[00:01]  * hallyn out - maybe bbl
[00:01] <stgraber> hmm, and now my tty works fine ...
[00:01] <stgraber> hallyn: I'll do some more tests and e-mail you the results. I don't have a working C implementation though (and I believe you'd be much faster at doing that than I'm ;))
[01:04] <Tohuw> I'm reinstalling Ubuntu Server 11.10 on a system where I had previously setup an LVM to use software RAID 1. I no longer want to use software RAID, but now whenever I attempt an install, the installer reports it has found one or more disks containing a RAID configuration. How can I remove this configuration? I have reformatted the drives... I also ran dmsetup and cleared the device mapper... what am I missing?
[01:05] <cloakable> Blank the drives, iirc.
[01:06] <cloakable> Just zero the drives for a bit, enough to wipe the raid headers.
[01:06] <Tohuw> cloakable: what would be the most reliable way to do this? I have the server install and I can get to a shell from there using the "Rescue a broken system"
[01:07] <kerframil> with the 1.0 metadata format, the superblock is at the end
[01:07] <kerframil> anyway, surely the installer can be told to shut up and proceed (it can via preseed, at least)
[01:08] <Tohuw> I told it not to activate the raid devices, which lets me partition as I would expect, but the  resulting installation does not have a working GRUB
[01:08] <Tohuw> it blanks out before even displaying what stage it is loaading
[01:09] <cloakable> Tohuw: "dd if=/dev/zero of=/dev/<hard disk device>" iirc.
[01:09] <cloakable> May take a while, depending on disk size.
[01:10] <Tohuw> 2x 1.5 TB, SATA II... coffee breaks
[01:11] <cloakable> Mmhm.
[01:11] <cloakable> Run the command in parallel, leave it a while.
[01:12] <Tohuw> Can someone point me to  TFM to R so I can understand superblocks, and the way partitioning actually works below the filesystem level?
[01:13] <Tohuw> I understand RAID in a practical sense, and I understand LVMs, but not the mechanics; e.g. what's actually happening under the hood and where that is living in relation to what is actually visible to the OS
[01:20] <Tohuw> cloakable: what is the correct way to run this command in parallel? I'm having difficulty grokking parallel's page on GNU.org
[01:22] <cloakable> Tohuw: "dd if=/dev/zero of=/dev/<hard disk device> &" will run the command in the background until it completes.
[01:33] <Tohuw> cloakable: yeah, I ended up doing that... I was trying to get all fancy with it >_< clearly coffee time
[01:33] <cloakable> Haha
[03:55] <BuenGenio> hello, what are these  -- MARK --
[03:55] <BuenGenio>  messages that keep being logged in /var/log/messages?
[03:56] <savid> Has anyone seen any issues where virtual interfaces won't show up in ifconfig?
[03:56] <savid> The interface can be brought up, and it works, just not visible to ifconfig apparently
[04:22] <rf5> hello. im trying to run the apache in my virtual server i installed apache2 but still cant access the ip
[04:29] <fluvvell> I found linux-ip.net stuff to help me set up two internet connections that I have, but the examples show them coming in on two different interfaces, whereas I have two routers that are connected on the same lan.
[04:29] <fluvvell> I want all traffic coming in on one connection, to go back out the same connection.
[04:34] <Pupuser> Hello, Linux noob here. Does anyone know if I can use wubi to install ubuntu server, or any other ubuntu distro?
[04:42] <tash> has anyone had problems with apache2 and php5 on Ubuntu server 11.10?
[04:42] <rf5> i am having tash
[04:42] <rf5> mine is not even opening lol
[04:43] <twb> fluvvell: you can't have that
[04:43] <tash> I'm getting a white screen :)
[04:43] <tash> I moved my site from old server to this, and can't get it to load ... :(
[04:43] <tash> the only thing that I can get to work is the It Works! page
[04:43] <fluvvell> twb, I cant have two incoming internet connections?
[04:43] <twb> fluvvell: it's possible if you have an AS (i.e. you're an ISP).
[04:44] <rf5> i cant get it working tash.
[04:44] <qman___> not like that at least, you need a load balancer
[04:44] <tash> I don't know if I'm missing VirtualHost config or what
[04:44] <fluvvell> what do you mean by an AS
[04:44] <tash> anyone able to help rf5 and I out?
[04:44] <twb> fluvvell: otherwise you can approximate it with a lot of fiddly netfilter and ip rule rules, which are beyond your capabilities
[04:44] <rf5> what did u do tash, just apt-get install apache2?
[04:45] <tash> I did apt-get install php5 and it installed apache2 for me
[04:45] <tash> and then apt-get install mysql-server php5-curl curl
[04:45] <twb> fluvvell: each ISP gets an AS number.  The backbone of the internet has an entry in each backbone router for every AS.  An AS means you are an autonomous part of the internet, you answer to no one.  This allows you to populate GRE tables with multiple routes from <the internet> to you.
[04:45] <fluvvell> twb: http://linux-ip.net/html/adv-multi-internet.html, http://kindlund.wordpress.com/2007/11/19/configuring-multiple-default-routes-in-linux/
[04:45] <twb> I haven't investigated the cost, but getting an AS these days would probably cost in the tens-of-thousands range
[04:46] <twb> fluvvell: those URLs discuss the "appoximation that's over your head" approach
[04:47] <fluvvell> twb, I'm always keen to learn further.  I try not to talk up my abilities, but I understand the concept of routing tables.
[04:47] <twb> fluvvell: you should take this to #netfilter
[04:48] <fluvvell> twb, ok. Its always good to get some advise anyway.
[04:50] <twb> fluvvell: when you go there, I will tell you more
[04:50] <fluvvell> twb, thanks.
[07:18] <Guest22249> hey, I need help please... I have an apt-mirror server which has updated now, with the normal things and debian-installer, running 11.10. I want to net install my computers on the network using a USB, everything work 100%, but there is no ubuntu desktop, i have to ssh in and tasksel multiselect ubuntu-dekstop. How do i get it to install ubuntu desktop? currently using kickstart and preseed together, but it does not install ubuntu desktop. my prese
[07:18] <Guest22249> ed.cfg tells it to install ubuntu-desktop
[07:18] <Guest22249> ed?
[07:18] <Guest22249> please note i am a newbe
[07:18] <Guest22249> how do i tell the USB installer to use ed.cfg?
[07:18] <Guest22249> hey, I need help please... I have an apt-mirror server which has updated now, with the normal things and debian-installer, running 11.10. I want to net install my computers on the network using a USB, everything work 100%, but there is no ubuntu desktop, i have to ssh in and tasksel multiselect ubuntu-dekstop. How do i get it to install ubuntu desktop? currently using kickstart and preseed together, but it does not install ubuntu desktop. my prese
[07:18] <Guest22249> ed.cfg tells it to install ubuntu-desktop
[07:22] <_ruben> Guest22249: you're preseed file likely wrong then, also be sure to check the installer's logs for any oddities
[07:22] <_ruben> i dont do desktop preseeds, so dunno if there's any caveas
[07:22] <_ruben> caveats
[07:26] <armand-jhb> well... i took the preseed file from the ubuntu cd.... so it cant be wrong
[07:27] <armand-jhb> in there, there is no way that it can call other files
[07:27] <armand-jhb> the preseed file tells it to also install ubuntu-desktop... but nothing
[08:14] <Tribaal> HI all
[09:24] <koolhead17> hi all
[10:08] <linocisco> i have created apt.conf under /etc/apt. I dont want to reboot. how to refresh or update that entry in apt.conf?
[10:12] <Daviey> jamespage: do you know why we have a karmic PPA on https://jenkins.qa.ubuntu.com/view/Precise/job/precise-server-ec2/ARCH=i386,REGION=us-west-2,STORAGE=instance-store,TEST=cloud-config,label=ubuntu-server-ec2-testing/3/console ?
[10:13] <dns53> linocisco why would a reboot be required?
[10:13] <jamespage> Daviey: yes - the cloud-config specifies it as a manually added PPA
[10:13] <jamespage> we don't actually care that its karmic - just that its present (proves cloud-init did its stuff)
[10:14] <linocisco> dns53, because i have put some entries in apt.conf like proxy authentication
[10:15] <Daviey> jamespage: I wonder, is it a good idea to use one which is outside ubuntu developers control? :)
[10:15] <Daviey> mind you, it can't be uploaded it.. the release is EOL'd
[10:16] <dns53> linocisco i had assumed that it was always read when apt is used and it would be used on the fly
[10:16] <jamespage> Daviey: feel free to raise a bug - https://bugs.launchpad.net/ubuntu-server-ec2-testing/+filebug
[10:16] <jamespage> its good point but low priority IMHO
[10:21] <linocisco> dns53, no
[10:22] <linocisco> dns53, there is no apt.conf in 10.04 LTS server version
[10:23] <smb> There is no apt.conf by default, but it is honoured as soon as it is there
[10:24] <smb> (only using it to change the proxy myself)
[10:25] <smb> dns53, If you sort of did the same, the next time you run apt-get it is in effect (without rebooting)
[10:26] <dns53> that is what i have seen in the past, it will read an apt.conf first, then apt.conf.d/* then command line options to apt so it should not require anything to update
[10:27] <dns53> it has been a while since i had to use a proxy but it did just require updating /etc/apt/apt.conf with the settings
[10:28] <smb> dns53, right. Oh and I think I confused person asking and anwering. sorry
[10:28] <smb> linocisco, ^
[10:28] <linocisco> smb , what?
[10:28] <smb> linocisco, The file is not present by default but it will get used the next time you use any apt-get command
[10:29] <dns53> linocisco no need to reboot, just run apt again and it will use the settings,   there are also command line options where you could specify the proxy server if you want to test things before updating the config file
[10:30] <linocisco> dns53, apt again means ?? apt-get update or what?
[10:30] <smb> Or apt-get install
[10:30] <dns53> linocisco yes, apt-get, aptitude, synaptic etc
[10:31] <linocisco> dns53, the error i got is "extra junk at the end of the file
[10:31] <linocisco> "
[10:32] <dns53> linocisco with the config file or during a transfer?
[10:32] <smb> does the line in apt.conf look like
[10:33] <smb> Acquire::http::Procy "http://<host>/";
[10:33] <smb> *Proxy
[10:37] <linocisco> smb, I got it now
[10:37] <linocisco> syntax error
[10:37] <smb> linocisco, ok
[10:37] <linocisco> smb, are you samba expert?
[10:38] <smb> linocisco, not really. using it only a bit
[10:43] <linocisco> who is using vbox with windows to host  ubuntu server guest?
[10:45] <dns53> so you want to mount a windows share under ubuntu?
[10:48] <dns53> is your computer part of a windows domain?
[10:49] <linocisco> dns53, no. our server is NOvell
[10:49] <linocisco> dns53,  clients are windows XP
[10:50] <dns53> they still exist? i thought they died with the dinosaurs
[10:50] <linocisco> dns53, actually I want to setup ubuntu mail server and mail clients from windows will use
[10:50] <dns53> do you want to connect to the novell network?
[10:51] <linocisco> dns53, all clients XP are installed with Novell clients for widnows software to authenticate network drive and internet access. now
[10:54] <dns53> so you want the users to be able to login to email etc using their novell credentials?
[10:55] <linocisco> dns53, no. it is to test for my own with my test users
[10:56] <linocisco> dns53, I have only windows XP which has internet. now it hosts ubuntu server with vbox (NAT) setup. i have installed mail server role with internet site.
[10:58] <linocisco> dns53, before I start anything, I am trying to start "ping" from host to guest and guest to host
[10:58] <dns53> you don't want nat you want a bridge, you can ping up the chain but not down the chain in nat
[11:03] <linocisco> dns53, yes. I can ping from guest to host. but not host to guest
[11:03] <linocisco> dns53, what do I do?
[11:04] <linocisco> dns53, what type of adapter would be better or more suitable?
[11:07] <dns53> linocisco well you will need to set the vm to use bridge networking (other computers on the network can connect to the vm) or internal (only your computer can connect to the vm)
[11:10] <dns53> windows i believe you just need to shut down the vm, go to the networking config page and you should have the option of specifying the type of network
[11:14] <linocisco> dns53, in my setup, i dont need other computers to connect to my ubuntu mail server. Just host XP's mail clients thunderbird has to connect to Guest" ubuntu server.".
[11:17] <dns53> linocisco so configure the network interface in virtualbox to use internal networking
[11:19] <linocisco> ok
[11:28] <linocisco> dns53, so , i think i should make virtualbox host only adapter on my xp and ubuntu-server's eth0 in same subnet or network. right?
[11:32] <dns53> linocisco i believe so, just choose an ip that does not exist on your normal network
[11:33] <linocisco> dns53,  pretty sure. thanks alot
[11:35] <jamespage> lynxman, around - have a rabbitmq-erlang-client query for precise
[11:43] <linocisco> btw, how to give shutdown/reboot right to normal user or first user created on ubuntu server
[11:43] <linocisco> maning default user during installation
[11:43] <linocisco> meaning default user
[11:46] <dns53> you can add a user to the admin group,    sudo   usermod -a -G admin  username
[11:50] <lynxman> jamespage: sorry wasn't paying attention, shoot :)
[11:50] <jamespage> lynxman, OK; so I noticed that that rabbitmq-erlang-client is still at 2.5.0
[11:50] <jamespage> in precise
[11:51] <lynxman> jamespage: hmm I sent all the packages to zul to pump them to 2.6.1
[11:51] <lynxman> jamespage: at the Budapest rally
[11:51] <lynxman> jamespage: there's erlang client and stomp
[11:51] <jamespage> lynxman, OK - looks like this one got missed - I'll take a loop
[11:51] <linocisco> dns53, is "admin" a group name already created?
[11:51] <lynxman> jamespage: darn
[11:51] <lynxman> jamespage: it's on my PPA
[11:51] <jamespage> rabbitmq-stomp FTBFS at the moment
[11:51] <lynxman> jamespage: I'll ping zul later then
[11:52] <dns53> yes, the admin is the group that the first user gets that makes sudo etc work
[11:52] <lynxman> jamespage: it'll fail if erlang-client is not there, it depends on it
[11:52] <jamespage> yep
[11:52] <lynxman> jamespage: I'll ping zul when he's awake and get back to you as soon as it's sync
[11:53] <jamespage> lynxman, great - if he is busy I can upload for you
[11:53] <jamespage> I think friday is openstack upload day!
[11:54] <lynxman> jamespage: if you wish so... :) it's in https://launchpad.net/~lynxman/+archive/ppa/+packages
[12:18] <jamespage> lynxman, OK - I'll check with zul and upload
[12:19] <zul> hmmm?
[12:23] <Daviey> Anyone fancy giving me a peer review of a diff?
[12:23] <Daviey> http://pb.daviey.com/bWFR/
[12:24] <lynxman> jamespage: thanks :)
[12:24] <lynxman> zul: morning!
[12:24] <lynxman> Daviey: looks good, pretty straight forward
[12:25] <Daviey> lynxman: you missed something :)
[12:25] <lynxman> Daviey: ? it's just adding a patch innit
[12:26] <Daviey> lynxman: missed the bug number on the d/changelog :)
[12:26] <lynxman> Daviey: lol
[12:26] <lynxman> Daviey: it's on the patch
[12:26] <lynxman> Daviey: it's always required in both?
[12:27] <Daviey> nah, but the fact that there is one - means it should be in the cl
[12:28] <lynxman> Daviey: aah I see
[12:30]  * lynxman always pays not much attention to changelogs
[12:30] <jamespage> Daviey, looks OK to me - just uploaded a very similar one for dahdi-linux
[12:30] <Daviey> jamespage: yeah, you sponsored it right?
[12:31] <jamespage> yep
[12:31]  * Daviey strikes that off his todo :)
[12:31] <Kiall> Daviey: I wen't with the 3.2 mainline on oneiric due to constant CPU stalls and and kernel panics with the stock kernels...
[12:31] <Kiall> (I filed / added to bugs.. but nothing ever came of them sadly)
[12:31] <Daviey> Kiall: crikey, smb/apw ^^ is that news to you?
[12:32] <Kiall> This is on a bunch of HP DL165 G7's BTW..
[12:32] <Daviey> Kiall: can you point me to the bug numbers please?
[12:32] <ikonia> Kiall: are these the amd or xeon 165's ?
[12:32] <jamespage> zul: rabbitmq-erlang-client needs an upload for 2.6.1 - lynxman has prepared - OK if I upload for him?
[12:32] <Kiall> AMDs
[12:32] <smb> Daviey, not sure which stalls and panics actually. Running the oneiric kernel without those
[12:32] <ikonia> Kiall: there is a bug on the intel
[12:33] <Kiall> Will see If I can dig them out.. There were a few others filed that I didn't comment on though. Probably never going to find those ones again...
[12:33] <apw> Daviey, ?
[12:33] <zul> jamespage: yep
[12:34] <apw> Daviey, we use oneiric kernels routinely on servers, so i assume its not a generic issue
[12:34] <Kiall> Daviey: here is one .. https://bugs.launchpad.net/ubuntu/+source/linux/+bug/905219
[12:34] <Daviey> apw: Kiall is saying he had regular freezes with a class of HP servers on Oneiric.
[12:34] <apw> Daviey, and i assume as a server person you are doing lots of testing too and would have noticed
[12:34] <Kiall> Regular as hell ;) https://launchpadlibrarian.net/89219509/uptime-day.png
[12:34] <Daviey> apw: I was more wondering if you had come across bugs with it.
[12:34] <apw> Daviey, i am sure we arn't aware of it, unless you have been made aware and made either of us aware :/
[12:35] <Daviey> apw: Heh, ok.. (i don't triage incoming kernel bugs fwiw.
[12:35] <Kiall> Daviey: Another two https://bugs.launchpad.net/ubuntu/+source/linux/+bug/870168 and https://bugs.launchpad.net/ubuntu/+source/linux/+bug/881968
[12:35] <ikonia> Daviey: there is a HP bug filed also I'll see if I can get that, although it's on the Intels' not the AMD
[12:36] <ikonia> (so may be a different issue with similar symptoms)
[12:36] <apw> Daviey, indeed, and nor does anyone else, there are far tooo many to get any real depth on any of them
[12:36] <smb> first one seems to require netfilter involved
[12:36] <ikonia> it's the 165's and a certain blade model
[12:37] <Kiall> ikonia: I'm seeing those three issues on AMD 165's
[12:37] <ikonia> yes, the issue (same symptoms) is known to HP on the intels but not the xeons
[12:38] <jamespage> lynxman, rabbitmq-erlang-client uploaded
[12:38] <Kiall> not on AMD I assume? anyway.. the 3.2 mainline has been rock solid for me on the same hardware.
[12:39] <smb> And is this real hw? I see kvm_mmu blah in the first bugs dmesg
[12:40] <Kiall> Yea, Its real hardware hosting KVM vms
[12:44] <Kiall> Also - I actually know of another person who's been experiencing #905219 (the first one), but fairly infrequently and not on HP hardware.
[12:44] <Kiall> Let me dig out the email to see if the trace is identical...
[12:46] <linocisco> hi i made"   sudo   usermod -a -G admin  username" but still can't shutdown with that username
[12:46] <smb> apw, It would help if those bug reports would not have a mix of various traces. Some with the beginning, some from the end ...
[12:48]  * cwillu_at_work is impressed by the screenshot of a remote viewer of a remote viewer
[12:49] <Kiall> cwillu_at_work: that would be mine ;)
[12:49] <cwillu_at_work> Kiall, that "view" menu doesn't maintain more of a log? :p
[12:49] <Kiall> Nope!
[12:49] <cwillu_at_work> and you can't copy/paste out of it?
[12:49] <Kiall> Which bug was that BTW, because that was likely a dell server from memory
[12:50] <Kiall> No - Its a VNC session
[12:50] <Kiall> (Via a RDP session)
[12:50] <cwillu_at_work> Kiall, conveniently, both things which support shared clipboards :p
[12:51] <Kiall> This VNC server doesn't do clipboards ;)
[12:54] <Kiall> Daviey: re http://pb.daviey.com/bWFR/
[12:54] <Kiall> should that not have been a new file in /usr/src/iscsitarget-1.4.20.2/patches/compat-3.2.patch ?
[12:56] <linocisco> anybody help me to setup working email server?
[12:56] <ikonia> linocisco: what are you not sure on ?
[12:57] <linocisco> ikonia, i have just install ubuntu server on vbox and i chose mail server option during setup. now postfix is running at loading message
[12:57] <linocisco> ikonia, i want to make it running mail server
[12:58] <ikonia> linocisco: ok - so what part are you not sure about ?
[12:59] <linocisco> i read flurdy page. too complicated
[13:00] <Kiall> linocisco: there are plenty of guides online, It would be worth going through one of those.. Without a specific problem, IRC tends to be fairly useless :)
[13:00] <Kiall> This one looks decent https://help.ubuntu.com/community/PostfixBasicSetupHowto
[13:00] <linocisco> ikonia, i just want to know if I need internet to install any component to mail server
[13:01] <ikonia> linocisco: you will need the internet to install software from the repos and updates, or use the CD to get software off the CD
[13:01] <linocisco> ikonnia. now i just wanted to email from one user to another
[13:02] <linocisco> can i quickly make it done in simple way ?
[13:02] <ikonia> linocisco: ok - so you want to mail internally over the internet, or internally on the same server ?
[13:02] <linocisco> ikonia, first only between internally
[13:03] <linocisco> ikonia, from user1 to user2 and vice visa
[13:03] <ikonia> linocisco: ok, so when you setup the mail server there was an option that allows you to send internet over the internet or internally, which did you select
[13:04] <linocisco> ikonia," internet site"
[13:04] <ikonia> linocisco: why did you select "internet" if you wanted to mail internally ?
[13:04] <linocisco> ikonia, first i thought it might be easiler
[13:04] <ikonia> linocisco: why ??? what reason would it be easier ?
[13:05] <ikonia> linocisco: the whole point of that menu is to ask you what you want to do, and you select the option you need
[13:05] <ikonia> linocisco: you wanted to "mail internally" yet you selected "mail internet", I cannot grasp how that would be thought of as an easier option than the option that did exactly what you wanted
[13:06] <linocisco> ikonia, actually in real setup with virtual box, it is not easier to give mail access to host connected to internet from guest ubuntu server
[13:06] <linocisco> ikonia , so should I choose dpkg-reconfigure postfix and setup  again?
[13:07] <ikonia> linocisco: to be honest, I think you need to look at what you actually want to do, then research what you need to do to set that up, then ask questions on anything you are unsure about or something that is not working
[13:08] <linocisco> ikonia,  i agree you will think so
[13:08] <linocisco> ikonia, i want to get it done fast. i chose my final dream setup
[13:09] <ikonia> linocisco: getting it right/secure is the priority
[13:09] <ikonia> linocisco: getting something done "fast" as the priority for a mail server is not a good model
[13:09] <linocisco> ikonia,  i m not seting up production server now. i just want to feel myself i setup working test mail server
[13:10] <ikonia> linocisco: again right/secure is still (in my view) what you need to focus on
[13:10] <ikonia> anyway, it's up to you, good luck
[13:11] <linocisco> ikonia, i wish you could suggest what to install and how to configure
[13:11] <ikonia> linocisco: you've been given a link
[13:11] <ikonia> linocisco: if you want to learn how to setup a mail server, get researching rather than asking "what do I type"
[13:11] <linocisco> ikonia, that is for internet site
[13:11] <ikonia> linocisco: apply thought and change approach for internal
[13:12] <linocisco> ikonia, i m linux newbie. so please kindly understand me
[13:12] <ikonia> linocisco: I do understand you
[13:13] <linocisco> ikonia, i run dpkg-reconfigure postfix. should I choose locally?
[13:13] <ikonia> linocisco: apply thought process in your head. Do you want to send mail locally or over the internet ?
[13:13] <linocisco> ikonia, it has "no config", "internet site", "internet with smart host", "locally"
[13:14] <linocisco> ikonia, locally in meaning. but not emailing from system to root
[13:14] <linocisco> ikonia, i want locally between two users
[13:15] <ikonia> linocisco: root is a user, so again, think it through, do you want to send mail over the internet
[13:15] <linocisco> ikonia, not now
[13:15] <ikonia> linocisco: do you want "no configuration"
[13:16] <linocisco> ikonia, just between two users at least and i can sleep well tonight
[13:16] <ikonia> ok - I give in now
[13:16] <ikonia> I shouldn't have to spoon feed at this level for a mail server, good luck
[13:16] <linocisco> ikonia, if i choose no config, what do i do next?
[13:17] <ikonia> linocisco: you don't choose that, I'm not going to help you with this. Please ask someone else
[13:17] <linocisco> ikonia, thanks anyway
[13:17] <ikonia> no problem, good luck
[13:17] <linocisco> ikonia, having no physical server sucks
[13:17] <ikonia> it makes no difference
[13:28] <linocisco> what is the error "update-alternatives"warning: forcing reinstallation of alternative /usr/bin/bsd-mailx because link group mailx is broken"?
[13:45] <lynxman> jamespage: can I ask you a couple ipxe related questions? :)
[13:48] <jamespage> lynxman, fire away!
[13:49] <lynxman> jamespage: when I do the upstream merge, you told me I could do that and my changes to the debian structure in a single commit? Or do two commits but keep the package version the same?
[13:50]  * jamespage scratches his head
[13:50] <jamespage> I think that you get one commit for the import of the upstream tarball
[13:50] <jamespage> and then you have todo another for the packaging changes
[13:51] <lynxman> jamespage: so two commits but keep just one package version in changelog then, right?
[13:51] <jamespage> yes
[13:51] <jamespage> bzr commits != changelog entries
[13:52] <jamespage> lynxman, it might work out like that once the package gets imported to the official branch post upload
[13:52] <lynxman> jamespage: yeah of course, I just wasn't sure what was the procedure on this one
[13:52] <jamespage> so takes as many commits as you like on your working branch - they get flattened post upload
[13:53] <lynxman> jamespage: fair enough, thanks :)
[13:54] <Tixos> hi, is anybody knowledgeable in DNS that could help me configure BIND?
[13:56] <linocisco> hi all , I am following "https://help.ubuntu.com/community/PostfixBasicSetupHowto". but i have question on "Setting Postfix Support for Maildir-style Mailboxes"
[13:57] <linocisco> when I  su - fmaster,
[13:57] <linocisco> MAIL=/home/fmaster/Maildir,
[13:57] <linocisco> mail  , I found no mail
[13:59] <linocisco> what do I do?
[14:07] <Daviey> Kiall: No, it's a distro patch.
[14:07] <Kiall> Ah - fair enough
[14:07] <Kiall> I had thought they were distro patches too, considering they are only used by debian/rules
[14:07] <Kiall> (a different kind of distro patch that is..)
[14:09] <Kiall> I had figured that style of patch was used to allow for a broader range of kernel support in a single package.. with a normal patch, it always gets applied regardless of the target kernel version
[14:18] <linocisco> i got problem with netcat
[14:19] <linocisco> netcat: getaddrinfo: Name or service not known
[14:19] <linocisco> "netcat: getaddrinfo: Name or service not known" . what do i do??
[14:21] <linocisco> Ursinha-lunch, hello
[14:23] <linocisco> hi all
[14:23] <linocisco> netcat: getaddrinfo: Name or service not known
[14:24] <greppy> linocisco: what is the command you are trying to use with netcat?
[14:25] <linocisco> greppy, netcat mail.mydomain.com 25
[14:26] <greppy> does mail.mydomain.com exist?
[14:27] <linocisco> greppy, it is on server
[14:27] <linocisco> greppy, not registered domain name on ICANN
[14:29] <linocisco> soren, hi bro
[14:29] <linocisco> soren, as a long term geek, could you help me?
[14:31] <greppy> linocisco: so, can you do "ping mail.mydomain.com"?
[14:31] <linocisco> greppy, from CLI ?
[14:31] <greppy> yes.
[14:32] <linocisco> greppy, unknown host
[14:32] <greppy> well, there is your problem with netcat as well.
[14:34] <linocisco> greppy, so what do I do?
[14:34] <linocisco> greppy, actually I want to setup postfix only one local user to another
[14:34] <linocisco> greppy, first. before I can setup internet based mail
[14:35] <greppy> when you installed postfix it should have asked if you wanted to be a local only mail site, if you selected that, it should work.
[14:36] <linocisco> gre
[14:39] <linocisco> greppy, i run dpkg-reconfigure postfix with locally
[14:39] <linocisco> greppy, stil netcat is not working
[14:39] <greppy> locally it won't be listening on port 25
[14:40] <greppy> because it's a local only server
[14:40] <linocisco> greppy, so what to test if I m setting up mail server from one user to another like user1@mydomain.com to user2@mydomain.com
[14:41] <greppy> from a shell, "echo 'testing' | mail user2@mydomain.com"
[14:41] <greppy> then check the logs in /var/log/mail.* to see what happens.
[14:45] <linocisco> greppy, "echo 'testing' | mail user2@mydomain.com" shows nothing
[14:46] <greppy> right.
[14:46] <greppy> that should end up with an email sent to user2
[14:46] <linocisco> greppy, if mail server setup is so difficult, let it be
[14:47] <linocisco> greppy, i will just try how to give shutdown right to default server user
[14:47] <linocisco> greppy, "sudo   usermod -a -G admin  username" doesnot work
[14:48] <greppy> linocisco: so the user can't do "sudo shutdown"?
[14:49] <linocisco> greppy, yes. it said need to be root
[14:49] <greppy> k
[14:49] <greppy> is the admin group in your /etc/sudoers file?
[14:50] <linocisco> greppy," sudo   usermod -a -G admin  username", admin group should exist?
[14:50] <greppy> sudo visudo
[14:50] <greppy> or, add that user to the sudo group instead of the admin group.
[14:51] <greppy> also, you will need to log out that user and log back in for the change to take affect, when you put a user in a new group.
[14:51] <linocisco> greppy, i typed sudo visudo, and i found sudoers file , what to do next?
[14:51] <greppy> check to see if there is an 'admin' group specification
[14:52] <linocisco> greppy, where to check?
[14:52] <linocisco> greppy, %admin=ALL
[14:52] <greppy> ok, so that should wokr.
[14:53] <linocisco> greppy, my default user is not found in that sudoers file
[14:53] <linocisco> greppy, how to add default user to admin group ?
[14:53] <linocisco> sudo   usermod -a -G admin  username is not enough??
[14:54] <greppy> that adds them to the group
[14:54] <linocisco> greppy, but i didn't see that user name in sudoer file
[14:54] <greppy> if you type 'id' in a shell, it should list the groups.
[14:54] <greppy> it won't be.
[14:54] <greppy> the user will be in the /etc/group file
[14:55] <greppy> if you don't see 'admin' in the output of 'id', then you need to log out and log back in.
[14:56] <linocisco> greppy, i tried logout and logi  n. i found my username is in the list twice
[14:57] <greppy> ok
[14:57] <greppy> so what happens when you type 'id'? does 'admin' show up as one of the groups?
[14:58] <Daviey> roaksoax: around?
[14:59] <linocisco> greppy, uid=1000(myusername) gid=1000(myusername) groups=4(adm), 20 (dialout), 117(admin)
[14:59] <linocisco> greppy, 1000(myusername)
[14:59] <linocisco> greppy, is that correct?
[15:00] <greppy> yup, so myusername is in the admin group, so 'sudo reboot' should reboot the box.
[15:02] <linocisco> greppy, not ok
[15:02] <linocisco> greppy, rebooting is done. not ok yet
[15:02] <linocisco> greppy, thanks anyway
[15:02] <linocisco> greppy, cu
[15:17] <roaksoax> Daviey: im here
[15:20] <Daviey> roaksoax: sorry, can i re-ping in 40 mins, on a call.
[15:21] <roaksoax> Daviey: sure
[16:08] <Tixos> can anyone help with this error, i installed apache2 and php as you would, so i dont see why there would be permission issues?
[16:09] <Tixos> PHP Notice:  session_start(): ps_files_cleanup_dir:
[16:11] <Daviey> zul: bug 930139 is fixed in both ks and ksl?
[16:11] <zul> just ks ill get to the ksl after im done the uploads
[16:15] <Daviey> zul: cool
[16:16] <Daviey> adam_g: How are the charms looking for ksl?
[16:22] <Tixos> can someone please paste me their cron for php5.3
[16:23] <Tixos> /etc/cron.d/php5
[16:24] <undecim> What does it mean when my (10.04) server responds to a ping, but I can'
[16:24] <undecim> t ssh to it, and there is no sign in /var/log that the system even booted
[16:25] <Tixos> how are you seeing /var/log if you cant ssh in
[16:25] <undecim> Took the hard drive out
[16:25] <undecim> looking at it with a SATA-USB adapter
[16:25] <Tixos> lol
[16:25] <undecim> It's clearly using the static IP from /etc/network/interfaces. I changed it and the IP changed
[16:25] <Tixos> how do you know if ssh was running
[16:26] <Tixos> what error did you get from trying to connect via ssh?
[16:26] <undecim> Because every other time I started the system (though admittedly only twice) ssh started
[16:26] <undecim> This time the system doesn't even seem to be booting, except that it responds
[16:26] <Tixos> so boot it again
[16:26] <undecim> connection refused
[16:26] <undecim> nmap indicates all ports are closed
[16:27] <Tixos> how have you installed this?
[16:27] <lynxman> jamespage: question for you, what would be the best way to clean a copyright file?
[16:27] <Tixos> what changes have you made?
[16:27] <Tixos> if it worked before, it indicates you broke something :)
[16:27] <jamespage> lynxman, update it? hmm
[16:27] <jamespage> licensecheck -r * is a good start
[16:27] <lynxman> jamespage: cool ty
[16:27] <undecim> 10.04 server amd64, sda1 for / and sdb, sdc in RAID 1 for /home
[16:28] <hallyn> SpamapS: jodh: say, exactly which script sections in an upstart job are run with 'set -e'?
[16:28] <hallyn> only pre-start?
[16:28] <undecim> Tixos: Yes, I'm sure I broke something, but that's what I'm trying to find
[16:28] <undecim> Tixos: What boggles my mind is that the network is started, but nothing is logged
[16:28] <lynxman> jamespage: I'm trying to solve lintian complaining about ipxe source: syntax-error-in-dep5-copyright paragraph 6 after the field
[16:28] <cjs226> running into an issue where omrelp doesn't send all logs to the remote server even though *.* is used.  for example, php errors aren't sent.  any ideas?
[16:29] <jamespage> lynxman, read http://dep.debian.net/deps/dep5/
[16:29] <jamespage> its probably an empty line in a license block
[16:29] <lynxman> jamespage: k
[16:30] <hallyn> no, i guess 'script' is also
[16:31] <hallyn> and post-stop.  ok, not sure why i thoughtonly pre-start did.  oh weel
[16:31] <SpamapS> hallyn: all
[16:34] <hallyn> SpamapS: yup, thanks :)
[16:36] <lynxman> jamespage: ipxe pushed to the branch ready for review (again)
[16:38] <jamespage> lynxman, OK - lemme take a look
[16:43] <roaksoax> Daviey: so what's up?
[16:53] <Tixos> can someone please paste me their PHP cron job !
[17:07] <jamespage> lynxman, all looks OK apart from the get-orig-source target - mind if I take a stab at getting it working right?
[17:08] <lynxman> jamespage: oh please do :) both Daviey and I worked on that one but somehow it wasn't 100% right on the spot
[17:08] <lynxman> jamespage: will be happy to copycat your optimisation for future packages *whistles*
[17:09] <jamespage> lol
[17:09] <jamespage> its not helped by the crap version numbering
[17:09] <jamespage> sorry  I should have said 'bad'
[17:10] <lynxman> jamespage: since the upstream maintainer didn't want to change it I couldn't myself :/
[17:15] <hallyn> SpamapS: if job b is 'start on starting a', will b start before a's pre-start script section?
[17:16] <SpamapS> hallyn: yes
[17:16] <hallyn> excellent.  thanks.
[17:16] <SpamapS> hallyn: b will start before a's goal is even changed to 'start'
[17:27] <hallyn> SpamapS: and pre-start runs after 'goal' is set to start?
[17:30] <hallyn> stgraber: http://people.canonical.com/~serge/lxc.debdiff is working for me so far.  do you have time to do the ureadahead+udevtrigger changes?  (then i'll try the console changes in lxc source, and a kernel patch for remount,ro)
[17:31] <SpamapS> hallyn: right.
[17:34] <stgraber> hallyn: any reason not to keep --no-restart-on-upgrade?
[17:35] <hallyn> stgraber: because the upstart jobs are smart enough not to mess up (i hope)
[17:35] <hallyn> stgraber: lxc.conf won't shut down containers unless entering runlevel [016],
[17:35] <hallyn> and lxc-net.conf won't shut down network if there are attached nics
[17:35] <hallyn> so, it wouldn't hurt, but i don't think it's needed
[17:35] <jamespage> lynxman, OK I think this works  - http://paste.ubuntu.com/836806/
[17:36] <lynxman> jamespage: and it looks a lot smoother than mine :)
[17:36] <jamespage> the rules file was trying to parse an iso/unix date from the version string which did not exist
[17:36] <jamespage> this just grabs the latest snapshot and increments that awkward prefix number used in the upstream version
[17:37] <stgraber> hallyn: I "think" I'd prefer that we drop these checks from the job and use --no-restart-on-upgrade. Because I can see some people expecting "stop lxc" to stop their containers and "start lxc" to start them
[17:37] <stgraber> hallyn: with the current jobs, they'd have to figure out they need to set RUNLEVEL
[17:37] <jamespage> lynxman, so 1.0.0+git-3.55f6c88 -> 1.0.0+git-4.d3630b6
[17:37] <jamespage> based on the snapshot now...
[17:38] <lynxman> jamespage: sounds good to me
[17:38] <jamespage> OK
[17:38] <lynxman> jamespage: all the changes I did are good anyway
[17:38] <jamespage> yep
[17:38] <hallyn> stgraber: hm.  guess that makes sense
[17:39] <hallyn> SpamapS: so if b is 'stop on stopped a', will b's post-stop happen after a is stopped?  (I assume so)
[17:40] <SpamapS> hallyn: yes
[17:41] <SpamapS> hallyn: init/job.c job_next_state is quite easy to read the state progressions, and then job_change_state shows what happens when each one changes
[17:42] <hallyn> stgraber: so what do you think lxc-net.conf should do at post-stop if lxcbr0 has attached interfaces?  exit with failure?  forcibly detach?
[17:43] <hallyn> i'll continue to just not shut it down and NOT return error
[17:44] <hallyn> SpamapS: ok
[17:44] <stgraber> hallyn: I belive the current behaviour of lxc-net is correct
[17:48] <jamespage> lynxman, all looks good (did a quick test as well)  uploading now
[17:48] <lynxman> jamespage: cool, thank you sir :)
[17:49] <jamespage> lynxman, np - thanks for being patient with my constant feedback!
[17:51] <lynxman> jamespage: my pleasure
[17:51]  * lynxman goes back to his jamespage shrine to pray some more
[17:51] <smoser> utlemming, was just thinking as this bit me.
[17:51] <jamespage> SpamapS, hey
[17:51] <smoser> ubuntu-cloudimg-query uses https to cloud-images.ubuntu.com
[17:51] <jamespage> nice spot on the jenkins reload problem
[17:52] <jamespage> could be an issue with java server stuff generally TBH
[17:52] <smoser> thats good (to avoid spoofing), but for the data transfer, it means that it wont be proxied (http_proxy=) which sucks.
[17:52]  * jamespage ponders whether zookeeper does the same thing
[17:52] <smoser> so just thinking that with the new data, including checksums and signatures, hopefully we can get the important data over https and the bulk over http
[17:53] <utlemming> smoser: that is what I was just thinking
[17:54] <utlemming> smoser: the new meta-data coming over https will give the authenticity with checksums, which should be more than adiquete to discover a man-in-the-middle
[17:54] <utlemming> smoser: for the bulk data
[17:55] <smoser> right.
[17:55] <smoser> so maybe we need to publish 'mirrors' at /query2 ?
[17:55] <SpamapS> jamespage: HUP needs to be handled. :)
[17:56] <SpamapS> jamespage: enjoy your weekend, we'll discuss that next week. :)
[17:56] <smoser> do you think ? so you could actually discover that http://cloud-images.ubuntu.com is a http mirror of https://cloud-images.ubuntu.com
[17:58] <utlemming> smoser: I am not sure I follow your line of thought here
[17:58] <smoser> th file data has some 'file_path' right?
[17:58] <utlemming> smoser: yes
[17:59] <smoser> and you have to somehow know that that file_path is to be appended to https://cloud-images.ubuntu.com/
[17:59] <smoser> to actually get a full url
[17:59] <smoser> but in /query, should we have data that indicates mirrors
[18:00] <smoser> authoritative_mirror: https://cloud-images.ubuntu.com/
[18:00] <utlemming> smoser: ah, that makes sense
[18:00] <smoser> mirrors: http://cloud-images.ubuntu.com, http://cloud-images-data.s3.amazonaws.com
[18:00] <smoser> something like that.
[18:01] <utlemming> ok, I follow now
[18:11] <hallyn> stgraber: all right, new http://people.canonical.com/~serge/lxc.debdiff appears to be working perfectly  (shoulda done this ages ago)
[18:14] <hallyn> arguing with myself over whether to push this, or wait until i fold in another change
[18:18] <stgraber> hallyn: looks good
[18:18] <hallyn> man, all these snprintfs in lxc start should be checked for overrun...
[18:19] <roaksoax> smoser: ping
[18:20] <smoser> here roaksoax
[18:20] <roaksoax> smoser: bug #930233 so... while I don't have any argument against using squid-deb-proxy, back when we decided to go with squid directly was because using default s-d-p created lots of archive errors so we decided to go with our own custom squid confug
[18:21] <roaksoax> smoser: the squid config is based on the config used internally IIRC
[18:21] <smoser> that was commonly believed
[18:21] <smoser> but i dont think true
[18:22] <smoser> anyway... it makes all the sense in the world to get it done right in one place.
[18:22] <roaksoax> smoser: i agree with you
[18:22] <roaksoax> smoser: either way, s-d-p is just squid3 + custom config + avahi for auto discovery
[18:22] <roaksoax> right?
[18:22] <smoser> right.
[18:22] <smoser> yeah, the avahi is a wart to us
[18:23] <smoser> but...
[18:24] <roaksoax> smoser: i don't have a problem switching completely to stock s-d-p as long as we can ensure that nothing breaks
[18:24] <roaksoax> and we save ourselves the hasle of having to ship a custom squid config
[18:25] <smoser> sounds like a plan to me.
[18:25] <smoser> the one thing i wish squid could do that i don't think it can is to say "this group of base urls are all the same"
[18:26] <smoser> i think apt-cacher-ng can do that.
[18:26] <smoser> ie, http://us.archive.ubuntu.com/ubuntu and http://archive.ubuntu.com/ubuntu are the same
[18:26] <smoser> but squid does not know that.
[18:27] <roaksoax> smoser: right, but isn't it that archive.u.c is in the UK while us.a.u.c is actually a mirror in the US?
[18:28] <smoser> that would make sense, wouldnt it
[18:28] <smoser> :)
[18:28] <smoser> they're the same.
[18:28] <smoser> but the knowledge of "just act like these are the same" is beneficial
[18:28] <roaksoax> interesting
[18:28] <smoser> as if you have one machine that uses archive.ubuntu.com and most use us.archive.ubuntu.com
[18:28] <smoser> you get no benefit on the one machine
[18:29] <smoser> and duplicate data all throughout your squid cache
[18:29] <smoser> (even if they weren't *actually* round robin mirrors anyway)
[18:33] <roaksoax> smoser: yeah
[18:39] <hallyn> stgraber: doing the lxc console and tty move during lxc startup doesn't work because /dev gets remounted by the container itself...
[18:39] <hallyn> stgraber: i think we might just have to leave that bug open until next cycle and use it to drive device ns
[18:41] <stgraber> hallyn: hmm, no, /dev is mounted by lxc and left alone by mountall
[18:43] <mgw> is there a way to keep the orchestra-server installer from doing an immediate import of various ISOs upon install?
[18:46] <hallyn> stgraber: well hmm.
[18:47] <hallyn> d'oh.  lemme try again
[19:20] <undecim> How do I automatically start my RAID before partitions are mounted?
[19:23] <WarDekar> hey I'm running 10.04LTS with command-line only and am having trouble figuring out how to connect to a PPTP server and route all network traffic through it, does anyone have any good links on how to set it up? there are so many on google and every one i've tried hasn't worked
[19:33] <kirkland> smoser: https://bugs.launchpad.net/ubuntu/+source/orchestra/+bug/930233 :-)
[19:33] <hallyn> stgraber: all right, well i think i have all the devices set up right, but either getty or login appears to be unhappy with 'lxc/console'
[19:33] <kirkland> smoser: that's where we started :-P
[19:33] <hallyn> stgraber: d'oh
[19:33] <hallyn> i'm a loser today.  i know the problem.
[19:34] <smoser> kirkland, yeah. i know.
[19:34] <smoser> i sent some flame bait to an internal list, hoping to get some people to tell me "you're doing it all wrong"
[19:34] <smoser> and what i got back was "really... you should be using squid-deb-proxy"
[19:34] <hallyn> stgraber: id on't have any speed issues like you had.  works fine for me.  only issue is, lxc/* needs to be added to securetty, not sure where we want to do that
[19:35] <smoser> it makes sense to have one thing that is done right.
[19:35] <hallyn> lemme wrap this up and put a bow on it and hand it over to you
[19:35] <kirkland> smoser: fair enough, i *so* wanted squid-deb-proxy to "just work"
[19:35] <smoser> smb, you still around ?
[19:35] <kirkland> smoser: but it "just didn't"
[19:35] <kirkland> smoser: so hopefully you'll have more success than me there
[19:35] <kirkland> smoser: good luck ;-)
[19:35] <smoser> smb, if so.. bug https://bugs.launchpad.net/ubuntu/+source/linux-ec2/+bug/929941
[19:36] <smoser> kirkland, well, the issue that made me stay away from it was no ppas
[19:36] <smoser> but that has since been easily configurablelly resolved.
[19:36] <smoser> and then i think we *thought* we saw more missmatch issues with it compared to some other solution
[19:36] <stgraber> hallyn: yeah, I'm really not sure what went wrong with the speed configuration here ... anyway, if it works the way you did it, yeah!
[19:36] <smoser> but i doubt we had hard data on that.
[19:37] <undecim> can someone tell me wtf is going on with this /proc/mdstat? http://pastebin.com/6rJ4DUFr
[19:37] <kirkland> smoser: I think we also needed support for s-d-p in the server installer itself
[19:38] <kirkland> smoser: whereas it was pretty trivial to just set an http proxy in d-i
[19:38] <kirkland> smoser: but that may have been solved since then too
[19:38] <smoser> s-d-p is just a squid proxy
[19:39] <hallyn> stgraber: http://people.canonical.com/~serge/lxc2.debdiff   that works (on top of http://people.canonical.com/~serge/lxc.debdiff for upstart),
[19:39] <hallyn> stgraber: but like i say there's the issue of /etc/securetty
[19:41] <hallyn> stgraber: (if we decide to go through with it we can squash them into one commit of course)
[19:41] <hallyn> while you stew on that i'll go see about the kernel patch
[19:44] <hallyn> although, maybe we're better off just using an apparmor mount rule to say "you can't mount --remount /"
[19:48] <stgraber> patch looks good, for securetty it's indeed problematic especially as our default user is root ...
[19:51] <stgraber> for 12.04, it's easy, we can simply add them to securetty in the login package
[19:52] <stgraber> as usual previous releases and other distros are the problem
[19:52] <stgraber> can we make the feature optional requiring a flag in the container config?
[19:52] <stgraber> then we could have it off by default in lxc itself and turn it on for lxc-ubuntu for >= precise
[19:56] <hallyn> or, we could take the opportunity to add 'ubuntu:ubuntu' everywhere and recommending doing 'su root' with password root
[19:56] <hallyn> that would sync with the ubuntu-cloud template
[20:00] <stgraber> hallyn: I guess we should do that on top of the rest, people upgrading their machine to Precise won't understand why they can't enter their old containers and some people will also wonder why they can't login as root after setting a password to root
[20:05] <hallyn> all right - i just hate adding cruft into the config namespace if we dont' have to
[20:07] <stgraber> agreed though this one seems worth it as it's a high potential for regression on upgrades not to mention being confusing to users
[20:07] <hallyn> lag.  hate lag.
[20:08] <stgraber> hallyn: can you confirm that adding lxc/console lxc/tty1 lxc/tty2 lxc/tty3 and lxc/tty4 to /etc/securetty works for you?
[20:08] <hallyn> stgraber: ok.  it'll make the patch a bit more involved.  do you think we should push the upstart patch now?
[20:08] <stgraber> if so, I'll upload login
[20:08] <stgraber> yeah, I guess you can upload the upstart change now, that'll give more time for people to complain about potential regressions :)
[20:09] <hallyn> upload login?
[20:09] <hallyn> oh
[20:09] <hallyn> stgraber: yes, it works for me.
[20:11] <stgraber> ok, will upload login in a few minutes then
[20:15] <mgw> so… is it possible to use in-target late command to run a puppet manifest that installs additional packages?
[20:20] <hallyn> hm, kernel patch may have been even simpler than i'd hoped.  (but let's see if it compiles :)
[20:25] <smoser> utlemming, you have thoughts on boto 2.2.1 ?
[20:26] <Daviey> smoser: have you smoked it?
[20:26] <smoser> for 10  minutes.
[20:26] <smoser> :)
[20:27] <smoser> during development of https://review.openstack.org/#change,4025
[20:27] <Daviey> development?
[20:28] <Daviey> i don't follow, that isn't your code?
[20:28] <smoser> well, when i was doing that, i was using new boto.
[20:28] <smoser> thats not my cde ?
[20:28] <smoser> it has my name on it.
[20:28] <Daviey> ah, sorry - i opened the wrong tab
[20:28] <smoser> phew!
[20:28] <Daviey> ETOOMANY gerrits
[20:28] <smoser> i was concerned i'd forgotten my name again
[20:29] <smoser> new boto has that hot new function everyone is talking about
[20:29] <smoser> https://github.com/boto/boto/pull/432
[20:30] <Daviey> that is a must have
[20:32] <mgw> is there any reason an in-target script that installs orchestra-server would hang indefinitely? I finally got the install to continue by logging in on another console and killing the dpkg command
[20:34] <smoser> mgw, rabbitmq?
[20:34] <mgw> smoser: rabbitmq?
[20:34] <smoser> i sweare i saw a hang like that the other day, but dont remember when
[20:34] <smoser> oh.. nah it wasnt in install, it was thorugh devstack.
[20:35] <smoser> mgw, the dpkg can take a logn time.
[20:35] <smoser> but it should not hang permenently.
[20:35] <hallyn> stgraber: do you think we should make it "lxc.ttydir = lxc" (more flexible) or just "lxc.devlxctty = 1" (0 by default)?
[20:35] <mgw> there was zero outputto /var/log/dpkg.log for 30 min
[20:35] <mgw> and nothing to syslog either
[20:35] <jdstrand> hallyn: hey, if I am using juju, am I supposed to expect that the container is going to come up right if lxc-start is confined with apparmor?
[20:35] <mgw> after doing a chroot
[20:36] <mgw> to the installed system
[20:36] <hallyn> jdstrand: on precise?  yes
[20:36] <hallyn> well,
[20:36] <mgw> smoser: I'll paste the dpkg command
[20:36] <hallyn> probably, though i don't know what weirdness juju introduces i suppose
[20:36] <jdstrand> let me try to disable the apparmor profile to see if that helps
[20:36] <hallyn> but we're nto doing any pathname-based restrictions right now
[20:37] <stgraber> hallyn: we can go with the flexible option (assuming not setting it will just default to nothing => /dev)
[20:37] <jdstrand> hallyn: are there any limitations with using lxc inside of kvm?
[20:37] <mgw> smoser: https://raw.github.com/gist/4d81aa0b55fe2f01ac49/43f981b1594ef3fbd0ea5a9ffe625959988c6d0e/gistfile1.txt
[20:37] <mgw> everything I checked had already been installed
[20:37] <hallyn> jdstrand: none
[20:37] <smoser> mgw, do you have a client that pastebins' to gists?
[20:37] <stgraber> hallyn: shadow uploaded for the securetty change
[20:37] <mgw> no, why?
[20:37] <smoser> just curios
[20:37] <smoser> because i've wanted one
[20:38] <smoser> :)
[20:38] <mgw> smoser: wget ?
[20:38] <mgw> or curl
[20:38] <mgw> ;-)
[20:38] <mgw> 2012-02-10 11:37:04 status half-configured ubuntu-orchestra-provisioning-server 2.21-0ubuntu2.1
[20:38] <mgw> that was the last line of dpkg.log
[20:39] <jdstrand> hallyn: should I be using a precise guest or is precise host enough?
[20:40] <mgw> from another server where I did the orchestra install outside of the ubuntu installer, i see there's one more line:
[20:40] <hallyn> jdstrand: older guests are supposed to work
[20:40] <mgw> 2012-02-07 17:16:03 status installed ubuntu-orchestra-provisioning-server 2.21-0ubuntu2.1
[20:40] <mgw> smoser: would it be normal to take 30+ min between those two log entries?
[20:40] <smoser> mgw, i dont know.
[20:41] <smoser> its the kind of thing you only pay attention to when it fails
[20:41] <Daviey> soren: around?
[20:41] <smoser> one thing that could have gone wrong.. there is no libperl5.12 now
[20:42] <smoser> mgw, but also possible tha tthere is a quesiton being asked in debconf.. i'n not sure if that'd cause hang or not.
[20:42] <smoser> roaksoax, ^
[20:42] <Daviey> zul: are we not using UNRELEASED pcoket for openstack staged uploads?
[20:42] <zul> Daviey: yes we are
[20:42] <Daviey> zul: ah, great - thanks
[20:43] <mgw> smoser: I think it would, but this exact puppet manifest I've installed many times from a regular shell and there are no debconf questions
[20:43] <smoser> mgw, it might not in the installer
[20:43] <smoser> but if it is, then we should fix that.
[20:43] <smoser> DEBIAN_FRONTEND=non-interactive and force conf old
[20:44] <smoser> mgw, you did this in late_command ?
[20:44] <smoser> via cobbler?
[20:44] <smoser> i'm being prompted for nagios3-cgi
[20:45] <mgw> DEBIAN_FRONTEND=non-interactive and force conf old
[20:45] <mgw> What does that do?
[20:46] <smoser> makes debconf not ask you annoying questions
[20:46] <mgw> does that need to be done in-target/
[20:46] <mgw> ?
[20:47] <smoser> http://paste.ubuntu.com/837052/
[20:48] <smoser> mgw, i'm not sure actually. i know at one point i pulled haair out
[20:48] <smoser> and fought to get the in-target environment set up well.
[20:48] <mgw> one sec, posting my late command
[20:49] <soren> Daviey: Um... Ok.
[20:49] <soren> Daviey: 'sup?
[20:49] <mgw> smoser: https://gist.github.com/088aab0ed8bd3ee8dc75
[20:49] <Daviey> soren: Did you say you were unhappy with using UNRELEASED pcoket in d/changelog?
[20:50] <mgw> maybe I should add what you pasted in my in-target subshell
[20:53] <smoser> mgw, well, what i suspect has happened is that some process has stdin or stdout open
[20:53] <smoser> and by doing that, the chroot (via in-target) wont exit
[20:53] <mgw> hmm
[20:54] <mgw> so something puppet is installing perhaps
[20:54] <smoser> mgw, i asked specifically about rabbitmq because i'd seen that there
[20:54] <smoser> bug 878600
[20:55] <jdstrand> stgraber: is lxc-ls supposed to require sudo?
[20:55] <stgraber> jdstrand: no
[20:55] <mgw> smoser: thanks; any suggestion (once I get back to that point in the install process) how to determine what, if anything, has stdin/out open?
[20:55] <jdstrand> stgraber: /usr/bin/lxc-ls: line 35: cd: /sys/fs/cgroup/cpuset///lxc: Permission denied
[20:56] <jdstrand> that is indeed root owned and not world readable
[20:56] <guntbert> are there known issues when installing in a XEN PV vps? I keep getting dropped to busybox with the error message "ALERT! /dev/disk/by-uuid/....... does not exist. In fact /dev/disk-by-uuid   doesn't exist itself..
[20:56] <stgraber> jdstrand: hmm, it works here ... (no permission denied)
[20:56] <stgraber> hallyn: ^
[20:56] <guntbert> after the reboot after installing, that is...
[20:57] <stgraber> though lxc-list seems to require sudo for some reason (it doesn't crash but it doesn't show the running containers)
[20:57] <hallyn> smoser has seen this before
[20:57] <hallyn> stgraber: it's supposed to show running containers without sudo
[20:57] <hallyn> jdstrand: do you have cgroup-bin or cgroup-lite installed?
[20:58] <jdstrand> I do have cgroups-lite installed
[20:58] <smoser> mgw, you can swear a lot.
[20:58] <smoser> that might help
[20:58] <smoser> :)
[20:58] <smoser> mgw, i think it was just plain bisecting stuff that got me to finding rabbitmq.
[20:58] <hallyn> jdstrand: uptodate precise for sure?
[20:58] <smoser> mgw, its probably a daemon process running (and see what is running with 'ps' the list is likely not that long)
[20:59] <jdstrand> hallyn: the host might be out of date. let me try that
[20:59] <smoser> hallyn, i tihnk i fixed this in a branch somewhere.
[20:59] <hallyn> smoser: fixed it in what?
[20:59] <smoser> ah. i fixed the  lxc-ls bug
[20:59] <smoser> but you nack'd it.
[20:59] <jdstrand> hallyn: cgroup-lite is what I want though?
[20:59] <hallyn> yeah
[21:00] <stgraber> stgraber@castiana:~$ lxc-info -n qatracker01
[21:00] <stgraber> lxc-info: failed to get state for 'qatracker01': Permission denied
[21:00] <smoser> but i  must have not really submitted it.
[21:00] <stgraber> hallyn: ^
[21:00] <stgraber> hallyn: that's why it doesn't show up in lxc-list for me
[21:00] <hallyn> oh, lxc-info
[21:00] <smoser> jdstrand, do you have btrfs ?
[21:00] <jdstrand> smoser: I like my data :P
[21:00] <jdstrand> smoser: that means 'no' :)
[21:01] <hallyn> stgraber:lxc-ls doesn't use lxc-info
[21:01] <stgraber> hallyn: indeed and lxc-ls works for me
[21:01] <smoser> jdstrand, well, good, then you are hitting it outside of my code, that i thought might be cuasing it.
[21:01] <stgraber> hallyn: I said above that I couldn't reproduce jdstrand's issue with lxc-ls but that I seem to have an issue with lxc-list not showing running containers
[21:02] <stgraber> hallyn: I guess we'd need to make lxc-list either not use lxc-info or have it require root
[21:03] <hallyn> i feel a certain resentment having lxc-list dumped on me
[21:03] <hallyn> but more to the point, i can't find where lxc-info comes from
[21:03] <soren> Daviey: Yes.
[21:03] <hallyn> oh there
[21:03] <soren> Daviey: I mean.. it's fine (and expected) for work-in-progress stuff.
[21:06] <hallyn> stgraber: yeah, bc lxc-info require perms to get state appears to be a feature (in src/lxc/state.c)
[21:06] <Daviey> soren: Did you encounter one that got uploaded?
[21:08] <soren> Daviey: Perhaps I'm getting thrown off by the fact that ~openstack-ubuntu-packagers/nova/ubuntu isn't the packaging branch you're using.
[21:08] <soren> Daviey: ..beucase I think that's where I saw it.
[21:09] <hallyn> jdstrand: i think your lxc would have to be older than 0.7.5-3ubuntu6 to have the error you're getting.  or older than 0.7.5-3ubuntu8 to get something similar
[21:09] <hallyn> else the bug perlexes me
[21:09] <hallyn> perplexes even
[21:09] <smoser> hallyn, bah. you're just pretending to not see things.
[21:09] <jdstrand> well, I've blown away the vm and trying another
[21:09] <smoser> i reported with something newer than that.
[21:10] <hallyn> smoser: wer eyou also using juju?
[21:10] <smoser> juju just uses lxc
[21:11] <hallyn> huh?
[21:11] <mgw> smoser: thanks for your help, i'll let you know if i figure anything out
[21:11] <smoser> mgw, fwiw, i just tried http://paste.ubuntu.com/837080/
[21:11] <smoser> on a fresh instance
[21:11] <smoser> (or 'to' a fresh instance)
[21:11] <soren> Daviey: I'm beginning to sound like a broken record... but where is your packaging branch again?
[21:11] <smoser> and it came back
[21:12]  * jdstrand is downloading 0.7.5-3ubuntu22 now
[21:12] <smoser> so, mgw what i'm saying is I didn't see any issue with one of those packages launching a daemon incorrectly.
[21:12] <mgw> ok
[21:12] <mgw> smoser: at the moment I'm running my same script but with puppet installing fewer packages
[21:13] <smoser> mgw, oh.
[21:13] <smoser> one thing you could do
[21:13] <smoser> is just close stdin and stdout  and stderr
[21:13] <smoser> then they *couldn't* be left open
[21:13] <smoser> (ie, close them before running anything in shell)
[21:13] <mgw> right… how do you do that?
[21:14] <soren> Ah, yes.
[21:14] <mgw> (or rather, How do I do that0
[21:14] <mgw> )
[21:14] <soren> Daviey: found it.
[21:14] <smoser> mgw, sh -c ' exec 0<&- 1>&- 2>&- ;'
[21:14] <smoser> that closes stdin out and err
[21:15] <mgw> smoser: thanks
[21:15] <smoser> for you you might just want to close stdin and redirect stdout and err to a file even
[21:15] <smoser> but one inside the chroot
[21:15] <mgw> yeah —ll just add that to the beginning of my bash -c line in late command
[21:15] <soren> Daviey: What's the logic behind naming it "essex" rather than "precise"?
[21:16] <smoser> sh -c ' exec 0<&- 1>/root/in-target.stdout 2>/root/in-target.stderr ; .... '
[21:16] <smoser> mgw, ^
[21:17] <smoser> mgw, but if a package/daemon is misbehaving that will just hide it
[21:17] <hallyn> jdstrand: say, can you pastebin output of /proc/mounts and ls -l /sys/fs/cgroup and ls -l /sys/fs/cgroup/cpuset?
[21:17] <smoser> and i'd like to fix it.
[21:18] <hallyn> jdstrand: (i couldn't reproduce here with new oneiric container, but of course i'd have to start with fresh cgroup mounts to recreate bad mkdir of /sys/fs/cgroup/cpuset/lxc)
[21:20] <jdstrand> hallyn: ok, I got my precise vm up to date, rebooted, then install lxc etal and now it seems to be working
[21:20] <jdstrand> $ lxc-ls
[21:20] <jdstrand> jamie-local-0-template
[21:20] <jdstrand> I had a kernel update in there-- maybe it was that
[21:20] <hallyn> jdstrand: ok if it happens again with uptodate precise please open a bug.
[21:20] <hallyn> possible
[21:20] <jdstrand> hallyn: thanks and thanks for trying to reproduce
[21:21] <hallyn> actually apparmor *coudl* be it with an olde rkernel, though not for cgroup accesses...  <shakes head>
[21:21] <jdstrand> I removed the lxc-start profile
[21:21] <jdstrand> *shrug*
[21:25] <jdstrand> hallyn: meh, now I see it :(
[21:25] <hallyn> without even a reboot?
[21:25] <jdstrand> $ lxc-ls
[21:25] <jdstrand> jamie-local-0-template	jamie-local-wordpress-0
[21:25] <jdstrand> /usr/bin/lxc-ls: line 35: cd: /sys/fs/cgroup/cpuset///lxc: Permission denied
[21:26] <jdstrand> notice, before I could see jamie-local-0-template fine
[21:26] <jdstrand> I tried to start jamie-local-wordpress-0 via juju, and got this error
[21:26] <jdstrand> /sys/fs/cgroup/cpuset///lxc is 700
[21:27] <hallyn> ok juju must mess with it somehow
[21:27] <hallyn> especially if it worked before, then you ran juju, and it change the perms
[21:27] <jdstrand> I tend to agree
[21:39] <hallyn> stgraber: http://people.canonical.com/~serge/lxc4.debdiff
[21:39] <hallyn> stgraber: with that, if you add 'lxc.devttydir = lxc', then console + tty{1-4} get set up in /dev/lxc/
[21:40] <hallyn> stgraber: (i could'nt just to 'lxc.ttydir' bc of shortxcomings in the confile parsing)
[21:40] <jdstrand> hallyn: if I destroy the service (the lxc instance), lxc-ls works again, but the directory is still 700
[21:40] <hallyn> jdstrand: yes, lxc doesn't chmod it, it just specifies a good mode when it mkdirs
[21:41] <hallyn> jdstrand: you can just chmod it by hand
[21:41] <hallyn> chmod 755 /sys/fs/cgroup/*/lxc
[21:41] <jdstrand> hallyn: well, I'm saying I get back to seeing jamie-local-0-template
[21:41] <jdstrand> I think it is a different code path that is being taking
[21:41] <hallyn> all in lxc-ls?
[21:42] <jdstrand> hallyn: yes in lxc-ls
[21:42] <stgraber> hallyn: cool!
[21:42] <hallyn> stgraber: do you midn testing that debdiff - and maybe adding a 'adduser + chroot' for ubuntu:ubuntu in the lxc-ubuntu.in template, if you're ok with the patch?
[21:43] <hallyn> jdstrand: can you open a bug against juju+lxc with very precise reproduction instructions?
[21:43] <stgraber> hallyn: yeah, I can test it and do the template changes
[21:46] <hallyn> stgraber: thanks!
[21:47] <mgw> smoser: on a different (related) subject — do you know if there is a debconf option to *not* run orchestra-import-isos on install?
[21:47] <mgw> actually, nm
[21:47] <mgw> bug 892328
[21:47] <smoser> mgw, yes. ther eis.
[21:47] <smoser> yeah.
[21:47] <stgraber> hallyn: does setting "lxc.ttydir = " mean the same thing as not setting it?
[21:48] <mgw> smoser: but that's not in a stable release yet :-(
[21:48] <stgraber> *devttydir
[21:48] <hallyn> stgraber: heh, i didn't test that.  ideally yea
[21:48] <smoser> stable, what fun is that.
[21:48] <mgw> hey, i'm pushing things going with non LTS
[21:49] <mgw> We might be bleeding edge company, but we still have to keep our systems sta ble
[21:49] <mgw> btw, my late command ran fine without the orchestra install — now i've added the redirects and am trying again with orchestra
[21:51] <jdstrand> hallyn: bug 930430
[21:51] <hallyn> jdstrand: thanks
[21:52] <Daviey> zul: why do we build-dep on python-all-dev ?
[21:52] <Daviey> (for nova)
[22:02] <mgw> smoser: awesome, it worked!
[22:02] <mgw> would it make sense to patch in-target to automatically do the redirect?
[22:06] <mgw> actually it didn't work
[22:09] <mgw> smoser: it caused everything after it to simply not run
[22:09] <mgw> (after the redirect)
[22:10] <smoser> mgw, i suspect you had some syntax error.
[22:10] <smoser> but i dont know where it would have gotten placed.
[22:10] <smoser> (ie, where youd' see the output of the error.
[22:10] <smoser> i have to run..
[22:10] <smoser> sorry
[22:10] <mgw> smoser: ok
[22:10] <mgw> ty
[22:11] <smoser> later.
[22:11] <stgraber> hallyn: running test build now, did quite a few more changes in the template
[22:12] <stgraber> hallyn: for some reason we were modifying /etc/dhcp/dhclient.conf to basically set the same value "<hostname>" in it means, use whatever is the current hostname
[22:12] <stgraber> hallyn: I also made the code removing tty[56] to be < Precise
[22:13] <stgraber> hallyn: and moved another udev config change to also < Precise (this one might need fixing in udev if it still affects us)
[22:14] <hallyn> stgraber: sounds good.  now are you also adding the to the template to add 'lxc.devttydir = lxc' for precise+ contaienrs?  cause i didn't...  or should we leave that up to the user for now?
[22:15] <stgraber> hallyn: yep, I'm also setting devttydir for >= Precise
[22:16] <hallyn> and adding ubuntu:ubuntu for all releases?  or did you decide against that?
[22:16] <hallyn> (it sort of seems like unnecessary tweaking, but...)
[22:17] <stgraber> ubuntu:ubuntu for all releases, guessing what's the sudo group based on the release
[22:23] <hallyn> stgraber: oh, ok - i was figuring we'd count on su and ignore sudo, but that works
[22:23] <hallyn> stgraber: thanks a lot.  ttyl
[22:52] <hayer> How can I give the www-data user access to the program hddtemp?
[23:18] <stgraber> hallyn: seems to work fine, uploading to my ppa now
[23:23] <hallyn> cool
[23:23] <hallyn> whats ff date again?  still need to get the ureadahead and friends ... :(
[23:23] <stgraber> hallyn: 16th
[23:24] <hallyn> phew
[23:55] <stgraber> hallyn: lxc built and published in my experimental ppa if you want to give it a try