[00:17] <bfreis> Hi, what should I do to install Ubuntu Cloud Infrastructure using the live image on a server? It looks like it is simply broken, it simply cannot boot: it gets stuck on a loop saying that it can't find /dev/sr0, and later on it drops to a BusyBox shell. How do people manage to install it?!
[00:17] <twb> sr0 is the CD drive
[00:17] <twb> Is this during the grub install phase
[00:18] <bfreis> it is after grub, I think (after I select "normal" boot over "fail safe" boot -- the same thing happens with "fail safe" boot)
[00:19] <bfreis> and I'm booting from a USB stick
[00:19] <bfreis> (the same USB stick I used to install Ubuntu Server 11.04 on many other machines)
[00:22] <bfreis> must I use a CD to install it? or is there any workaround?
[00:22] <bfreis> I've searched a lot, there's a bug reported on launchpad related to this, but it is marked as fixed
[00:23] <bfreis> also, on some forums people suggested disabling the floppy drive on the bios, which I did (even if I have no floppy drive), but the problem is still there
[03:36] <slicslak> i seem to recall a metapackage that has compile tools in it, linux headers, etc. -- basically the stuff needed for building source
[03:37] <twb> build-essential
[03:40] <John24Doe> so.... whats the topic?
[03:41] <slicslak> yea, that's it, thx twb
[03:43] <twb> slicslak: also "apt-get build-dep foo", or if you are building a kernel module see dkms or (worse) module-assistant
[03:47] <John24Doe> hellooooo any one therrrrrre?
[03:47] <twb> John24Doe: do you have a real question?
[03:48] <John24Doe> yes i do..
[03:48] <twb> John24Doe: ask it.
[03:48] <John24Doe> i just got into the ubuntu server and its great so far but i have a question about the email  service..
[03:49] <TimR> are you trying to setup a email server john24doe?
[03:50] <John24Doe> ive noticed that there are a bunch of email services.. so my question  is what is the best, simple and secure service out there?
[03:50] <twb> John24Doe: service that does what?
[03:50] <TimR> well there is postfix+mysql+dovecot
[03:50] <TimR> there is iredmail you can use
[03:50] <twb> TimR: you misspelled "LDAP" as "mysql"
[03:50] <TimR> no I dint
[03:50] <TimR> didnt
[03:51] <TimR> there is a setup as mysql for backend managemnt
[03:51] <TimR> managment*
[03:51] <TimR> trust me I have ran postfix+mysql+dovecot before
[03:51] <TimR> intill I switched it all over to iredmail+mysql setup
[03:52] <John24Doe> ..
[03:53] <TimR> it depends how much you know how to run a linux mail server john24doe
[03:54] <twb> John24Doe: we cannot suggest a solution unless you describe more clearly what you are trying to achieve.
[03:54] <TimR> that is true what twb is saying
[03:54] <John24Doe> lol thats the problem...  as i said im new to the linux server world.
[03:54] <John24Doe> sorry
[03:54] <TimR> for a newbie I would use iredmail+mysql
[03:55] <John24Doe> so postfix supports encryption
[03:55] <John24Doe> ?
[03:55] <John24Doe> ok ill check out iremail
[03:55] <TimR> what do you mean encryption like in SSL?
[03:55] <TimR> http://www.iredmail.org/
[03:55] <John24Doe> yeah
[03:55] <funkyHat> email encryption is done at the message level and is not dependent at all on a particular mail server, John24Doe
[03:56] <John24Doe> ok good to.   know sorry for the newb speech
[03:56] <funkyHat> SSL is also possible (and yes postfix and dovecot both support that)
[03:58] <funkyHat> That gives you an encrypted connection from you/other clients to the mail server, which could prevent someone from listening in on the data that goes back and forth at that level, but you can't guarantee that the recipient or sender is using encryption at their end, so the usefulness of that is limited
[03:59] <funkyHat> It's definitely worthwhile if you are likely to want to access your mailbox or send mail from an internet cafe or other public wifi
[04:00] <twb> funkyHat: that rather depends on what attack vectors you're guarding against.
[04:00] <twb> funkyHat: if I ran an internet cafe I would have physical key loggers installed on all the hosts, for example
[04:01] <twb> SSL helps against MITM but not against the endpoint being compromised, which is basically a given for an internet cafe
[04:01] <funkyHat> twb: I was thinking of an internet cafe with wireless and your own laptop, if you're using another computer to access the email you'd have to set up webmail at which point IMAP/SMTP encryption becomes irrelevant
[04:01] <jandrusk> All they have to do is running sslstrip. SSL is in really bad shape these days.
[04:01] <twb> funkyHat: ah ok
[04:01] <twb> jandrusk: that from dsniff or something?
[04:02] <jandrusk> Part of Backtrack, but it works by sending all of the SSL requests through the host running sslstrip.
[04:02] <John24Doe>  hey twb im going with postfix
[04:02] <twb> John24Doe: so presumably you want to run an MTA?
[04:03] <John24Doe> yes
[04:03] <twb> postfix is a good choice and is the recommend MTA by Ubuntu
[04:03] <twb> Read the relevant Ubuntu Server Guide sections if you haven't already
[04:03] <funkyHat> postfix is the best choice unless it doesn't do something you want (which is unlikely), in which case exim4 is probably the best choice
[04:04] <John24Doe> ok will do. thanks twb
[04:05] <twb> Yes, exim4 is the other main camp (especially over in Debian); anyone talking about sendmail or cyrus is a greybear or a Red Hat refugee
[04:05] <twb> *greybeard
[04:09] <funkyHat> jandrusk: so sslstrip relies on users not noticing the url is wrong?
[04:10] <RoyK> twb: last I checked, cyrus was a mail store, not an mta...
[04:11] <twb> RoyK: yes, but dovecot beats cyrus
[04:11] <twb> RoyK: I was disgusted to find recently that I had a postfix+dovecot instance out at a customer, where cyrus's sasl-bin was being used by postfix :-/
[04:11] <RoyK> some say so, yes, but still, neither are MTSs :P
[04:11] <RoyK> s/MTS/MTA/
[04:12]  * RoyK never made friends with cyrus' sasl
[04:12]  * funkyHat never heard a good word about it
[04:13] <jandrusk> funkyHat: sslstrip does not care what the URL is. It is essentially doing a MIM to sniff the data.
[04:14] <funkyHat> http://www.thoughtcrime.org/software/sslstrip/ the description suggests it's sniffing regular HTTP traffic and replacing links to HTTPS URLs with similar looking HTTP URLs or HTTPS ones if it has signed certs for them
[04:14] <funkyHat> It doesn't appear to be an actual working exploit on SSL
[04:15] <funkyHat> (yes, it's probably good enough to fool most people, but still not an actul SSL exploit)
[04:15] <twb> funkyHat: ah, thanks for the fact-check
[04:16] <funkyHat> jandrusk: ^
[04:17] <jandrusk> Hmm. It's been a while since I played with it, but was pretty sure I saw some CCN's from Amazon, GMail, etc... You may be right.
[04:19] <jandrusk> Pretty much all of the DLP (Data Loss Prevention) vendors clone the server certificate and then sniff the data in order to inspect it.
[04:20] <twb> "data loss prevention"
[04:20] <twb> Is that some sort of euphemism for JEDGAR?
[04:20] <jandrusk> lol
[04:20] <twb> "Hi this is the NSA for $10/mo we will back up your secrets to our secure data warehouse"
[04:20] <jandrusk> It's security technology that allows you to inspect network traffic for SSN's, Credit Cards Numbers,etc..
[04:21] <jandrusk> Or I should say, Corporations.
[04:21] <jandrusk> For $5/mo I'll give you access to the NSA's secure data warehouse.
[04:23] <twb> so basically dsniff
[04:25] <jandrusk> Right.
[06:52] <dckirba> Hello everyone, how are you? Is it ok to ask what may be a beginner's question in this channel?
[06:56] <TimR_> whats the question?
[06:59] <lifeless> dckirba: yes it is
[07:00] <koolhead17> hi all
[07:11] <TimR_> why did ubuntu take out syslog in 10.04.3 lts for?
[07:12] <TimR_> the reason why I am asking because webmin is not seeing syslog on the system
[08:02] <angelete2> hi
[08:02] <angelete2> i have such a ghost inside my ubuntu server
[08:03] <angelete2> i have this line in my crontab: #47 3   * * 1   root    test -x /usr/sbin/anacron || ( cd / && run-parts --report /etc/cron.weekly )
[08:03] <angelete2> so it's commented
[08:03] <angelete2> but it still executes
[08:04] <Jeeves_> angelete2: Have a look at /etc/cron.d/
[08:04] <angelete2> and even worse, it executes not on mondays but on tuesdays
[08:04] <angelete2> Jeeves_: what should i look for?
[08:06] <jibel> SpamapS, with the change to mysql yesterday, all mysql-* packages are now completely removed on upgrade included mysql-server, is it what is expected ?
[08:09] <Jeeves_> angelete2: About the same command
[08:36] <dckirba> hi, can anyone give me a bit of help setting up ubuntu 11.10 as a dhcp server for a small office?
[08:44] <angelete2> dckirba: what do you want to do exactly?
[08:45] <dckirba> angelete2: Hi! Small office network. Trying to set up Ubuntu 11.10 on one computer to provide IP addresses through DHCP, share files, internet
[08:45] <dckirba> angelete2: Having trouble configuring dhcp-server
[08:47] <dckirba> angelete2: I have install icp-dhcp-server and configured the files so that it listens to eth0 for requests and have specified a range of IP addresses to use as a pool
[08:48] <dckirba> angelete2: But plugging other computers into the network results in no communication with the dhcp server
[08:48] <dckirba> angelete2: in fact I can't even ping the server if I give another computer a static IP address
[08:50] <RoyK> dckirba: what ip addresses and subnet mask?
[09:01] <Daviey> uksysadmin: How are you getting on with your preseeding?
[09:01] <Daviey> (morning btw)
[09:02] <uksysadmin> Hi Daviey - I've succeeded!
[09:02] <uksysadmin> I'm at the stage now where I'm looking at what I want to achieve and improving it - that's a better place to be than scratching my head on the basics.
[09:03] <Daviey> uksysadmin: Great! Are you using orchestra, or plain preseeds?
[09:03] <uksysadmin> Ok - its a mix of the two, I guess.  It is done using Orchestra - but essentially its not using any of the features that Orchestra was designed for - its just cobbler and preseeding at this stage
[09:04] <uksysadmin> As I'm specifically looking at OpenStack - it naturally makes sense to look at achieving this with juju
[09:04] <Daviey> uksysadmin: interesting, i'd love it if you could write up what you are doing?  Is that possible?
[09:05] <uksysadmin> The issue is that I've a fairly tried and tested script that I've been using to roll out my environment - which is great for doing unattended installs, but I know it can be improved to fit with more "ubuntu friendly ways.
[09:05] <uksysadmin> sure - I'll be writing this up
[09:13] <Daviey> uksysadmin: great!
[09:17] <RoyK> erm...
[09:17] <RoyK> roy@tor:~$ ls -lh /proc/kcore
[09:17] <RoyK> -r-------- 1 root root 128T Nov 25 10:17 /proc/kcore
[09:17] <_ruben> nice :)
[09:54] <koolhead11> hi all
[09:57]  * koolhead11 wonders if everyone is out shopping on black friday!!
[10:05] <uksysadmin> hello koolhead11
[11:01] <koolhead11> hola uksysadmin
[12:01] <Daviey> jamespage: Hola, is https://blueprints.launchpad.net/ubuntu/+spec/servercloud-p-complex-deployment-testing comeplete?
[12:01] <jamespage> Daviey: yes
[12:02] <Daviey> jamespage: it does make the complexity look pretty simple... we'll crack that out in a day, right? :)
[12:02] <jamespage> Daviey: it needs to be taken in context of the other associated blueprints
[12:03] <Daviey> ah, yes - right
[12:03] <Daviey> thanks jamespage
[12:04] <Daviey> cmagina: How is https://blueprints.launchpad.net/ubuntu/+spec/servercloud-p-arm-storage-testing looking?
[13:02] <koolhead11> zul: around
[13:50] <zul> good morning
[13:54] <Mum29> Hey all. I was wondering, Currently i have a setup on a dotdeb server, they use latest mysql, When ubuntu 12.04 with mysql 5.5.xx it wont be the latest. So will there be any problems for them to move to ubuntu due to this ?
[13:55] <zul> jamespage/Daviey: so openstack jenkins
[13:55] <jamespage> zul: fire away
[13:56] <zul> jamespage: so we will want to test the trunk and stable releases for diablo until essex beccomes stable/essex
[13:56] <zul> so this includes horizon, swift, nova, quantum, keystone, and glance
[13:57] <jamespage> zul: so is trunk actually essex? but not yet stable?
[13:57] <zul> so i guess the idea is to have a ppa on commit trunk and maybe a weekly ppa for stable since testing stable can probably be done weekly
[13:58] <jamespage> and I assume we want to test diablo on oneiric and trunk/essex on precise?
[13:58] <zul> right trunk is actually essex
[13:58] <zul> correct
[13:58] <zul> so my idea is to re-use a lot of the openstack-ci scripts that is already done
[13:58] <zul> for creating the tarballs and ppas, etc, etc
[13:58] <jamespage> zul: absolutely
[13:59] <zul> thoughts?
[14:00] <zul> Daviey: ^^^
[14:00] <Daviey> hola
[14:01] <Daviey> yeah
[14:01] <zul> jamespage: but ill need help setting up the jenkins jobs
[14:01] <jamespage> zul: right
[14:01] <jamespage> just looking at openstack-ci
[14:02] <Daviey> I think having, nova-trunk-testing produces per commit packages, and nova-trunk-stable contains packages which passed the CI run
[14:02] <Daviey> Which means on success, jenkins copies the packages between the PPA's
[14:02] <zul> yeah thats what i was thinking as well
[14:03] <Daviey> jamespage: You were thinking?
[14:04] <jamespage> Daviey: 'passed the CI run' == Completely deployed and tested on hardware or something else
[14:04] <jamespage> ?
[14:05] <Daviey> jamespage: yeah
[14:05] <zul> Daviey: right but the diablo-stable doesnt go through any ci afaik
[14:05] <jamespage> zul: No reason why not
[14:05] <Daviey> jamespage: so jenkins tests the -testing PPA, and copies to -stable if it succeeds
[14:05] <jamespage> (famous last workds)
[14:06] <zul> jamespage: i meant for upstream
[14:06] <jamespage> ah
[14:06] <jamespage> I see
[14:06] <Daviey> jamespage: zul suggested last night that we could also do on-commit of the stable branch, as an SRU candidate.
[14:06] <zul> Daviey: i was thinking more like stable-testing and copies to -stable and then that is the SRU candidate
[14:06] <jamespage> OK
[14:06] <Daviey> zul: did you draft up your thoughts?
[14:07] <jamespage> so lemme just work this through
[14:07] <zul> so if people want to test it earlier before the SRU canidate is uploaded then they can
[14:07] <jamespage> So for Precise/Essex
[14:07] <jamespage> 1) New PPA somewhere - nova-trunk-testing
[14:07] <Daviey> jamespage: do you disagree with the mail i sent a while ago, for the approach?
[14:07] <Daviey> (high level)
[14:09] <Daviey> jamespage: perhaps we should copy Debian's naming style of experimental -> testing, rather than testing->stable..  We probably shouldn't call trunk stable.
[14:09] <zul> So this is what I was thinking
[14:09] <zul> Precise/Essex
[14:10] <zul> ppa for trunk (called i dont know what)
[14:10] <zul> and we do on commit testing in that ppa
[14:10] <zul> Oneiric/Diablo
[14:10] <Daviey> jamespage / zul: Fancy throwing this into etherpad?
[14:10] <jamespage> yesJenkins will detect changes on the
[14:10] <zul> Daviey: sure
[14:11] <jamespage> yes
[14:11] <zul> i think i might ttx's input as well
[14:11] <jamespage> I'll create one
[14:11] <ttx> .oO0Oo.
[14:11] <jamespage> http://pad.ubuntu.com/precise-openstack-testing
[14:12] <koolhead11> i am yet to get dash working, am using the truck/daiblo
[14:12] <koolhead11> :D
[14:12] <ttx> JAMES PAGE! Hello sir.
[14:12] <jamespage> hey ttx!
[14:13] <ttx> I should hang out here more often.
[14:13] <koolhead11> hey ttx
[14:13] <xranby> jamespage: hi i found something interesting this week.   since the default pthread stack size are 8Mb on arm in combination with lp861296   made java only able to start around 200 thread before running out of memory. this can manifest itself as a stall
[14:14] <xranby> jamespage: now we have a fix for lp861296     and it makes it possible to run at least 50% more threads
[14:14] <xranby> so we can run around 300 threads.
[14:14] <jamespage> xranby: thats great news!
[14:14] <xranby> by fixing the jvm sourcecode i can reduce the amount of memory allocated by pthread
[14:14] <xranby> and bump this up to around 7000 threads
[14:15] <xranby> jamespage: so i am working with robert to reduce thread memory consumption so that we can safely run thousands of threads
[14:16] <xranby> jamespage: you can try the pandaboard kernel attached to  https://bugs.launchpad.net/ubuntu/+source/linux/+bug/861296
[14:16] <xranby> and check if it makes your jenkins process more happy while it are checking for updates
[14:18] <jamespage> xranby: sure - leave it with me - I'll update the bug report
[14:21] <zul> Daviey: your presence is required
[14:21] <jamespage> zul: does the openstack-ci toolset have tools for detecting build success/failure?
[14:21] <zul> jamespage: i think so ill have to poke at it
[14:21] <zul> if not we can add it and then push it back upstream
[14:26] <Daviey> zul: oh?
[14:27] <zul> Daviey: well we would like your input
[14:31] <jamespage> zul: so I suggest that all of the individual project logic is encapsulated in 'ubuntu-openstack-ci' or whatever
[14:31] <jamespage> that way you can dev locally fine
[14:31] <zul> agreed
[14:31] <jamespage> we only use Jenkins to trigger builds and report on results
[14:32] <zul> but the trunk ppa will have swift, nova, glance etc etc etc
[14:32] <jamespage> yes - so jobs for them all
[14:32] <jamespage> openstack-essex-<component>-trunk
[14:32] <jamespage> where component == swift, nova, glance etc etc etc
[14:33] <zul> right
[14:35] <jamespage> zul, Daviey: I assume we can have a lp account in the right groups for the Ubuntu Jenkins Openstack user?
[14:35] <zul> i was thinking something like ~ubuntu-openstack-testers
[14:36] <Daviey> jamespage: pretty confident
[14:37] <jamespage> so we might want a team for the testing notifications (with a list) and a specific user who will do PPA uploads, send email etc...
[14:40] <jamespage> zul: I think we should just mimic behaviour for diable/essex re preocess
[14:41] <zul> jamespage: yeah its just one more extra step for copying to stable trunk to sru candidate trunk i think
[14:41] <zul> s/trunk/ppa/g
[14:41] <jamespage> why not consider 'stable' to be the SRU candidate PPA?
[14:41] <zul> jamespage: good point
[14:42] <zul> less work
[14:42] <jamespage> yes - and easily templated - makes setting up the a new release really easy
[14:43] <zul> right the only concern i have for the sru testing is that we dont have access to github yet and the bzr git-plugin doesnt do branches afaik
[14:43] <zul> so maybe it would be easier to generate tarballs in this case
[14:44] <zul> what do you think?
[14:44] <jamespage> zul: might take a little time but lets get access to github setup
[14:45] <jamespage> zul: the bzr mirror deadens the pace of commits anyway so it not ideal
[14:45] <zul> ok so we will definently need access to github
[14:49] <zul> Daviey: thumbs up with the pad thingy?
[14:58] <Daviey> zul: sorry, was OTP
[14:58] <Daviey> will read again
[14:58] <zul> no worries
[15:01] <SpamapS> jibel: removed, or replaced by 5.5 versions?
[15:02] <SpamapS> jibel: (re the mysql server packages)
[15:03] <Daviey> SpamapS: Did you say there was a transition page tracking this?
[15:04] <SpamapS> yes... http://people.canonical.com/~ubuntu-archive/transitions/libmysqlclient.html
[15:04] <Daviey> zul / jamespage: looks good, do have one question - on the pad
[15:04]  * SpamapS typed that from memory... may be wrong
[15:04] <Daviey> SpamapS: great! thanks
[15:05] <zul> SpamapS: still building on arm?
[15:05] <Daviey> zul: Nah, SpamapS FTBFS on arm :)
[15:06]  * SpamapS is big endian.. its always a problem
[15:06] <Daviey> SpamapS: you need more power.
[15:06] <SpamapS> 5.5 takes about 8-9 hours to build on the armel builders
[15:07] <jibel> SpamapS, on upgrade it installs mysql-client-5.5 mysql-client-core-5.5 mysql-server-5.5 mysql-server-core-5.5 but removes  mysql-client-5.1 mysql-client-core-5.1 mysql-server-5.1 mysql-server-core-5.1 included  mysql-server
[15:07] <jibel> which might be a problem for future upgrades
[15:08] <SpamapS> jibel: what version? The latest version re-added the mysql-server meta-package which should be installed
[15:08] <SpamapS> 5.5.17-4ubuntu5 I think added that
[15:08] <zul> Daviey: answered
[15:09] <hallyn> Daviey: so for https://blueprints.launchpad.net/ubuntu/+spec/servercloud-p-trusted-cloud...  how about i reduce that to 1 item about "follow along with evm development"?
[15:10] <jibel> SpamapS, 5.5.17-4ubuntu5
[15:10] <jibel> Broken mysql-server:amd64 Depends on mysql-server-5.5 [ amd64 ] < none -> 5.5.17-4ubuntu5 > ( database )
[15:10] <jibel>   Considering mysql-server-5.5:amd64 0 as a solution to mysql-server:amd64 0
[15:10] <jibel>   Removing mysql-server:amd64 rather than change mysql-server-5.5:amd64
[15:11] <hallyn> Daviey: in other words I think it's worth spending a few spare cycles helping it along, but our eye is to something useful for 14.04, not 12.04.
[15:11] <Daviey> hallyn: Right, i agree with that.. I suspect as is, it would end up being defered anyway
[15:12] <Daviey> hallyn: keep the WI's that are already DONE
[15:12] <hallyn> Daviey: ok.
[15:12] <hallyn> jjohansen: the seccomp2 in our kernel now, that doesn't allow fork after starting seccomp2 still, right?
[15:12] <SpamapS> jibel: *HM*
[15:13] <SpamapS> jibel: that may have been a wonky archive... mysql-server is no longer provided by mysql-server-5.5 ... so apt shouldn't be considering that change
[15:15] <jibel> SpamapS, k, I'm running the upgrade again
[15:15] <jibel> SpamapS, logs are here https://jenkins.qa.ubuntu.com/job/precise-upgrade/PROFILE=server-tasks-amd64,label=upgrade-test/lastBuild/
[15:15]  * SpamapS hugs jenkins
[15:17] <SpamapS> jibel: hm, the Breaks: on mysql-common may have unintended consequences.. hrm.
[15:22] <zul> Daviey: ok SRUs uploaded
[15:26] <Daviey> zul: keystone?
[15:26] <zul> Daviey: keystone, nova and glance
[15:27] <Daviey> zul: great!
[15:27] <zul> well see if they get accepted ;)
[15:28] <Daviey> SpamapS: would you be able to look at zul's SRU's soonly please? :)
[15:28]  * SpamapS will have a look at them today for sure
[15:28] <SpamapS> have not done my SRU duties in 2 days.. its time
[15:28] <Daviey> \o/
[15:33] <tjaalton> jamespage: hey, I'm merging curl-7.22.0-3 from unstable, hope that's ok
[15:39] <Daviey> Does anyone here mind if someone else hijacks a merge where they touched it last, and do not have an open "please merge" bug assigned?
[15:39]  * Daviey really doesn't mind, but i wondered what others felt about it?
[15:39] <hallyn> Daviey: uh, not me...
[15:39] <hallyn> "Git ur dun" I say
[15:42] <RoAkSoAx> Daviey: i don't mind it either
[15:42] <Daviey> ta
[15:43] <jamespage> tjaalton: thats fine - thanks for checking
[15:44] <tjaalton> jamespage: thanks
[16:08] <Daviey> SpamapS: who else is working on the mysql transition?
[16:10] <SpamapS> Daviey: nobody
[16:11] <Daviey> SpamapS: Are you cracking through the rebuilds?
[16:11] <SpamapS> I mean other than the occasional saving throw from our dungeon master cjwatson
[16:11] <SpamapS> Daviey: all no-change rebuilds are done. the things left are broken and need slight changes because of multi-arch
[16:11] <SpamapS> Daviey: also I seem to have broken upgrades so I'm fixing that
[16:12] <zul> RoAkSoAx Daviey: the cobbler patches got accepted upstream
[16:12] <RoAkSoAx> zul: cool
[16:12] <Daviey> SpamapS: I just did a test build for mythtv.. about to upload?
[16:12] <zul> no one cares about mythtv come on! :)
[16:13] <Daviey> gah /me uploads
[16:14] <SpamapS> Daviey: builds are fine. Its just upgrades of mysql-server from oneiric/lucid -> precise that are broken.. mysql-server gets removed.
[16:14] <Daviey> oh joy
[16:15] <SpamapS> Daviey: hmm, mythtv didn't show up on my dctrl-grep ...
[16:16] <Daviey> SpamapS: only checking main?
[16:17] <SpamapS> Daviey: no, checked them all I believe.. hrm
[16:17] <Daviey> O_o
[16:19] <SpamapS> libmysqlclient16-dev
[16:19] <SpamapS> Daviey: that will need fixing
[16:19] <Daviey> crappers. thanks.
[16:22] <ringods> Hello, following the info on JeOS and vmbuilder, but get stuck at the problem mentioned here http://ubuntuforums.org/showthread.php?t=1807125
[16:22] <ringods> No answers available unfortunately. Any hints?
[16:30] <cjwatson> SpamapS: I started on libdbi-drivers but got depressed at yet another crappy configure script that thinks it knows better than the linker
[16:31] <cjwatson> Daviey: ah, good, I was about to hunt down why mythbuntu image builds were broken
[16:40] <SpamapS> cjwatson: yeah they all seem to do that. :-P
[16:41] <SpamapS> cjwatson: I have libdbi-drivers working locally here actually
[16:41] <cjwatson> ok, good
[16:41] <SpamapS> it was broken because it build-depends on mysql-server too
[16:42] <SpamapS> It really is too bad how autotools works that it just lets bad code sit out there forever
[16:48] <TimR_> hey micheal aka genii-around
[16:50] <genii-around> TimR_: Hello.
[16:51] <TimR_> I did have issues last night
[16:51] <genii-around> TimR_: Ah. At what stage?
[16:51] <TimR_> At this point hsfconfig needs to be made so that we can apply the patch:
[16:51] <TimR_> sudo make install
[16:52] <TimR_> that is where I got lost at
[16:52] <genii-around> TimR_: I just got an incoming email from you actually...
[16:52] <TimR_> alright
[16:56] <genii-around> TimR_: Yes, the hsfconfig part is what configures the driver for your current kernel, etc. Is it failing to run or you are not sure what values to give it? For linux source build directory that matches your running kernel... to put: /usr/src/linux-headers-$(uname -r)
[16:57] <TimR_> I didnt even get to start that process yet because I didnt know what directory I should be working out of
[16:57] <genii-around> TimR_: ~/Linuxant/hsfmodem-7.80.02.04full/
[16:58] <TimR_> alright let me see if I can get it going
[16:59] <genii-around> TimR_: Work is requiring me on and off, ping if you have problems and I'll get back to you as soon as I can.
[16:59] <TimR_> ok
[17:02] <SpamapS> jibel: mysql-server removal problem found .. fix pending
[17:02] <TimR_> alright I got where its asking what build directory that matches your running kernal
[17:05] <TimR_> then its also saying no pre-build modules for ubuntu 10.04.3 linux-2.6.32-35-generic i686-smp
[17:26] <RoyK> anyone around that can help me with a wierd memory issue? with 64GB memory, seems 56GB is in use by processes, 40GB is swapped out, and ~zero free, also after adding buffers+cache .... http://paste.ubuntu.com/749481/
[17:32] <jamespage> zul: I've done the basic Jenkins job setup in the lab; I'll look at the openstack integration tests next weel
[17:32] <jamespage> have a good one
[17:40] <eagles051387> hey guys has anyone in here gotten xen working on 11.10
[17:48] <koolhead17> eagles051387: did you searched before asking same here?
[17:49] <eagles051387> koolhead17: i have compared the configuration i have to another server which is running debian and uses bridged networking and a slightly older version of the hypervisor and i had it set the same yet its not picking up my bridged networking configuration in the xend configuration file
[17:51] <koolhead17> eagles051387: i just did a search and got this http://www.beyondlinux.com/2011/11/02/install-xen-4-1-and-setup-your-cloud-os-on-ubuntu-11-10/
[17:51] <koolhead17> see if it helps
[17:51] <eagles051387> thanks
[17:52] <koolhead17> !xen
[17:55] <eagles051387> koolhead17: already looked there that is incomplete and very out dated
[17:55] <koolhead17> even the documentation ? :(
[17:56] <eagles051387> the link ubottu gives
[17:56] <eagles051387> as well the 11.10 documentation doesnt have any documentation on xen
[17:58] <eagles051387> koolhead17: just been told that using xend-config.sxp isnt the right way to configure things now
[17:58] <eagles051387> in regards to xen and bridged networking
[18:14] <mtaylor> Daviey: around?
[18:15] <mtaylor> Daviey, zul: https://bugs.launchpad.net/ubuntu/+source/glusterfs/+bug/785158
[18:15] <Daviey> mtaylor: o/
[18:17] <mtaylor> Daviey: coworker just pointed me at that - I'm trying to help them figure out how to better interact with canonical directly (so that it doesn't just involve forwarding bugs to me) but in the meantime, having a broken gluster seemed potentially bad :)
[18:17] <Daviey> mtaylor: This isn't #canonical-server btw :)
[18:18] <mtaylor> Daviey: of course not! I'm merely pinging you because a broken gluster is bad for #ubuntu-server ;)
[18:20] <Daviey> mtaylor: okay, thanks!
[18:22] <ersi> there's a #canotical-server? :O
[18:23] <Daviey> no.
[18:23] <ersi> heh.
[18:23] <Daviey> mtaylor: Are you aware of the SRU process?
[18:24] <Daviey> (not sure the suggested fix is correct)
[18:24] <mtaylor> Daviey: I'm aware that it exists, but don't believe I've ever done it
[18:27] <zul> Daviey: that work around in the past has been used
[18:30] <Daviey> mtaylor: seeing that in just natty?
[18:31] <mtaylor> Daviey: it's possible - I'm just now discovering that the servers in question are all running natty ... so this might be about getting someone to learn SRU process and proposing fixes to be backported
[18:31] <mtaylor> I'm also questioning the sanity of fixating on natty
[18:31] <zul> yes because oneiric is alot better :P
[18:33] <Daviey> mtaylor: Ah, hoped it was confirmed or not on oneiric
[18:37] <KurtKraut> How do I truly change the default language (locale) in the Ubuntu server? I'm getting this errors during aptitude: perl: warning: Setting locale failed.
[18:37] <KurtKraut> perl: warning: Please check that your locale settings:
[18:37] <KurtKraut> 	LANGUAGE = (unset),
[18:37] <KurtKraut> 	LC_ALL = (unset),
[18:37] <KurtKraut> 	LC_CTYPE = "pt_BR.UTF-8",
[18:37] <KurtKraut> 	LC_COLLATE = "pt_BR.UTF-8",
[18:37] <KurtKraut> 	LC_MESSAGES = "pt_BR.UTF-8",
[18:38] <KurtKraut> 	LANG = "en_US.UTF-8"
[18:38] <KurtKraut>     are supported and installed on your system.
[18:38] <KurtKraut> perl: warning: Falling back to the standard locale ("C").
[18:38] <KurtKraut> locale: Cannot set LC_CTYPE to default locale: No such file or directory
[18:38] <KurtKraut> locale: Cannot set LC_MESSAGES to default locale: No such file or directory
[18:38] <KurtKraut> locale: Cannot set LC_ALL to default locale: No such file or directory
[18:38] <KurtKraut> Oops, sorry for pasting.
[18:38] <eagles051387> KurtKraut: i am as well
[18:38] <eagles051387> not sure how to fix them though :(
[18:40] <ersi> eagles051387, KurtKraut: /etc/environment if I'm not mistaken
[18:40] <ersi> There you can change what the system-wide locale is
[18:40] <ersi> atleast previously
[18:40] <eagles051387> ersi: im on a clean install and im seeing this
[18:40] <eagles051387> thats what doesnt make sense
[18:40] <eagles051387> clean install of 11.10
[18:41] <ersi> That I have no idea about. Try checking /etc/environment
[18:41] <ersi> Clean install of 11.10 worked just fine for me, I got en_US.UTF-8 as system wide locale
[18:42] <ersi> KurtKraut: You could set those on either the command line or in your bash profile btw, if you wish to just set it for your user
[18:43] <KurtKraut> ersi, this is a very common problem. If I purchase a VPS or dedicated server with Ubuntu, the datacenter realizes that I'm from Brazil and tries to configure the pt_BR locale. But they don't include the locale packages in the default install and seems to partially set the locale.
[18:43] <ersi> Sounds like the VPS provider is silly
[18:44] <ersi> so.. install the br-langpack? :)
[18:46] <KurtKraut> ersi, it doesn't happen with one VPS provider. It happens to all, event AWS Amazon.
[18:46] <KurtKraut> ersi, I can name about 10 providers that provokes such behaviour.
[18:47] <ersi> It's all about how the default image is
[18:48] <ersi> If you do isolate it to the standard Ubuntu install, file/search for a bug about it :)
[18:50] <eagles051387> ersi: care to explain my issue with locals on a clean install and not in a virtual machine
[18:51] <ersi> I have no idea what so ever why you have that problem
[18:52] <mtaylor> Daviey: not confirmed on oneiric to the best of my knowledge
[18:53] <ersi> but if you can isolate it, I'm sure the developers/maintainers are interested.. note down what you do when you install and file a bug if it doesn't meet your expectations
[18:53] <mtaylor> KurtKraut: one sec - I have the fix documented somewhere...
[18:53] <mtaylor> KurtKraut: it's quite annoying
[18:53] <ersi> or search if there's anything filed already
[18:54] <mtaylor> KurtKraut: a) you need to ensure that you have a locale selected and that locale is in /etc/locale.gen (in your case, I believe it probably is)
[18:55] <mtaylor> KurtKraut: next, run the program "locale-gen"
[18:55] <KurtKraut> mtaylor, thanks, this doesn't fix. I've found a fix by trial and error.
[18:56] <mtaylor> KurtKraut: yeah? what was your fix?
[18:56] <Daviey> mtaylor: Sure?
[18:56] <KurtKraut> mtaylor, the problem is none of these locale commands (ex.: dpkg-reconfigure locales;locale-gen) where able to create and store the system variables LC_COLLATE, LANG and LANGUAGE.
[18:56] <KurtKraut> mtaylor, I've manually set them by trial and error in /etc/enviroment and at least now I don't see that flood of locale erros.
[18:57] <mtaylor> Daviey: as in, the folks locally aren't running oneiric, so they can't verify if they are seeing the problem on oneiric or not
[18:57] <mtaylor> Daviey: so I have no idea if it's on oneiric - although I would like to find out
[18:58] <KurtKraut> But I'm not sure at all if I've put the correct variables in /etc/enviroment: http://pastebin.com/JcrxSCvH
[18:58] <Daviey> mtaylor: did you see Marc's update?
[18:58] <Daviey> he added a bug task pointing to the upstream issue.
[18:59] <mtaylor> well - you should not need to put things in /etc/environment - BUT - if that works for you, awesome!
[18:59] <mtaylor> Daviey: just saw just now :)
[18:59] <mtaylor> Daviey: that's good to know at least
[18:59] <Daviey> mtaylor: but yes, it looks like s/2/1/ looks like a good fix for natty.
[19:00] <mtaylor> Daviey: and a thankfully not-very-intrusive one
[19:01] <Daviey> :D
[19:01] <Daviey> my favourite flavour
[19:07] <TimR_> does anybody know how to get the syslog to show up in webmin
[19:07] <TimR_> for 10.04.3lts server
[19:09] <genii-around> !webmin
[19:10] <TimR_> see I had this issue with couple other desktops I have
[19:11] <TimR_> unless there is someway I can read the logs from webgui
[19:15] <zul> mtaylor: https://jenkins.openstack.org/job/python-novaclient-tarball/58/console
[19:33] <eagles051387> !grub2
[19:41] <koolhead17> eagles051387: :)
[20:37] <hallyn> Daviey: around?
[20:38] <hallyn> Actually maybe I should ask in #ubuntu-devel in general.
[21:14] <chilicuil> hi there, does anyone know of a one line dhcp server?, something like this: $ ./give_ip 10.0.0.2 -i eth0 (to give the ip 10.0.0.2 to the first one who request it trought the eth1 interface), for http I can do $ sudo python -m SimpleHTTPServer 80
[21:22] <Daviey> hallyn: yup
[22:17] <Guest70665> help
[22:19] <RoyK> !ask
[22:21] <SpamapS> besides its, "HAAALP"
[22:21] <SpamapS> ^^ the magic word
[22:21] <SpamapS> uvirtbot: your mother was a toaster
[22:31] <genii-around> I guess the toaster talk threw them off
[22:33] <hallyn> stgraber: are you still around?
[22:34] <stgraber> hallyn: yep
[22:34] <hallyn> stgraber: https://code.launchpad.net/~serge-hallyn/ubuntu/precise/libcgroup/fix-cgrouplite-fstab, it's a trivial fix, i don't have perms to upload.  Woudl you mind?
[22:37] <stgraber> hallyn: fixed (you have upload rights now)
[22:37] <stgraber> this package should have been in the ubuntu-server package set, we probably missed it when creating it
[22:38] <hallyn> stgraber: oh!  i thought (someone said) it was bc it was in universe
[22:38] <hallyn> stgraber: cool, thanks.
[22:38] <stgraber> oh, that can explain why it wasn't included in the package set initially indeed
[22:39] <hallyn> does it mean it shouldn't be?
[22:39] <stgraber> there's nothing wrong with having package sets contain mixed main and universe packages
[22:40] <stgraber> my problem is rather with having lxc (the binary package) and cgroup-lite in universe, I think that at least for the LTS they should be in main as it's something we're pushing quite a bit, at least on ARM
[22:41] <hallyn> soren: don't know wha thappened that time.  http://people.canonical.com/~serge/kill-etherboot/etherboot_5.4.5.dsc should be up now.
[22:41] <hallyn> oops, sorry about that
[22:42] <stgraber> hallyn: just did a quick check, it's not the only package that's not in main (squid, mime-tools and feedparser are in the same packageset but not in main), so I really don't feel bad about adding libcgroup to that package set then :)
[22:42] <hallyn> stgraber: ok, cool, thanks
[22:44] <mtaylor> zul: fixed. re-running.
[22:45] <mtaylor> SpamapS: sup
[22:51] <Daviey> hallyn / stgraber / jjohansen : At some point, would you mind documenting what we can hope for in 12.04 in regards to security for lxc?
[22:52] <SpamapS> mtaylor: why oh why couldn't you have written pandora-build 10 years earlier?
[22:52] <SpamapS> mtaylor: and, er, hi
[22:52] <SpamapS> mtaylor: seems *every* project invented their own way to check for libmysqlclient
[22:52] <hallyn> Daviey: 'in regards to security' -> what exactly do you mean?
[22:52] <Daviey> hallyn: Setting expectations of the areas of weakness
[22:53] <Daviey> (and the areas of added win ofc)
[22:53] <mtaylor> SpamapS: that's because there is no good way of checking for libmysqlclient
[22:53] <hallyn> Daviey: basically, without user namespaces, which won't be there for 12.04, there's not much to say.
[22:53] <mtaylor> SpamapS: because all of the mysql install/release engineering is batshit
[22:54] <mtaylor> SpamapS: also, I've got a few outstanding pandora-build feature requests, so I may do a little bit more hacking here
[22:54] <SpamapS> mtaylor: at least since 5.0 there has been mysql_config --libs
[22:54] <hallyn> you can trust your containers for correctness and reliability, but don't give one to untrusted user.
[22:54] <mtaylor> SpamapS: but that doesn't always work
[22:54] <hallyn> Daviey, let's gather the troops on monday to chat about it :)
[22:54] <mtaylor> SpamapS: and has its own set of gotchas
[22:54] <Daviey> hallyn: But what /can/ we do to add secuity.. I read an old paper you wrote that suggested some ways many years ago... is that still valid?
[22:54] <SpamapS> mtaylor: only because people do things wrong. ;)
[22:54] <mtaylor> SpamapS: and gives you some shit that's not actually valid
[22:55] <mtaylor> SpamapS: and mysql_config --libs shouldn't be needed in the normal case if the lib were linked properly
[22:55] <SpamapS> mtaylor: I've never had problems with --libs .. have not used the others admittedly
[22:55] <hallyn> Daviey: as we can't stack selinux or smack on top of apparmor, it doesn't much apply
[22:55] <mtaylor> SpamapS: depends on if someone is installing from source, or if they have instaled one of the drop-in binary tarball dists
[22:55] <Daviey> hallyn: apparmor can't help?
[22:55] <hallyn> Daviey: we would need some extensions on apparmor to make it work
[22:55] <mtaylor> SpamapS: so you can't actually 100% count on it from an autotools perspective :)
[22:55] <hallyn> which is why i say let's talk monday with jjohansen (and maybe stgraber)
[22:56] <SpamapS> mtaylor: suck :-P
[22:56] <mtaylor> SpamapS: yup
[22:56] <Daviey> hallyn: I might be pushing your area here, but how much would need to be added to apparmor?
[22:56] <mtaylor> SpamapS: simple fix - use libdrizzle
[22:56] <Daviey> hallyn: as in weeks worth, or a couple of days?
[22:56] <SpamapS> mtaylor: then again, why would I expect mysql release engineering to understand.. they symlink libmysqlclient_r.so.18 to libmysqlclient.so
[22:57] <hallyn> Daviey: at least weeks.  we need the ability to specify a pathname relateive to procfs or sysfs, for instance ( no matter where it is mounted)
[22:57] <mtaylor> SpamapS: they also think that the cmake port of mysql is not without problems, and the cmake is not deficient
[22:57] <Daviey> hallyn: right, lets start the weekend and sort it Monday :)
[22:57] <hallyn> Daviey: thanks, bc i'm startgint to type really slowly
[22:57] <Daviey> hallyn: have a good-in'
[22:57] <brianherman> im trying to get xen running on ubuntu
[22:57] <brianherman> i get this error
[22:57] <Daviey> .. and everyone else
[22:57] <Daviey> o/
[22:58] <hallyn> you too :)
[22:58] <brianherman> WARNING!  Can't find hypervisor information in sysfs!
[22:58] <brianherman> i installed the hypervisor package
[22:59] <brianherman> is it because im not running the server kernel
[22:59] <hallyn> Daviey: and i'll send you the source for my googlechat->irc proxy :)
[22:59] <SpamapS> mtaylor: http://people.canonical.com/~ubuntu-archive/transitions/libmysqlclient.html .. pretty much all of those with all red X's is the people who wrote their own crappy mysql autoconf function
[23:00] <mtaylor> SpamapS: wow. that sucks for you
[23:01] <SpamapS> mtaylor: yeah, multiarch + libmysqlclient == sad panda
[23:03] <brianherman> nobody got xen experience?
[23:06] <SpamapS> brianherman: KVM has been the preferred hypervisor on Ubuntu for a while...
[23:06] <SpamapS> brianherman: Xen dom0 support only came back as of 11.10 I believe (maybe 11.04)
[23:06] <brianherman> yeah its in 11.10
[23:06] <brianherman> is it in the server kernel?
[23:07] <brianherman> how would i check
[23:07] <SpamapS> yes
[23:07] <brianherman> oh ok
[23:07] <brianherman> so its just my kernel
[23:07] <brianherman> i was running desktop
[23:07] <brianherman> sorry to bother you
[23:07] <SpamapS> Might also be in the desktop kernel
[23:07] <SpamapS> zul: ^^ got any tips for brianherman ?
[23:07] <SpamapS> brianherman: zul is our xen expert :)
[23:07] <brianherman> cool
[23:08] <SpamapS> might not be online right now tho
[23:10] <brianherman> im nuking my dads computer from orbit
[23:10] <brianherman> so
[23:10] <brianherman> i can wait lol
[23:10] <brianherman> thanks spamaps
[23:11] <SpamapS> mtaylor: hey guess what? API change in mysql 5.5's libmysqlclient. AWESOME
[23:11] <SpamapS> extern void my_free(void *ptr);
[23:11] <SpamapS> 5.5 has a real my_free function..
[23:11] <SpamapS> 4.1 its a macro.. with 2 args
[23:11] <SpamapS> >:|
[23:12] <brianherman> have you tried mariadb?
[23:12] <SpamapS> #define my_free(PTR,FLAG) _myfree((PTR), __FILE__, __LINE__,FLAG)
[23:14] <SpamapS> oh they're saying thats an internal function so its ok
[23:15] <mtaylor> SpamapS: yup. internal function - except that it's in the public header file
[23:15] <SpamapS> :)
[23:16] <mtaylor> SpamapS: also - they don't know how to mark symbols as public/private
[23:16] <SpamapS> since myodbc is "part" of mysql, I guess they figured its ok to call my_free
[23:16] <mtaylor> SpamapS: which might have something to do with their move to cmake - except that they weren't marking symbols as private properly before the move, so I can't blame it on that
[23:16] <mtaylor> SpamapS: well, you _have_ to call my_free (in theory) if you call my_malloc
[23:17] <mtaylor> SpamapS: BUT - libmysys is an "internal" library
[23:18] <mtaylor> ok. I'm going to go spend money on electronics
[23:23] <Daviey> RoAkSoAx: I just looked at 51_koan_grub2_instead_of_grubby.patch .. that is neat, but doesn't allow download of kernel/initrd at grub runtime, right?
[23:23] <RoAkSoAx> Daviey: don't remember if it does it before adding the grub entry and saves it in a temp dir, or after
[23:23] <RoAkSoAx> in runtime
[23:24] <SpamapS> mtaylor: just say no to tamagotchi :)
[23:24] <Daviey> RoAkSoAx: but i can't declare for use at grub time, use this next-server? right?
[23:25] <zul> brianherman: you are to run a -server kernel
[23:25] <Daviey> RoAkSoAx: The thought being, it should also be able to boot from local disk, on no change reboots?
[23:27] <RoAkSoAx> Daviey: yes you can reboot fromm local disk on no change reboots
[23:27] <RoAkSoAx> Daviey: it just adds an entry to grub
[23:28] <Daviey> RoAkSoAx: right, but i'm looking at the scenario where people are using booting locally to grub, with the first entry being $something (I was thinking ipxe).. which pulls down the config from cobbler, and either localboot or re-install.
[23:29] <Daviey> The tricky part is, also careing for the situation where we don't have access to dhcpd.. so we pre-declare the cobbler server.
[23:29] <RoAkSoAx> Daviey:right
[23:29] <RoAkSoAx> Daviey: the --replace-self is "pre-declaring" aswell
[23:29] <Daviey> RoAkSoAx: but you need to do that before every reboot?
[23:29] <RoAkSoAx> grabs the initrd, linux, and gets the kickstart based on the system or profile
[23:30] <RoAkSoAx> Daviey: yes before every reboot
[23:30] <RoAkSoAx> Daviey: i mean, the idea of --replace-self is just to replace an installation
[23:30] <RoAkSoAx> Daviey: for that particular, already deployed, system
[23:30] <Daviey> RoAkSoAx: Isn't that de-centralising reinstalls?
[23:31] <RoAkSoAx> Daviey: you could say so, but either way, it is a way of reinstalling regarles cobbler system has pxe boot enabled or disabled
[23:31] <RoAkSoAx> Daviey: cause it will force the reinstallation
[23:32] <Daviey> RoAkSoAx: Hmm, ok - do you think the same could be achieved via grub-ipxe?
[23:34] <RoAkSoAx> Daviey: i'll have to look at grub-ipxe first and then see what we can achieve with both of the tools
[23:34] <RoAkSoAx> and in what situations to use them
[23:35] <Daviey> RoAkSoAx: pxe-kexec is also of interest.
[23:35] <RoAkSoAx> Daviey: ya
[23:36] <Daviey> RoAkSoAx: The thing that concerns me about ipxe, is that i'm not sure you can declare a script outside of the binary.. i think it needs embedding.. but i'm not certain
[23:36] <Daviey> this makes life less interesting.
[23:36] <RoAkSoAx> Daviey: yeah.. well I guess we'll have to play with it
[23:36] <Daviey> http://ipxe.org/scripting
[23:36] <Daviey> rocking!
[23:57] <zul> ooh i like the ipxe scripting