[00:16] <glosoli> oh I reinstalled server one thing I releasied
[00:16] <glosoli> what can be wrong with not getting auto complete ?
[00:16] <ewook> glosoli: ie no tab completion with bash?
[00:17] <glosoli> yes
[00:17] <ewook> type bash
[00:17] <ewook> and try again
[00:17] <ewook> if it doesn't work, somehow your default bash-settings doesn't contain auto completion.
[00:18] <glosoli> it kinda works, but won't autocomplete for words like apt-get install for install exactly
[00:18] <glosoli> probably because of root ?
[00:19] <ewook> install shouldn't be covered by completion. it's part of input to the command, not a command.
[00:19] <ewook> I might be wrong.
[00:21] <ewook> I am wrong
[00:22] <ewook> that's plain weird.
[00:22] <ewook> (not being wrong, but the completion not working)
[00:23] <glosoli> ewook: hah ok
[01:59] <edgy> hi, how can I tell whether the installed OS is ubuntu server or ubuntu desktop?
[02:00] <glosoli> dpkg -l | grep gnome
[02:00] <glosoli> :)
[02:03] <edgy> glosoli: but may be whoever installed it, didn't install gnome or just removed it, can't I tell from the kernel or something?
[02:05] <glosoli> edgy: sorry might be not :)
[02:06] <glosoli> not sure
[02:06] <edgy> ok let me explain my problem
[02:06] <edgy> I noticed in my desktop when a kernel update is available, a file /var/run/reboot... is created, now in my server I don't get that file so I don't know when to reboot, I thought may be the server does it in a different way
[02:10] <glosoli> wha can be the problem that apache won't start after reboot
[02:15] <qman__> edgy, if you installed a kernel update, you need a reboot
[02:15] <qman__> by default, a simple apt-get upgrade will not update the kernel
[02:16] <edgy> qman__: there unattended-upgrades that can boot and i am going to try it
[03:58] <redactd> fyi, the problem i was experiencing yesterday with signed mail not being sent and the tls errors etc was due to our cisco firewall. changed this to be less restrictive and mail is flowing smoothly now :_)
[04:03] <patdk-lap> oh, the evil dns and smtp *fixup*
[04:30] <genii-around> Holy crap bug-bot!
[06:46] <jibel> jamespage, do you mind if I kill the job natty-server-ec2-daily ? it's been running for 22h
[06:55] <jamespage> jibel, not at all
[06:55] <jibel> ta
[09:05] <jibel> jamespage, there is a problem with automated tests of the latest server images. I filed bug 1028301
[09:05] <jibel> manual installation works fine
[10:15] <jamespage> jibel, thanks for letting me know
[10:16] <jamespage> jibel, I normally do a 'call for testing' email to ubuntu-server around about now - are the server ISO images in a good state for testing?
[10:16] <jamespage> manual testing that is....
[10:18] <jibel> jamespage, I did a couple of manual installations on amd64 (default install, LVM) and excepted this bug the new squashfs install seems to be working
[10:19] <jamespage> jibel, is there a good source of information about the squashfs install that I could reference - it will be of interest to people testing.
[10:21] <jibel> jamespage, the best place would be thetech overview but there is nothing there. Daviey would know
[10:21] <jamespage> Daviey, ^^  info re squashfs install for call for testing alpha3?
[10:22] <Daviey> I will add something to that
[10:23] <Daviey> But really, people shouldn't need to notice the difference
[10:23] <Daviey> other than slightly faster install
[10:23] <Daviey> One think roaksoax will be best tasked to do i think, is provide a preseed for netinstall that uses squashes
[10:26] <jamespage> Daviey, do the existing automated test cases which use preseeding need to be updated then?
[10:40] <Daviey> jamespage: Should be unimpacted
[10:42] <Daviey> jamespage: but seems not
[10:43] <Daviey> jibel / jamespage: do you have capacity to look into failing jenkins?
[10:43] <jamespage> Daviey, jibel already raised a flag about that - see backlog
[10:44] <jibel> Daviey, [11:05] <jibel> jamespage, there is a problem with automated tests of the latest server images. I filed bug 1028301
[10:45] <jamespage> 'Copying data to disk' - nice
[10:50] <jamespage> jibel, are we waiting on a re-spin to resolve that then?
[10:52] <koolhead17> hello all
[10:53] <jamespage> hey koolhead17
[10:53] <jibel> jamespage, we are
[10:53]  * koolhead17 bows to jamespage 
[10:53] <koolhead17> Daviey: i need my cookies
[10:54] <jibel> right after next publisher run which should happen in less than 15 minutes
[10:54] <koolhead17> jamespage: howdy?  that hadoop charm was for presentation by mark i guess at oscon
[10:55] <jamespage> koolhead17, I think m_3 used the stock hadoop charm from the charm store - not sure would have to check
[10:55] <jamespage> its the one we used for scale testing during UDS Q
[10:56] <koolhead17> jamespage: well juju is awesome
[10:56] <koolhead17> :)
[10:57] <jamespage> jibel, I'll do the call for testing once thats working again
[11:03] <koolhead17> BTW i missed all folks during OSCON
[11:47] <glosoli> how to  make directory writable by webserver ?
[11:47] <Jeeves_> chown www-data:www-data ${dir}
[11:51] <glosoli> Jeeves_: thanks
[11:51] <Jeeves_> Please not that that's not very smart, security wise.
[11:52] <Jeeves_> But there are no other options
[11:53] <glosoli> Jeeves_: I understand, it's likely to happen when using something build up on php
[11:53] <glosoli> :)
[12:06] <zul> good morning
[12:24] <glosoli> As far as I understand Devcot is the most proper way of getting IMAP running on Ubuntu Server
[12:24] <glosoli>  ?
[12:25] <mardraum> dovecot? it's fine software, but there are a lot of mature imap solutions out there.
[12:26] <glosoli> mardraum: can you name some ?
[12:26] <mardraum> nope, use google
[12:27] <glosoli> mardraum: I did already, checking by priority in Google Search results, Dovecot seems to be the most widely used one
[12:27] <mardraum> sure, if that's important to you, use it?
[12:27] <glosoli> ah ok..
[12:33] <pmatulis> glosoli: dovecot is well liked in the ubuntu community, as is postfix
[12:34] <glosoli> pmatulis: ok, I like things which are widely used even though sometimes they are not the most proper ones, but it helps getting support more quickly than trying to build  bike from group up
[12:35] <glosoli> from ground up"
[12:35] <pmatulis> glosoli: dovecot is a perfectly proper IMAP server and postfix is a perfectly proper MTA
[12:35] <glosoli> I am not sure what's MTA
[12:36] <Pici> !mta
[12:39] <glosoli> Pici: thanks, I always forget there are bots who can answer with proper commands typed
[12:39] <jibel_> jamespage, quantal server builds are back to normal
[12:45] <zul> jamespage:  ping so i was thinking about the openstack database migration qa stuff..
[12:46] <zul> jamespage: so i was thinking something like the following
[12:47] <zul> jamespage:  compare the database to a previous run of the test, with the current test, if there is changes, restore the old database, update the database and then run the devstack exercises
[12:48] <m_3> koolhead17: yes, the hadoop charm was the one from the charmstore... lp:charms/hadoop
[12:49] <koolhead17> m_3: awesome :)
[12:53] <koolhead17> https://bugs.launchpad.net/openstack-manuals/+bug/1028365
[12:53] <koolhead17> can anyone tell me if its new dep inside 12.04/essex
[12:53] <koolhead17> or more about what "guestmount" is meant for
[12:53] <jamespage> jibel_, \o/
[12:54] <jamespage> right - everyone test now!
[12:54] <jamespage> ...
[12:54] <jamespage> lol
[12:56] <zul> jamespage: but i am testing! :)
[12:56] <zul> koolhead17:  its like nbd
[12:58] <koolhead17> zul: so is it needed as dep with nova ? someone reported sayining nova needed it
[12:58] <koolhead17> i have asked him about his env
[12:58] <zul> koolhead17:  no its another option
[12:59] <koolhead17> https://bugs.launchpad.net/openstack-manuals/+bug/1028365
[12:59] <koolhead17> i have asked him 4 his env
[13:02] <jamespage> zul, lemme just send this email and I'll be with you
[13:06] <zul> jamespage:  no now! :)
[13:09] <jibel_> jamespage, ouch, minimal server install pulled ubuntu-standard
[13:11] <jamespage> jibel_: side effect of the installer changes?
[13:11] <jibel_> likely
[13:11] <jamespage> Daviey, ^^ poke minimal virtual is very broken....
[13:12] <Daviey> jamespage: How so?
[13:12] <Daviey> oh
[13:12] <Daviey> i see
[13:12] <jamespage> wrong kernel, wrong modules, wrong base - in fact all wrong!
[13:13] <Daviey> jibel_: that is something that is reasonable to release note.. i'm not going to bust a gut over that i think
[13:15] <patdk-wk> *release note* minimal installs have been supersized
[13:16] <jamespage> disk space is overrated anyway
[13:17] <patdk-wk> it's more the cron jobs and other stuff though :)
[13:20] <jamespage> if anyone in channel would like to help with testing ISO images for ubuntu quantal alpha 3
[13:20] <jamespage> please see https://lists.ubuntu.com/archives/ubuntu-server/2012-July/006359.html
[13:20] <jamespage> zul: OK - now
[13:20]  * jamespage reads backscroll
[13:20] <zul> jamespage: okies
[13:25] <patdk-wk> getting great speeds right now, no one else download the iso :)
[13:28] <patdk-wk> ok I'm done :)
[13:28] <patdk-wk> that was a fast two iso's
[13:29] <jamespage> zul, that would infer that we care about database migrations on every package update/commit during the dev release?
[13:31] <zul> jamespage: kind of..i think it more infers that we care if the database schema changes than we care because the database schema doesnt change on every packages commit/update
[13:32] <jamespage> zul, OK - I think I get it
[13:32] <jamespage> so we test upgrades *when* a fresh schema is different from the previous tests schema snapshot?
[13:33] <zul> jamespage:  yeah
[13:33] <zul> jamespage:  i think that database schema tests can be done on a kvm thingy
[13:34] <jamespage> zul, how where you thinking of handling creation of the image to then upgrade?
[13:34] <zul> jamespage:  i was thinking of keeping a sql dump in a bzr branch
[13:34] <jamespage> ah - one sec - I see - we just install the data snapshot before we try to upgrade
[13:35] <zul> right
[13:35] <jamespage> install rather
[13:35] <jamespage> zul, it works for me
[13:35] <koolhead17> .0
[13:36] <zul> jamespage: good because i started writing this up yesterday :)
[13:36] <jamespage> zul, \o/
[13:36] <zul> jamespage: just have to do the all in one scripted kvm install
[13:36] <jamespage> I still need to work on that jenkins/auto package build test stuff
[13:38] <zul> i totally forgot about that
[13:40] <koolhead17> zul: so if one runs qemu he needs guestmount i suspect not incase of KVM
[13:40] <zul> koolhead17: no if nbd fails, it will try using guestmount and if libguestfs is not installed then it will just fail
[13:42] <jamespage> zul, yeah - I 'forgot' to
[13:42] <zul> jamespage: heh
[13:42] <koolhead17> ooh. so running openstack in Qemu will need one of them. i just closed making it invalid for doc as we tested it on KVM :(
[13:42] <zul> how convient
[13:42] <jamespage> but I promised I would do it - tracking down test failures in glibc + google-perftools killed my time for a few days....
[15:05] <pmatulis> jamespage: it would be nice if the columns on [1] could be explained.  what do these mean, for the 'amd64 server' row:
[15:05] <pmatulis> 4/18 (1)2/4 (4)1/1 (1)
[15:05] <pmatulis> [1]: http://iso.qa.ubuntu.com/qatracker/milestones/226/builds
[15:06] <jamespage> pmatulis, so 4/18 (1)
[15:06] <jamespage> 4 tests completed
[15:06] <jamespage> out of 18 tests in total
[15:06] <jamespage> with 1 failure
[15:06] <pmatulis> ah ok
[15:07] <pmatulis> and the 3 bugs cover those 6 failures?  ideally?
[15:10] <jamespage> pmatulis, yes - patdk-wk and I overlapped and confirmed each others iscsi test failures
[15:10] <pmatulis> jamespage: i see, good.  btw, i just fired up a normal and preseed install.  so far so good
[15:15] <jamespage> pmatulis, +1 and thanks for helping out! much appreciated
[15:48] <a1fa> has anyone played with vlan trunks to vbox  via ubuntu server?
[16:24] <gitesh> Hi. I just install ubuntu server 12.04LTS Two days back, here.
[16:25] <gitesh> The first thing that I wanted to start Internet , I am trying but I couldn't .
[16:25] <gitesh> Can someone please help me?
[16:26] <zul> just ask the question
[16:27] <gitesh> How do I connect ubuntu server to Internet ?
[16:31] <gitesh> I have wired connection.
[16:33] <RoyK> gitesh: do you have an ip address set? run ifconfig - then check if you have a default gateway - netstat -rn
[16:33] <RoyK> check dns config - cat /etc/resolv.conf
[16:34] <gitesh> Royk. ok, I collect that info. brb.
[16:34] <RoyK> if all looks well, try to resolv something, like "ping bbc.com" (
[16:35] <RoyK> not all hosts answer to ping requests - I tried cnn.com first, it doesn't, bbc.com does
[16:35] <gitesh> I should be master in  server internet configuration. Coz it's first of all thing should be done, i guess.
[16:49] <Threshold460> hello
[16:50] <Threshold460> guys what is the server address so i can connect via my client please?
[16:52] <Thresholder> hello guys i am trying to install ubuntu server and got stuck..anyone could help me out please?
[16:53] <patdk-wk> !ask
[16:57] <gitesh> RoyK, I found nothing.
[16:59] <gitesh> also, for default gateway, netstat -rn : I found an empty table.
[17:00] <Thresholder> ok sorry...i was trying to install vnc in ubuntu server 12.04 but i didn't exit root as per the guide i'm following and now I get - mkdir: cannot create directory `.vnc': File exists - error
[17:00] <gitesh> RoyK, also dng config was an emty file.
[17:00] <Thresholder> This is the guide I am following http://www.torrent-invites.com/seedbox-tutorials/181210-ubuntu-server-w-gui-permenant-auto-connecting-ssh-tunnel-fail2ban-etc.html
[17:01] <RoyK> gitesh: what do you have in /etc/network/interfaces?
[17:02] <gitesh> RoyK, just a minute :)
[17:07] <gitesh> RoyK: about etc/network/interfaces There were two lines : 1>auto lo 2> iface lo inet loopback
[17:08] <RoyK> og
[17:08] <RoyK> oh
[17:09] <RoyK> try adding 'auto eth0' and 'iface eth0 inet static' or with 'inet dhcp' - google for ubuntu network config
[17:09] <RoyK> gitesh: also - check with ifconfig -a that ubuntu actually sees your nic
[17:10] <gitesh> Can someone point me somewhere?, I want to learn. gateway , router , etc/
[17:10] <RoyK> !network
[17:10] <RoyK> erm
[17:10] <RoyK> wrong
[17:10] <RoyK> !interfaces
[17:10] <gitesh> RoyK: Okay, i do auto eth0.
[17:10] <RoyK> gitesh: does ifconfig -a show it
[17:10] <RoyK> ?
[17:11] <rdw200169> gitesh Linux Routers by Tony Mancill, if you can survive that, you'll never need help again
[17:11] <RoyK> gitesh: http://kurl.no/XvO7
[17:12] <gitesh> rdw200169, RoyK. Thanks for your supports.
[17:12] <gitesh> I will have to check ifconfig -a
[17:20] <gitesh> rdw200169: 450 pages of pdf  :)
[17:20] <rdw200169> gitesh aw, c'mon its a good book, Mancill deserves a little cash ;)
[17:22] <RoyK> you don't need that book to make networking work on ubuntu
[17:22] <RoyK> a quick google goes a long way
[17:22] <gitesh> rdw200169: pain, pain. I will read and try directly. need days. Thanks for the link :p
[17:23] <rdw200169> gitesh all things worth doing are generally difficult
[17:24] <gitesh> RoyK, ok. Yet.I want for ubuntu only.brb
[17:34] <gitesh> RoyK: about ifconfig -a , eth0      Link encap:Ethernet  HWaddr 00:15:c5:4a:16:5a
[17:34] <gitesh> Rx and Tx errors
[17:34] <RoyK> well, just set a static ip address on that
[17:35] <RoyK> then test
[17:36] <gitesh> RoyK: ok. How do I get my static IP?
[17:37] <RoyK> see above
[17:37] <RoyK> or google it
[17:37] <RoyK> like I said
[17:38] <RoyK> you won't get a tutorial on irc - you might get some good hints, and you may as well get some kicking if you ask too many stupid questions
[17:38] <gitesh> ok, sorry.
[17:50] <savid> Is there a way to install third-party nginx modules without rebuilding nginx from source?  Anyone know of any PPA's that include optional third-party modules?
[17:57] <RoyK> savid: any module in particular?
[17:58] <savid> RoyK, fancyindex
[17:59] <RoyK> hm... perhaps http://bit.ly/MUMu8c
[18:10] <glosoli> I checked webmin isn't in the repos should I use deb package provided by webmin team or use tar gz ?
[18:13] <RoyK> !webmin
[18:14] <pdtpatrick> Question - anyone having problems with precise using PXE  + preseeding? I have a Dell R410 server with onboard RAID controller. I'm then setting LVM on the mirrored drive via preseed. However, after server finishes and restart, it never boots and gets stuck at initramfs, complaining that it cannot find root filesystem.
[18:14] <glosoli> !panel
[18:15] <glosoli> !webpanel
[18:16] <RoyK> glosoli: hint number one: Learn the commandline - it's not hard, and it will take you a long way, far longer than any web-based admin tool
[18:16] <glosoli> RoyK: I am not sys admin :) and doing this job not because I want to...
[18:16] <RoyK> !ebox
[18:17] <genii-around> !ebox
[18:17] <genii-around> Heh, too slow
[18:17] <RoyK> :)
[18:17] <glosoli> can it work with dovecot mess ?
[18:17] <RoyK> no idea
[18:17] <RoyK> if you want something easy to manage for email, try zimbra
[18:17] <RoyK> not supported officially by ubuntu, but works fine if on a separate machine or vm
[18:18] <glosoli> ah..
[18:18] <RoyK> just don't mix zimbra with an existing ubuntu install with apache and postfix and all - it's not easy...
[18:18] <RoyK> glosoli: but then - if you have an ubuntu server, setting up a vm is a no-brainer
[18:20] <glosoli> RoyK: yeah, well the only thing I am lost at now is dovecot nto being able to find proper documentation of how to create an user, as the project itself of dovecot is bad docummented (I am in risk to be called stupid)
[18:20] <RoyK> glosoli: the only "hard" part, is to setup a bridge to allow for direct routing, but then again, it's not that hard, just a little fiddling with /etc/network/interfaces
[18:21] <RoyK> glosoli: dovecot isn't badly documented, but it's documented for sysadmins, which is the bad part
[18:21] <RoyK> glosoli: do you have a spare ip address on this server?
[18:21] <glosoli> RoyK: ah yes, it your comment makes more sense, and doesn't hurt them too much :)
[18:21] <RoyK> glosoli: all sysadmins have been non-sysadmins at some time ;)
[18:21] <glosoli> RoyK: nah it's container, virtual dedicated server
[18:22] <RoyK> for use with most stuff, or just the email sever?
[18:22] <glosoli> well I am studying Software Engineering, sys admining is not meant to be my job, but some experience is good proba ly
[18:22] <glosoli> RoyK: OpenERP, PHP Webs, Email Server
[18:22] <RoyK> it definetely won't hurt you...
[18:22] <rdw200169> glosoli some?  I work with 'developers' all day every day; i write software too, but the sysadmin part is what makes me indispensable
[18:23] <RoyK> glosoli: ok - setting up a vm on that thing may be a bit hard :P
[18:23] <RoyK> rdw200169++
[18:23] <lamont> what zimbra does to the unpublished internals of postfix should never be done
[18:23]  * RoyK has been a sysadmin for 15+ years
[18:23] <rdw200169> glosoli just look at the median income for a Senior Linux SysAdmin in the NYC market and compare that with a Java Developer
[18:23] <glosoli> rdw200169: well depends on company probably, if serious projects I prefer having serious sys admin who does it for life :)
[18:23] <RoyK> lamont: erm - have they patched it and not released sources?
[18:24] <bac> hallyn: i have a question about https://bugs.launchpad.net/ubuntu/+source/lxc/+bug/924281 when you have a moment.
[18:24] <glosoli> rdw200169: and how does it compare
[18:25] <rdw200169> glosoli my income is … nice.  very nice.  I have recruiters banging down my door frequently.  I have job security.  Java Developers?  They churn those out by the thousands at every university in the country.  Linux?  Nobody teaches that anymore...
[18:26] <glosoli> rdw200169: in my college people has linux lectures to be honest
[18:26] <glosoli> :)
[18:26] <RoyK> rdw200169: nice as in what?
[18:26] <rdw200169> glosoli And when I say 'Linux' I really mean 'Unix' and all that.
[18:27] <lamont> RoyK: ISTR that they just called internal functions in ways that were not nice, and subject to breaking on any point release.
[18:27] <glosoli> in btw, your income is nice as in what ?
[18:27] <lamont> though maybe they stopped doing that
[18:28] <rdw200169> glosoli don't bash the benefits of SystemsAdmin experience; even for a developer its a very very nice bonus
[18:28] <hallyn> bac: ask away.  (running tests elsewhere, will answer soon)
[18:29] <glosoli> rdw200169: I don't :) as I said the experience itself is good, but you hadn't said yet how does it compare to java dev salary
[18:29] <bac> hallyn: the bug is marked fixed release from back around march.  i'm still seeing the exact symptoms and need clarification/advice from you as to if it should be expected to work.
[18:29] <bac> hallyn: this is with a precise host and a precise container
[18:30] <rdw200169> glosoli i leave that to your research; the internets are filled with statistical data per region on income and cost of living adjustments
[18:30] <glosoli> rdw200169: Sys Admin salary median in usa is 72 000  year, for medium developer its 86000 :)
[18:30] <glosoli> talking about median for medium developer ,not for senior
[18:31] <rdw200169> glosoli and you're throwing in the regular Sys Admins.  Windows Admins are a dime a dozen.  Focus on Senior Linux Systems Administrator please
[18:31] <skrite> hey all, i have some quick questions about mysql-cluster. we want to start with three machines, and if one crashes, how difficult is it to set up it's replacement?
[18:32] <rdw200169> skrite: 1) specify a master with the best write performance. 2) the other two servers should be slaves 3) don't even think about master-master replication scenarios
[18:32] <skrite> rdw200169: right, but what about mysql-cluster ?
[18:32] <rdw200169> skrite: ah right
[18:32] <skrite> rdw200169: i found that master-master was going to create headaches and early graves
[18:32] <skrite> :)
[18:33] <rdw200169> skrite: personally i despise MySQL with the passion of a thousand suns; but I digress, Paying Oracle is a mortal sin
[18:33] <glosoli> rdw200169: still Senior Java developer gets more compared to Senior Linux Admin :)
[18:33] <rdw200169> skrite: as such, i have no idea
[18:34] <rdw200169> glosoli: then go be a java developer.
[18:34] <glosoli> rdw200169: meanwhile doesn't differ to much and I understand your point :)
[18:34] <hallyn> bac: it's possible that you need to use the apparmor policy taht supports nesting
[18:34] <hallyn> bac: but, i'd say open up a bug
[18:34] <skrite> rdw200169: yes, i just have 5 years of code that rely on mysql
[18:34] <glosoli> rdw200169: I don't do Java lol.. :)  and I don't live in USA :)
[18:35] <hallyn> bac: oh, yes.  you definately need to use the nesting policy
[18:35] <rdw200169> glosoli one thing to consider is this new thing called 'DevOps', which is like both in one
[18:36] <glosoli> rdw200169: I understand your point, don't count me for these who don't value, I value the experience :) But people here in Lithuania where I live thinks sys admining is some easy fun.. so you got paid little for it :)
[18:38] <bac> hallyn: i'm not sure what that entails.  hint?
[18:39] <rdw200169> skrite: yeah, i know where you're coming from with that.  MySQL in the Lamp days was an epidemic that has survived into a plague for many operations.  Hence, the whole NoSQL movement to handle massive amounts of overly-relational MySQL data to ease the relational burden
[18:40] <gary_poster> bac, googled: http://www.stgraber.org/2012/05/04/lxc-in-ubuntu-12-04-lts/
[18:40] <gary_poster> look for "Container nesting"
[18:40] <skrite> rdw200169: he he. yeah. i have lots of ideas if i were to build this system again.
[18:41] <hallyn> bac: sorry, i'm waiting for my laptop to finish updating to see if the new lxc ships with teh policy you need or not
[18:41] <hallyn> just another min
[18:41] <stgraber> hallyn: quantal does
[18:41] <hallyn> stgraber: precise-updates?
[18:41] <rdw200169> skrite: all things considered, though, since you're using MySQL-Cluster, i've been reading up on the data sheet, and its obvious you get some sort of replicable redundancy, both within the cluster and to another cluster off-site
[18:41] <gary_poster> stgraber, is the one from your blog post the one we should use in lucid?
[18:41] <gary_poster> or should we steal from quantal?
[18:42] <gary_poster> oh
[18:42] <gary_poster> precise-updates would be cool
[18:42] <stgraber> we can't really justify introducing a completely new apparmor policy as an SRU
[18:43] <stgraber> and quantal has a completely different structure for the apparmor profiles
[18:43] <hallyn> drat, yeah, not in precise-updates
[18:44] <stgraber> I'm kind of planning to start maintaining an lxc backport in precise-backports once we have the API changes landed in quantal
[18:44] <hallyn> stgraber: that reminds me, i'm hoping wed night to upload some quantal bugfixes;  if you have anything small you want to queue up in the bzr tree i'll look for that before merging
[18:44] <gary_poster> ah ok stgraber, hallyn.  So stgraber's blog post is the best source?
[18:44] <hallyn> gary_poster: yeah
[18:44] <skrite> rdw200169: mysql and mysql-cluster use the same query language, right? so if there was an ultimate failure, restoring from a mysqldump (or similar) would be possible. This is good. just wondering what goes into recovery because i know that a minimum of three machines are required
[18:44] <gary_poster> cool, thank you hallyn
[18:44] <stgraber> so then you'll have the packages in precise-proposed for the fixes without any new feature, and precise-backports for the new cool stuff
[18:44] <hallyn> stgraber: i suspect most ppl would use backports :)
[18:45] <hallyn> also might be worth considering a store for container policies
[18:45] <stgraber> hallyn: make sure to upload to -proposed (alpha3 freeze)
[18:45] <rdw200169> skrite: one would assume so, but i haven't used it, so i don't know ;)
[18:45] <skrite> rdw200169: i see, ok, thanks
[18:46] <patdk-wk> I used to, but too many times it would break me, so I use my own ppa, that I backport to
[18:46] <hallyn> stgraber: yup
[18:46] <stgraber> hallyn: yeah, I'm expecting most people will want the new cool stuff, though I know at least some want no risk of regression and stability, so having both -updates and -backports should cover that.
[18:46] <patdk-wk> also means I keep up with security updates correctly, as backports don't
[18:46] <hallyn> stgraber's backport will :)
[18:47] <micahg> well, backports is on by default now, but pinned lower, so it's not all or nothing, you can choose which apps you want from -backports
[18:47] <hallyn> stgraber: how will you handle the versioning?
[18:47] <micahg> starting in natty it's pinned lower and oneiric it's on by default
[18:47] <stgraber> yeah, I'd probably be pushing to -backports instead of my PPA once we start doing that. Currently my PPA is usually 30min or so behind what we push to quantal, so we should be fine for any security/bugfix ;)
[18:47] <patdk-wk> I'm still on lucid, as precise .1 isn't out yet, and still have upgrade issues
[18:48] <micahg> stgraber: I think you know enough backporters to push things through quickly if need be :)
[18:49] <stgraber> hallyn: I'd have to check the backport policies again as it's been a while since I last read them, my guess is <quantal version>~12.04.x or something similar (so much higher than precise but lower than quantal)
[18:49] <hallyn> ah yeah
[18:49] <micahg> yep, that's the current versioning scheme
[18:49] <hallyn> of course you'll have to disable seccomp (and such).  assuming i get that enabled
[18:49] <micahg> err...~ubuntu12.04.X
[18:50] <stgraber> hallyn: well, we might be able to backport libseccomp too. Precise's kernel supports seccomp2 right?
[18:50] <hallyn> no
[18:50] <hallyn> or if it does, a different api.
[18:50] <stgraber> ah, that's annoying :)
[18:50] <hallyn> i *think*.  could be wrong
[18:50] <stgraber> is libseccomp clever enough to detect kernels that don't support it and deal with that?
[18:51] <micahg> hrm?  I think it does
[18:51] <hallyn> well it'll at least return -1 on seccomp_init(0.
[18:51] <stgraber> if so, we probably should backport libseccomp too, so that when the quantal kernel is backport to precise (as part of the LTS hardware enablement), we get the feature to work for free
[18:51] <stgraber> *backported
[18:51] <hallyn> then users just can't enable seccomp through the config.  yeah maybe it's no big deal.
[18:52] <hallyn> i'm sick of bisecting, btw
[19:07] <Dulcin> do i have to reboot ufw somehow for ports to become open?
[19:07] <Dulcin> after adding rules*
[19:14] <hallyn> Dulcin: did you add rules with the ufw command?  then no, i don't think so
[19:14] <hallyn> you can confirm with 'iptables -L' which will show the active rules
[19:14] <hallyn> i'm pretty sure ufw is active as soon as you commit
[19:22] <jdstrand> Dulcin: 'sudo ufw enable'
[19:22] <jdstrand> that's it. do that before or after adding commands
[19:22] <jdstrand> err
[19:22] <jdstrand> adding rules
[19:23] <jdstrand> Dulcin: if you edited /etc/ufw/*rules, then you will want to do 'sudo ufw reload'
[19:27] <Dulcin> hallyn jdstrand, apparently it wasn't ufw acting up, but my mysql my.cnf which only allowed localhost connections
[19:27] <Dulcin> but thanks for the tips
[19:28] <jdstrand> np
[19:28] <jdstrand> gotta listen to connect :)
[19:48] <skrite> hey all, still reading up on and trying to decide some things about mysql-cluster. what forum would be good to post some basic questions to? i want to know a couple of things before i get started.
[19:54] <Daviey> adam_g: Happen to have a screenshot of Ubuntu's Horizon handy?
[19:55] <adam_g> Daviey: i do not off hand
[19:55] <adam_g> Daviey: one sec, i can get one
[19:59] <Daviey> adam_g: thanks!
[20:02] <skaet> jamespage - what's the lastest on https://bugs.launchpad.net/ubuntu/+bug/1028458 ?   valid bug or not?
[20:02] <smoser> hallyn, kvm in precise seems to have broken mouse grab?
[20:03] <jamespage> skaet, its a bug but I don't think it should block anything
[20:03] <patdk-wk> skaet, it seems me and jamespage where hitting two different bugs, but thought we where hitting the same :)
[20:03] <jamespage> skaet, its def not critical as I first thought
[20:04] <hallyn> smoser: with the latest update?
[20:04] <skaet> jamespage,  ok.   I'll remove it from the respin consideration list.    Is it worth a release note?
[20:04] <jamespage> skaet, I would say so
[20:04] <skaet> patdk-wk,  do I need to worry about yours?  ;)
[20:04] <jamespage> skaet, the server looks like it failed to boot
[20:04] <cyclicflux> Yooooo whats happening?!?
[20:04] <patdk-wk> skaet, defently :)
[20:04] <patdk-wk> mine just errors to initrd, and won't boot
[20:05] <skaet> patdk-wk,  number?
[20:05] <smoser> hallyn, i don tknow when that would have been
[20:05] <hallyn> smoser: works for me
[20:05] <patdk-wk> skaet, same
[20:05] <hallyn> are you using SDL?
[20:05] <skaet> lol,  misread the earlier comment.  gotcha.
[20:05] <skaet> ok,  release note it is.
[20:06] <patdk-wk> more though, I think it might have more to do with /run not being mounted or something
[20:06] <patdk-wk> based on that line 505 error
[20:06] <smoser> hallyn, you're sure?
[20:06] <patdk-wk> but it does mount, just takes longer than expected
[20:06] <smoser> i just ran kvm from a desktop iso
[20:06] <adam_g> Daviey: is there a specific panel you want?
[20:06] <patdk-wk> atleast my guess
[20:06] <smoser> if i click in the window, it does not capture my mouse at all.
[20:06] <smoser> mouse moves freely inside and out, and 'ctrl-alt-f1' goes to desktop rather than vm
[20:07] <hallyn> smoser: does it grab the mouse, but clicks don't go through?
[20:07] <cyclicflux> I had a quick question, this application menu crap is driving me nuts on the ubuntu11.10, and I want to upgrade to pangolin(I think) 12.04, to get the Unity HUD going. However, in my particular case it wants to remove MySQL, etc... some major components on my development machine.  This is technically not a server install however I over time have installed a number of components via apt. I know that there is an addi
[20:07] <cyclicflux> tional upgrade utility for Ubuntu servers, which I will install, I just was seeing the best way to do it and get a few perspectives on it.
[20:07] <hallyn> smoser: which desktop iso?
[20:07] <smoser> precise
[20:07] <hallyn> depending on the guest graphics driver it certainly can be that it doesn't grab.  suse guests for instance do that
[20:07] <hallyn> precise-what?
[20:08] <smoser> release
[20:08] <hallyn> precise-desktop-amd64.iso?
[20:08] <smoser> yes
[20:08] <cyclicflux> hallyn, I believe the only precise is pangolin
[20:09] <hallyn> cyclicflux: yes but there are multiple isos, different installers
[20:09] <hallyn> smoser: yes, this is not new.  that's just what the graphics driver is doing.  it's 'advanced'.  (at least that's my understanding - and it's definately not new)
[20:09] <smoser> hm..
[20:09] <smoser> ok
[20:10] <hallyn> smoser: hm.  then again,
[20:10] <cyclicflux> hallyn, I just saw that in the prior messages
[20:10] <hallyn> i can't tell if mine is being very slow,
[20:10] <cyclicflux> lol!!
[20:10] <hallyn> smoser: are you able to click on things in the guest at all?
[20:11] <adam_g> Daviey: http://ubuntuone.com/4eWrBIDcedrpo0Q37Pn45v http://ubuntuone.com/1ysNFB5UD7c6rzu7pVFRwl
[20:11] <cyclicflux> smoser, you may try to make a manual xorg.conf, and put in mouse settings for it. But its likely due to an intel-graphics
[20:11] <smoser> hallyn, yes. clicking works fine.
[20:12] <hallyn> smoser: ok  (i realized i gave it 300M ram :)
[20:13] <Daviey> thanks adam_g
[20:13] <cyclicflux> smoser, I read that there are problems with the intel-based graphics cards and ubuntu12.04, as I was reading on conflicts when one upgrades or installs it.  Don't quote me on it, but I believe that a few of the articles stated that if using intel graphics card that it is best to stick with 11.10. I however would not know first hand.
[20:15] <hallyn> smoser: note that so long as i'm careful to click (to grab) and leave the mouse inside the sdl window, ctrl-alt-f1 does go to the guest for me
[20:16] <kyle__> I apt-get installed ubuntu-desktop, and now when I go to remove it, it's not removing all of it automatically.  Is that expected?
[20:18] <smoser> kyle__, yes
[20:18] <smoser> but you may be able to 'apt-get autoremove'
[20:18] <smoser> to get rid of all the rest
[20:19] <kyle__> Doesn't look like it's getting everything, but I'll see what it's done.
[20:21] <kyle__> Wow.  Lots and lots of stuff to remove still.  May be faster to just reinstall.
[20:21] <smoser> ah. yeah, it wouldn't work.
[20:22] <smoser> when you isntall something, its dependencies get installed. then when you remove it, only it gets removed.
[20:22] <kyle__> For a lot of packages it removes them, or apt-get autoremove will remove them. Ah well.
[20:23] <kyle__> I'm trying to get displaylink video to work, I thought maybe the desktop config stuff would do it automatically.
[20:23] <kyle__> Silly me, didn't do a thing.
[20:24] <Daviey> i would have expected it to work aswell
[20:24] <Daviey> IIRC, DisplayPort only works with non-free drivers.. but i could be wrong.
[20:24] <kyle__> Daviey: The idea of hot-plug video cards is not well supported.
[20:24] <Daviey> pass.. not much desktop expertise here.. :)
[20:25] <kyle__> displaylink == USB video adapter.  There are open source drivers, and they're all similar enough to work with that driver.
[20:25] <kyle__> Daviey: I'm making "Fear Of God" machines for my labs.  They pose as informational displays, but occasionally show a live feed of the security cameras
[20:26] <kyle__> It keeps the students from slacking off, and others from vandalising.
[20:26] <Daviey> kyle__: sounds like a nice project
[20:27] <kyle__> It is.  And buying an atom box with 1 head + 1 displaylink adapter was much cheaper than buying the cheapest dual head option.
[20:27] <kyle__> At least through my vendors.
[20:33] <miceiken> hi per is mad
[20:36] <jparker> anyone have a recommendation for installing a light gui?
[20:39] <kyle__> jparker: xfce4 is pretty full featured, and light.  Lighter would be oroborus, fluxbox, blackbox.
[20:39] <jparker> i basically just need to run firefox on it
[20:42] <genii-around> You can even just go with twm and xterm in your .xsession , then manually run firefox from there, etc
[20:45] <jparker> tried the xterm route but was getting errors with "DISPLAY is not set"
[20:46] <jparker> i even enabled xforwarding in sshd_config
[20:46] <genii-around> Ah, so forwarded x
[20:47] <genii-around> jparker: Probably need something like export DISPLAY=host-ip-here:0:0
[20:48] <genii-around> work, afk
[20:48] <trimeta> What's the best way to restart a possibly crashed sftp-server process? I was messing around with things and it seems the processes are zombies now, but sudo killall -9 sftp-server doesn't work.
[20:49] <trimeta> I've tried service ssh restart, but the processes are still there.
[20:52] <kyle__> Lure the processes with brains.
[20:53] <kyle__> trimeta: You probably need to find the process ids, and kill -9 them manually.  killall needs a pretty exact match iirc.
[20:54] <trimeta> kyle__: Nope, sudo kill -9 <PID> has no effect.
[20:54] <genii-around> Can't kill zombified processes
[20:54] <kyle__> Oh yea.
[20:55] <kyle__> humm.
[20:55] <kyle__> genii-around: Poke through /proc/<PID>, see what it's parent process is.
[20:56] <trimeta> Hmm, good idea.
[20:56] <trimeta> ...How do I do that?
[20:56] <trimeta> There's no /proc/<PID>/parent file...
[20:57] <guntbert> trimeta: zombie processes are dead already, does sudo service ssh start help?
[20:57] <trimeta> It doesn't. Although, maybe they're not really zombies? top doesn't say I have any.
[20:57] <trimeta> All I know is they can't be killed...and kill/killall don't return any sort of error message.
[20:59] <trimeta> Also, looks like sshd is the parent process...but I've done "sudo service ssh restart" multiple times.
[21:00] <genii-around> Maybe try both upstart and init.d
[21:01] <trimeta> The parent sshd process seems to persist regardless.
[21:01] <kyle__> trimeta: Look in /proc/<PID>/status, or ppid.
[22:40] <glosoli> hey folks I really came to an end after two hours trying to figure out, I got roundcube running on my ubuntu server, everything seems to be fine dovecot doesn't output any errors in logs, the main problem is I can send mail easily but can't get the mail sent to me it s like inbox not working, tryed to add it as account to some mail clients to check if problem is not roundcube so it isn't , also dovecote says everything fine, any ideas ?