library into the expected /usr/lib/erlang location
[21:59] <semiosis> guntbert: anyone can reproduce this error by creating an EC2 instance from ami-cef405a7 (the official 64-bit maverick image in us-east-1 for ebs-root) and do apt-get update, then apt-get upgrade
[21:59] <SpamapS> fun
[22:00] <semiosis> guntbert: which worked fine until this afternoon
[22:01] <guntbert> semiosis: in that case I suggest you file a bug report (although I don't know against what package) - the actual error is reported in lines 629,630
[22:01] <SpamapS> jbernard: also no []'s
[22:03] <semiosis> guntbert: how about package language-selector-common?
[22:04] <guntbert> semiosis: good idea :-) (I obviously was not thinking straight)
[22:04] <semiosis> guntbert: been there myself, thanks for the help :)
[22:04] <guntbert> semiosis: :)
[22:05] <SpamapS> jbernard: what shell syntax is that { } stuff? It doesn't work in dash.
[22:06] <semiosis> guntbert: https://bugs.launchpad.net/ubuntu/+source/language-selector/+bug/766534
[22:20] <hallyn> SpamapS: '[' is a program
[22:20] <hallyn> SpamapS: and upstart seems to find them just fine...
[22:21] <semiosis> so that language-selector bug is in state Fix Committed, any idea how long it takes to get from there into the repository?
[22:21] <hallyn> SpamapS: and oddly enough, '{' seem sto be working
[22:21] <hallyn> SpamapS: when you say 'doesn't work', do you mean it'll be unpredictable?
[22:22] <hallyn> SpamapS: in dash:
[22:22] <hallyn> $ [ -f /etc/aliases ] && { echo -n 'hi '; echo 'there'; }
[22:22] <hallyn> hi there
[22:22] <hallyn> :)
[22:23] <SpamapS> hallyn: [ grep works ?
[22:23] <SpamapS> hallyn: for me { } wasn't working
[22:24] <hallyn> /etc/init/ssh.conf and mountall.conf show exacmples of them too
[22:24] <SpamapS> hmm I probably forgot the trailing ;
[22:24] <hallyn> well, and you ahve to have a space
[22:24] <SpamapS> hallyn: I get [] .. but not    [ grep something ]
[22:24] <hallyn> that gets mundane
[22:25] <SpamapS> $ [ grep foo ] && { echo -n "foo "; echo "found"; }
[22:25] <SpamapS> [: 3: grep: unexpected operator
[22:26] <SpamapS> hallyn: and you see my point about using || not && ?
[22:26] <hallyn> no
[22:27] <SpamapS> hallyn: exit code of the statement needs to be 0
[22:27] <SpamapS> && means it was 0, then you did the stop ; exit
[22:27] <SpamapS> so otherwise it will be 1, and the set -e kicks in and stops your pre-start
[22:27] <RoyK> use both
[22:27] <hallyn> stupid -e
[22:28] <RoyK> $[[ `true` ]] && echo true || echo false
[22:28] <SpamapS> thats confusing when there's only one action needing to happen
[22:28] <SpamapS> grep -q "thing" || { stop; exit 0; }
[22:29] <hallyn> SpamapS: well it's a good thing I did this, these are all things which I didn't quite know, and which I *do* know have bitten me in lxc upstart jobs (but i never knew why)
[22:29] <hallyn> so, two things:
[22:29] <SpamapS> hallyn: you are not alone. ;)
[22:30] <hallyn> jbernard: SpamapS : let's push only the security fixes (not the upstart ones) for natty, and queue up the upstart ones for the first week of o-series being open
[22:30] <hallyn> SpamapS: let me do a merge request for the upstart tree, assign you as reviewer, and you reject it and put all of the above feedback in there?  That way I ahve a permanent record
[22:30] <jbernard> hallyn: sounds good, just pop the upstart patches off the top and you're good to go
[22:30] <SpamapS> hallyn: does it still work w/o the upstart job to make sure it starts at the right spot in the boot?
[22:30] <SpamapS> hallyn: good plan. :)
[22:31] <hallyn> SpamapS: no
[22:31] <hallyn> SpamapS: so it wont' play nice with libvirt
[22:31] <SpamapS> hallyn: that sux
[22:31] <hallyn> jbernard: will do
[22:31] <hallyn> SpamapS: yeah...
[22:31] <SpamapS> hallyn: libcgroup is still in universe right?
[22:31] <hallyn> yes
[22:31] <SpamapS> hallyn: worth it I think to fix the bug then
[22:32] <SpamapS> as much as I'm tearing your shell script apart.. its fairly straight forward as an upstart job goes. :)
[22:32] <hallyn> drat
[22:32] <hallyn> lemme translate for SpamapS
[22:32] <hallyn> "it's a pretty trivial upstart script to write.  You just messed up so badly, it make sme want to facepalm"
[22:32] <hallyn> :)
[22:33] <SpamapS> http://lolpics.se/pics/482.jpg
[22:34] <hallyn> weird, this is the first pentadactyl bug I've seen in months.  But when I click on SpamapS in 'choose reviewer', it has deselected the input field, and doesn't fill the input field in
[22:35] <hallyn> you, sir, have a pending review
[23:08] <hallyn> jbernard: it occurs to me, the 'grep "^cgroup" /proc/mounts" check in cgred's init script is redundant.  cgred won't run unless cgconfig has run (and succeeded), so it seems like it's guaranteed that cgroup is already mounted
[23:08] <hallyn> I'll just leave a comment by the code, rather than remove it now
[23:16] <hallyn> SpamapS: (I've gone ahead and pushed lp:~serge-hallyn/upstart/natty/libcgroup/upstart3 to start addressing your feedback, but will still re-compare to your feedback on the other tree)
[23:16] <hallyn> or would that be lp:~serge-hallyn/ubuntu/natty/libcgroup/upstart3 :)
[23:44] <SpamapS> hallyn: just update the branch you submitted for MP
[23:44] <SpamapS> hallyn: Still wading through a few other things
[23:45] <hallyn> SpamapS: hm, ok - i wasn't sure if htat would muck up your review
[23:47] <SpamapS> hallyn: yes the review is active basically forever
[23:47] <hallyn> sure, but do your comments get targeted at specific revisions?
[23:47] <SpamapS> hallyn: idea being I can mark it "needs fixing" multiple times then you just re set it to "Ready for review" when you think you've fixed it
[23:48] <SpamapS> yes my comments will show above revisions pushed
[23:48] <SpamapS> so if I comment now, and then you push, your push will come after the comment