[02:43] <lifeless> hallyn: around? mountall barfing again for me in lucid-lxc
[02:43] <lifeless> dpkg: error processing /var/cache/apt/archives/mountall_2.15.3_i386.deb (--unpack):
[02:43] <lifeless>  unable to make backup link of `./lib/init/fstab' before installing new version: Invalid cross-device link
[02:49] <twb> Urk
[02:49] <twb> lifeless: what's your /proc/mounts look like
[02:49] <twb> pref. both inside and outside the container
[02:50] <lifeless> twb: http://paste.ubuntu.com/850850/
[02:51] <lifeless> in, then out
[02:51] <twb> Hm, btrfs has a limit on ~30 hard links in one dir, this causes problems for at least git package upgrades...
[02:52] <twb> Where is it backing up lib/init/fstab *to*?
[02:52] <twb> It's dpkg doing it, right?  We are not talking about the postinst
[02:53] <lifeless> I assume so
[02:53] <lifeless> Unpacking replacement mountall ...
[02:53] <lifeless> dpkg: error processing /var/cache/apt/archives/mountall_2.15.3_i386.deb (--unpack):
[02:53] <lifeless>  unable to make backup link of `./lib/init/fstab' before installing new version: Invalid cross-device link
[02:53] <lifeless> Errors were encountered while processing:
[02:53] <lifeless> ls -l /lib/init/fstab
[02:53] <lifeless> -rw-r--r-- 0 root root 444 2011-08-11 18:27 /lib/init/fstab
[02:54] <lifeless> stat /lib/init/fstab
[02:54] <lifeless> ..
[02:54] <lifeless> Device: fc07h/64519d    Inode: 87044329    Links: 0
[02:54] <lifeless> regular file
[03:01] <twb> Is that saying it has 0 hard links ot 444 hard links?
[03:01] <twb> That's weird, the 0 should be a 1
[03:02] <twb> On my debian 2.6.32 system running btrfs, /etc/fstab for example has a 1 in that column, indicating one instance of the inode on the filesystem.  If it was hard-linked it'd say "2", but why would it ever say 0
[03:09] <patdk-lap> in what column?
[03:12] <twb> After the DAC, second column
[03:12] <lifeless> twb: 0 links
[03:13] <twb> It can't have 0 links
[03:13] <lifeless> twb: I'm not running btrfs :P
[03:13] <twb> If you just touch a file, that file has a link
[03:13] <lifeless> indeed
[03:13] <lifeless> just reporting what is there ;)
[03:14] <patdk-lap> what if it's a deleted file?
[03:14] <twb> patdk-lap: in that case maybe
[03:14] <patdk-lap> since btrfs doesn't really delete a file
[03:15] <twb> But if it's deleted it won't appear in the dirent
[03:15] <lifeless> what does btrfs have to do with it ?
[03:15] <patdk-lap> dunno, twb said running btrfs
[03:15] <twb> lifeless: I am just doing a knee-jerk reaction to btrfs + hard links
[03:15] <lifeless> sure, but I'm running ext4....
[03:15] <twb> That's not what the paste said
[03:16] <lifeless> thats /dev/sde-1
[03:16] <twb> Oh, /media/tb1-2 only, never mind all the noise
[03:16] <lifeless> my external drive with a copy of archive.u.c on it
[03:16] <twb> The 0 links is still very odd tho
[03:16] <lifeless> yup
[03:16] <twb> ohhhh
[03:17] <twb> it's a file bind mount or something -- look at line 14
[03:17] <twb> That's fucking WEIRD
[03:17] <lifeless> yeah
[03:17] <BuenGenio> hello
[03:17] <BuenGenio> after log rotation apache for some reason will stop writing to the new log file, until I do a manual restart (service apache2 restart)
[03:17] <twb> I've never seen that before
[03:17] <BuenGenio> any idea?
[03:17] <twb> BuenGenio: logrotate is supposed to tell apache to reload
[03:18] <twb> BuenGenio: if that doesn't happen automatically your logrotate <--> apache interaction is broken
[03:19] <twb> BuenGenio: this is because apache opens the log file and keeps writing to it forever, so after rotation it is still writing to the rotated file; with most daemons (dunno about apache) you send it a HUP which tells it to close and reopen the log file based on path, i.e. it opens the new log file
[03:21] <FainaUkraina> sorry got disconnected
[03:22] <FainaUkraina> so is it an Ubuntu thing with logrotate?
[03:22] <twb> FainaUkraina: no it's a unix-wide thing, everything that rotates logs needs to do this
[03:23] <twb> http://paste.debian.net/156995/
[03:23] <FainaUkraina> yes, I understood it needs a HUP, but why does ubuntu not send it?
[03:23] <twb> It should
[03:23] <twb> Presumably you broke it somehow
[03:58] <qman___> yeah, apache has proper log rotation in a default install
[03:58] <qman___> it should 'just work' unless you go and change it
[05:53] <funkysmell> hello. I have access to ubuntu server, but only via ssh. I have sudo access. I need to change the proxy settings on the server so I can use wget. any help would be awesome
[05:54] <funkysmell> where wold i find the proxy details?
[05:57] <twb> funkysmell: http_proxy=http://proxy:8080/
[05:58] <twb> CLI tools usually honour that environment variable; e.g. w3m, curl, wget do.  Most GUI browser do not, because they have stupid fat heads.
[05:58] <twb> As with other environment variables, you need to either put it in front of the command you want to run, or export it.
[06:17] <funkysmell> twb: Cheers, I'll try it an let you know
[07:58] <uvirtbot`> New bug: #930252 in squid3 (main) "squid3 crashed with SIGABRT in fatal_dump()" [Critical,Triaged] https://launchpad.net/bugs/930252
[08:16] <koolhead17> hi all
[09:08] <iclebyte> where could I grab the source code for the version of 'login' installed on ubuntu 10.04lts ?
[09:16] <uvirtbot`> New bug: #937613 in openldap (main) "Segmentation fault in libc" [Undecided,New] https://launchpad.net/bugs/937613
[09:50] <kai> iclebyte: apt-get source login ? or somesuch
[10:13] <RoyK> iclebyte: use dpkg -S to find the package containing the binary and then apt-get source to grab the source code
[10:21] <iclebyte> i got it thanks.
[10:22] <iclebyte> is it possible to lock telnet down to only specific users?
[10:22] <iclebyte> in.telnetd
[10:22] <larsemil> Fatal: device-mapper: mapped boot device cannot be on multiple real devices
[10:22] <larsemil> what to do ^^
[10:22] <larsemil> that is when running depmod
[10:33] <rbasak> jamespage: I've just written a tool that recurses down reverse build depenencies. The transition is now 206 packages.
[10:34] <jamespage> rbasak, 206?
[10:34] <rbasak> jamespage: maybe I've made a mistake
[10:34] <jamespage> rbasak, hrm - maybe
[10:34] <jamespage> that sounds huge
[10:35] <lynxman> morning o/
[10:36] <eutheria> is /var/backup for general purpose use?
[10:44] <TeTeT> after upgrading from 10.04 to Precise, the vm's managed through libvirt are gone. Do the config files need to be changed?
[10:48] <smb> TeTeT, which part was upgraded the VM host or the machine running them? And what VM type KVM or Xen?
[10:50] <TeTeT> smb: the host was upgraded, kvm in use
[10:51] <TeTeT> more details: http://pastebin.ubuntu.com/851214/
[10:52] <rbasak> jamespage: does this look reasonable? http://paste.ubuntu.com/851215/
[10:53] <smb> TeTeT, Ok, not sure where managed instances are saved in the KVM case. Though the virsh list seems to show some, but I guess those you did re-create?
[10:54] <smb> TeTeT, Oh, actually I guess they should be in qemu dir you listed... :-P
[10:54] <jamespage> rbasak, it looks like to much
[10:55] <smb> TeTeT, Let me check something on a local machine...
[10:55] <jamespage> we only need to transition packages that actually depend on libopenmpiXX
[10:55] <TeTeT> smb: yes, I created a few new ones
[10:56] <smb> TeTeT, Could you pastebin (or put onto chinstrap) one of the non-working configs?
[10:58] <TeTeT> smb: http://pastebin.ubuntu.com/851226/
[11:02] <smb> TeTeT, Hm, ok. Odd. At least looking quickly the format seems still the same here. Would there be anything interesting in /var/log/libvirt/libvirtd.log?
[11:04] <TeTeT> smb: full of this error type: 2012-02-20 07:58:57.850+0000: 1278: error : virDomainDefParseXML:7542 : unsupported configuration: Only the first console can be a serial por
[11:06] <smb> TeTeT, Hm, at least there seems to be the console section listed twice in the xml you pasted...
[11:06] <TeTeT> smb: seems the console has been changed
[11:06] <tjaalton> zul: hey, are you planning to merge samba from debian, 3.6.2 seems to include a bunch of bugfixes (.3 just one CVE)
[11:07] <TeTeT> smb: yep, the old vms use two time console, the new one serial and console
[11:08] <smb> TeTeT, ok that probably should be handled more gracefully on upgrade... I guess when you can verify that changing one of those to serial does fix it to be displayed this could be considered a bug in the upgrade...
[11:11] <TeTeT> smb: worked for one vm so far, haven't changed the others
[11:13] <smb> TeTeT, Ok, well have not heard about it yet. MAybe needs a bug still. Btu I try to remember to bring it up in the meeting today. If there is nothing coming up before
[11:13] <TeTeT> smb: but is filed, bug 931350, strangely I did not get bug emails from it
[11:13] <uvirtbot`> Launchpad bug 931350 in libvirt "vms missing after upgrade from Lucid to Precise" [Undecided,Incomplete] https://launchpad.net/bugs/931350
[11:15] <smb> TeTeT, Weird indeed...
[11:16]  * smb has to physically transfer his entity... (->afk)
[11:19] <lx-berlin> hi, anyone here? I have a ubuntu software RAID 1. I just booted the computer and everything was fine. Then i did a "sudo apt-get update" and then "sudo shutdown -r now". After the restart mdadm started a rebuild. This is not normal, right?
[11:44] <eutheria> i am struggling to load my system to generate a pgp key
[11:45] <eutheria> i wonder if it is because it is a vm
[11:46] <patdk-lap> well, you need random data
[11:46] <patdk-lap> and vm's have less random data
[11:47] <patdk-lap> cause more interrupts and other things, network traffi, ...
[11:47] <eutheria> would running stress help?
[11:47] <patdk-lap> dunno
[11:47] <eutheria> it hits the cpu and disk
[11:48] <eutheria> oh why didn't i generate it on the host
[11:48] <patdk-lap> rng-tools is the quik method
[11:52] <eutheria> well the rngtest seems to help
[11:54] <eutheria> oh weird the number went down then up
[12:08] <eutheria> wow that was difficult
[12:48] <zul> tjaalton: it would hvae been done yesterday but i was on holiday ill get to it today but its assigned to someone else
[13:00] <tjaalton> zul: cool, nice to hear (just happened to bump into the relnotes of 3.6.2, that's all)
[13:10] <jamespage> hallyn, utlemming - really like the ubuntu-cloud lxc template - works really nicely
[14:16] <husien> hello guys.. i'm trying to intall ubuntu server 11.10 into my VmWare workstation8. and i got tis --->> Warning :file:///cdrom/pool/main/v/vim/vim-tiny_7.3.154+hg~74503f6ee649-2ubuntu3_i386.deb was corrupt
[14:16] <husien> Warning :couldn't download package vim-tiny (ver 2:7.3.154 +hg~74503f6ee649-2ubuntu3 arc i386
[14:17] <husien> redownload the ubuntu server 11.10 and burn the iso with 3 diffrent software. still got the same problem.. ermmmmm...
[14:17] <pmatulis> hmm, just upgraded a lucid server and the language pack for Firefox was brought in...
[14:18] <zul> Daviey: hey did you have a look at the merge yet?
[14:19] <koolhead17> husien: can you do md5 checksum for the ISO integrity?
[14:22] <koolhead17> husien: https://help.ubuntu.com/community/VerifyIsoHowto  this doc will help you :)
[14:28] <eutheria> what would be the 'best' way to backup a mysql server?
[14:31] <koolhead17> eutheria: backup as in?
[14:32] <eutheria> well i was just going to mysqldump it all
[14:33] <eutheria> but i just thought i would ask the question and see if someone said, the magic tool that gives all options is
[14:33] <koolhead17> eutheria: i would use mysqldump too :
[14:47] <husien> i using winMd5sum to cheksum the Iso..  MD5SUM-->>881d188cb1ca5fb18e3d9132275dceda compare 881d188cb1ca5fb18e3d9132275dceda the result is same..
[14:51] <koolhead17> husien: where did you get  881d188cb1ca5fb18e3d9132275dceda ? From the ubuntu download page?
[14:52] <eutheria> koolhead17, i found a package that does some mysql back automation, i will see if it is any good
[14:52] <husien> http://releases.ubuntu.com/oneiric/MD5SUMS
[14:53] <koolhead17> eutheria: what is it called. i am curious too :)
[14:53] <eutheria> automysqlbackup
[14:54] <eutheria> it isn't doing much, but i just wanted something that i could do an apt-get install on and have backups done quickly
[14:54] <geekbri> I seem to be having an issue where a user with a nonstandard home directory in 10.10 isn't reading properly from .bashrc or .profile.  Anybody have any experience with this issue?  I tried setting JAVA_HOME in both of those and it is always blank when i sudo -i -u user or su - username.  If i set the JAVA_HOME once I am logged in as the user it works fine...
[14:54] <RoyK> eutheria: mysqldump > some-file, backup that file
[14:55] <koolhead17> husien: how about installing the package from repository, incase you have internet access on the VM once installed?
[14:55] <eutheria> RoyK, more like dbs = `mysql -e 'show databases` for db in dbs: mysqldump --opt --rountines etc etc
[14:56] <RoyK> eutheria: or --all-databases ...
[14:56] <RoyK> eutheria: I never specified opts - that's up to you ;)
[14:56] <eutheria> RoyK, i like to have a .sql file of each db,
[14:56] <eutheria> i just wanted something that would do that for me
[14:57] <eutheria> i've written that backup script so many times and forgot about it so many times :)
[14:57] <RoyK> it takes like two minutes to write that shell script...
[14:57] <eutheria> takes 10 seconds to apt-get install autobackupmysql
[14:57] <RoyK> heh - use a private source repository
[14:58] <RoyK> eutheria: plus the time to find it plus the time to configure it?
[14:58] <eutheria> configured out the box
[14:58] <eutheria> i know it know so done :)
[14:58] <eutheria> err now
[14:58]  * RoyK mutters something about laziness
[14:58] <eutheria> all good programmers are lazy
[14:59] <RoyK> eutheria: and all operations personel like to be in control
[15:00] <eutheria> i have things to do :)
[15:01] <eutheria> this is why i have a pfsense firewall these days and not iptables
[15:01] <eutheria> i went nuts and it is virtual pfsense firewall
[15:01] <husien> koolhead17: now i know what the problem.. its my VmWare..!!! damn.. lol :) .. ok i going to reinstall this first and reboot my pc... big thanks koolhead17...
[15:01] <koolhead17> hi robbiew
[15:01] <eutheria> ubuntu is my ldap/dns and db servers
[15:01] <robbiew> hey koolhead17
[15:01] <koolhead17> husien: your welcome :)
[15:02] <koolhead17> eutheria: whats final verdict then :D
[15:02] <Daviey> zul: not as yet, can you resend me the details pls?
[15:02] <eutheria> koolhead17, well it seems to do a good job, you get a daily, weekly and monthly folders
[15:03] <eutheria> ok i am going a little nuts using duplicity to make a backup of this data on top
[15:03] <zul> Daviey: sure lp:~ubuntu-server-dev/{nova,keystone}/debian-merge
[15:03] <eutheria> but i think what i will do with duplicity is have only 1 weeks worth of backups
[15:04] <Daviey> ta
[15:04] <eutheria> and this way i have 1 months worth of mysql dumps
[15:46] <uvirtbot`> New bug: #937869 in mysql-dfsg-5.1 (universe) "5.1.x security update tracking bug" [Undecided,In progress] https://launchpad.net/bugs/937869
[16:33] <uvirtbot`> New bug: #932466 in glance "glance db migrations deadlock against precise mysql server" [Low,Fix committed] https://launchpad.net/bugs/932466
[16:39] <smoser> Daviey, re bug 924375, i think i'm waiting on something from maas team on what i'm supposed to take in.
[16:39] <uvirtbot`> Launchpad bug 924375 in cloud-init "cloud-init should allow pre-seeding of ec2 datasource:Ec2:metadata_urls" [Medium,New] https://launchpad.net/bugs/924375
[16:39] <smoser> pre-seeding ec2 data source might be generally useful, but wouldn't be used.
[16:40] <smoser> the maas one is what is important (and there is nothing for maas now).
[16:46] <Daviey> smoser: just a url?
[16:46] <smoser> maas will need cloud-init to do oath i think
[16:46] <Daviey> a fully qualified url, not wildcard?
[16:46] <Daviey> smoser: oh, bollox
[16:46] <Daviey> smoser: ok, let me get back to you.
[16:48] <jMCg> Hey folks, I'm failing to build the latest qemu-kvm package (deb-src from precise) on oneiric: http://dpaste.com/706381/ -- anyone know a good reason why this is failing -- or the right channel to ask this?
[16:51] <jMCg> w00t. https://wiki.ubuntu.com/IRC/ChannelList?highlight=%28CategoryIRC%29 >> I guess #ubuntu-packaging
[16:54] <hallyn> jMCg: did you 'debian/rules build' first?  looks like the configure step wasn't done, so i suspect not
[16:54] <smb> SpamapS, Daviey About that CEPH thing: nobody from the kernel team had been aware. And given its feature freeze by now there probably needs to be a very good argument to get anything done, if it is required.
[16:55] <jMCg> hallyn: ACK.
[16:55] <SpamapS> smb: no changes are needed
[16:56] <SpamapS> smb: the large xattr stuff is an optimization
[16:56] <smb> SpamapS, Ah, ok. Leann is crafting an email to clearify things. But it sounds as those then get at least postponed
[16:56] <SpamapS> smb: the better optimization is to just use BTRFS, which is what upstream recommends.
[16:57] <SpamapS> smb: yeah, lets make sure it gets on the TODO for precise+1
[16:57] <smb> SpamapS, And assigned to something that we are aware off :)
[17:04] <BrixSat> hello i have seted up debian on a hp server with raid1 by software and after all the install i dont get grub :( (no boot at all)
[17:04] <BrixSat> *ubuntu 64bits :)
[17:37] <koolhead17> so new version of apache server releases :)
[17:44] <goddard> how can you test the connection speed of your server?
[17:44] <goddard> realiably :D
[17:44] <goddard> if they say you have a 100Mbit connection when can you call them a liar ?
[17:49] <andol> goddard: Well, first of all they really have to define to at what point they promise 100Mbit...
[17:51] <uvirtbot`> New bug: #937987 in squid3 (main) "squid is loosing cache data when restarting" [Undecided,New] https://launchpad.net/bugs/937987
[18:04] <goddard> andol: well first i have to test the speed because that is the funnest part
[18:15] <SpamapS> utlemming: hey are you aware of issues with the precise cloud images?
[18:15] <SpamapS> utlemming:  libc6-xen : PreDepends: libc6 (= 2.13-24ubuntu4) but 2.15-0ubuntu2 is to be installed
[18:16] <SpamapS> utlemming: looks like they maybe got created half-way through a glibc update?
[18:16] <utlemming> Spamaps: No, this is the first that I've seen of it.
[18:17]  * utlemming looks
[18:17] <utlemming> SpamapS: are you using yesterday's image or todays?
[18:18] <SpamapS> utlemming: checking
[18:18] <SpamapS> ami-5b4f9e32
[18:19] <SpamapS> in us-east-1
[18:20] <utlemming> that's the alhpa2 release image
[18:20] <SpamapS> interesting
[18:20] <utlemming> SpamapS: do you have a bug number on that?
[18:20] <SpamapS> we use http://uec-images.ubuntu.com/query/ in juju ..
[18:21] <SpamapS> utlemming: yes it was reported in juju first.. hang on
[18:21] <utlemming> what is the full query string that you're using?
[18:23] <SpamapS> utlemming: looking that up
[18:23] <SpamapS> utlemming: I think it might be released.. which would explain the issue
[18:23] <utlemming> SpamapS: ack
[18:24] <utlemming> SpamapS: that's what I am thinking
[18:24] <SpamapS> utlemming: is the alpha2 image known to be totally broken?
[18:24] <utlemming> SpamapS: not that I've seen
[18:25] <utlemming> what is the JuJu bug so I can get some context here
[18:25] <SpamapS> https://bugs.launchpad.net/juju/+bug/937889
[18:25] <uvirtbot`> Launchpad bug 937889 in juju "Hang on "juju status" with EC2 and Precise" [Undecided,New]
[18:26] <utlemming> SpamapS: logs?
[18:27] <SpamapS> utlemming: I've got an affected instance up
[18:28] <SpamapS> utlemming: imported your key on the instance I just msg'd you
[18:34] <uvirtbot`> New bug: #937941 in puppet (main) "Upstream have withdrawn 2.7.10 due to several regressions" [Undecided,New] https://launchpad.net/bugs/937941
[18:47] <utlemming> SpamapS: this looks like a problem with an interactive dialog that is triggered with an update/installation. I'm testing a fix right now
[18:51] <uvirtbot`> New bug: #938047 in mysql-5.1 (universe) "package mysql-server-5.1 5.1.54-1ubuntu4 failed to install/upgrade: subprocess installed post-installation script returned error exit status 1" [Undecided,New] https://launchpad.net/bugs/938047
[18:51] <sdferfx> Hello. I am trying to use Ubuntu AMIs to install Ubuntu 11.10, but I want to attach to my own EBS device. Can this happen? If not, how do I manually specify an ISO to EC2 for install?
[18:52] <Zermanno> Hi, i'm using top but i can't understand which time unit it use in the TIME column, are they seconds?
[18:55] <sdferfx> Zermanno, use htop
[18:55] <Zermanno> sdferfx, done! thanks!
[18:56] <patdk-wk> Zermanno, TIME+ is Min:Sec.secdecimals
[18:56] <Zermanno> tnx
[18:56]  * patdk-wk notes top and htop do the same
[18:57] <sdferfx> Yeah, htop wouldn't solve that question, it's just a lot better ;)
[18:57] <patdk-wk> prstat :)
[19:14] <uvirtbot`> New bug: #938044 in freeradius (main) "Unable to listen on IPv6" [Undecided,Invalid] https://launchpad.net/bugs/938044
[19:14] <SpamapS> utlemming: hm I wonder if install_packages installs recommends
[19:14]  * utlemming checks
[19:15] <Jeeves_> Sorry about #938044
[19:15] <Jeeves_> My bad :)
[19:20] <utlemming> SpamapS: the default for cloud images is to install recommends, but not suggests
[19:22] <SpamapS> utlemming: ok so that will be the same behavior.
[19:23] <SpamapS> utlemming: fix pending, looks pretty simple to change. :)
[19:23] <SpamapS> Heh. have to change 6 lines of code, and 10 tests
[19:24] <utlemming> SpamapS: a nice side effect is a faster boot...it looks like it'll shave two or three minutes off
[19:29] <SpamapS> utlemming: what will?
[19:29] <utlemming> SpamapS: consolidating the install of juju into the other packages list
[19:30] <SpamapS> utlemming: I want to believe you, but 2 - 3 minutes sounds extreme!
[19:30] <SpamapS> utlemming: is that because it happens in parallel with other stuff?
[19:31] <utlemming> I'm not sure...but in testing the fix, it was done quicker. Perhaps it was just me, but it was at least a minute or two.
[19:31] <uvirtbot`> New bug: #938078 in apache2 (main) "Problems with mod_rewrite since updating to 2.2.14-5ubuntu8.8 " [Undecided,New] https://launchpad.net/bugs/938078
[19:37] <SpamapS> utlemming: Alright, fix proposed. :)
[19:44] <zul> Daviey/rbasak: the nova console patch has been re-enabled
[20:02] <rbasak> zul: awesome - thanks!
[20:04] <Daviey> zul: cool
[20:21] <RoyK> !xen
[20:22] <jfbouchard> Hello, I have an issue with Ubuntu 10.04 + a guess Windows XP 32 bits in KVM. Install goes fine but when I start the OS, I get a nice black screen and the CPU goes to 100%
[20:22] <jfbouchard> After research online, there is a lot of occurance of this. Note that I don't find any valid error message. any idea ?
[20:24] <cyberdyn> can anyone point me to some docs or offer some suggestions on how to troubleshoot why my ubuntu 11.10 server (new install) isn't sending mail? / I've installed exim according to linode docs..: http://library.linode.com/email/exim/send-only-mta-ubuntu-11.04-natty
[20:25] <cyberdyn> i try to send mail with echo "This is a test." | mail -s Testing address@domain.com / and it doesn't get delivered.
[20:30] <RoyK> that xen article needs some cleanup...
[20:40] <stgraber> hallyn, jjohansen: Do you have an ETA for the new kernel/apparmor fixing the current LXC issues? I was trying to figure out a way of dealing with locales in containers and noticed that localegen seems to be blocked by apparmor here for no good reason
[20:41] <jjohansen> stgraber: I am going to send the pull request today
[20:42] <jjohansen> stgraber: I can build you a kernel with it if you would like, it will probably take a day or two to land
[20:43] <stgraber> jjohansen: I can wait for it to land if we're so close to having it in. I started deploying LXC on more and more of my servers and so far localegen and udev upgrades are the only two cases where I noticed the issue and I have workarounds for both
[20:44] <jjohansen> stgraber: okay, but if you change your mind just ping me
[20:44] <stgraber> hallyn: btw, the easiest way to deal with locales will probably be to do "chroot $ROOTFS /usr/share/locales/install-language-pack $LANG"
[20:45] <stgraber> hallyn: and probably messing with /etc/default/locale a little too (well, copying it should be the easiest)
[20:51] <hallyn> stgraber: heh, thanks, pinging jjohansen was on my todo list
[20:52] <hallyn> stgraber: ok, let me know if you want me to touch the locale stuff, else i'm assuming you'll do it if you feel it's needed
[20:53] <stgraber> hallyn: I'm not 100% sure it'll be needed yet, apparmor preventing localegen might be preventing it from just working. Anyway I'm getting error messages every time I login over ssh, so that's a good reminder that I need to deal with it ;)
[21:06] <lifeless> hallyn: hey, did you see my issue with mountall in backscroll ?
[21:06] <lifeless> hallyn: should I file a bug ?
[21:07] <hallyn> lifeless: i id not
[21:08] <hallyn> lifeless: I think that's basically un-fixable for lucid;  fixed for precise.
[21:08] <hallyn> stgraber may have a bug (fix released) that that's a dup of, even.
[21:08] <lifeless> hallyn: so, while I need a lucid lxc, I'm boned ?
[21:09] <lifeless> hallyn: (my host is precise)
[21:09] <hallyn> you can work around it, but we can't just make it work with no extra work
[21:09] <hallyn> simplest workarounds: (a) do it in kvm;  (b) just umount /lib/init/fstab before doing the upgrade
[21:10] <lifeless> b works for me ;)
[21:10] <hallyn> it's very unfortunte
[21:10] <hallyn> similar bugs happen when an upgrade tries to create /dev/console or somesuch
[21:10] <stgraber> lifeless: yeah, if you try to upgrade mountall, udev or makedev in a container your update will crash
[21:11] <stgraber> lifeless: for mountall, unmounting /lib/fstab is enough, for udev/makedev, the easiest is to shutdown the container, then chroot into it and upgrade, then boot the container again
[21:11] <stgraber> all of these are fixed in Precise but need changes inside the container that we can't really sru/backport
[21:14] <lifeless> thabks
[21:15] <hallyn> stgraber: it's not purty, but i've got about 6 pages of pdf (latex source) for a start to the lxc server guide.  let me know if you care to look over it, else i'll assume you'd rather gauge your eyes out (and not be offended).
[21:18] <stgraber> hallyn: sure I can certainly find some time to look at it
[21:19] <hallyn> ok, thanks.  just trying to decide how much to say about libvirt-lxc
[21:30] <hallyn> stgraber: hm, maybe it's too early to get feedback.  But http://people.canonical.com/~serge/lxc.serverguide.pdf
[21:34] <rbasak> hallyn: mind if I take a look? I've been setting up a precise server with lxc :)
[21:34] <hallyn> rbasak: absolutely!
[21:35] <rbasak> thanks!
[21:35] <hallyn> rbasak: if you want the latex source to make notes in, it's at http://people.canonical.com/~serge/lxc.serverguide.tex
[21:36] <rbasak> thank you!
[21:37] <hallyn> (jsut re-synced it;  i won't push to there again for now in the hopes of keeping your copies in sync :)
[21:39] <hallyn> (note - Daviey said he had someone who was going to polish it later, so i'm just trying to get all the content in and accurate, nto going for polish or eloquence)
[21:40] <rbasak> hallyn: so I had a possible issue yesterday - I have a btrfs root, but lxc-create -ing new containers seems to use space every time. Is this expected?
[21:41] <hallyn> as in, full rootfs space?
[21:41] <rbasak> yes
[21:42] <hallyn> not expected, unless btrfs strangely reports
[21:42] <hallyn> what does 'btrfs subvolume list' show?
[21:42] <hallyn> (or whatever command shows subvolume info)
[21:43] <hallyn> (note smoser coded and tested it so he might be better able to answer;  but it sounds wrong to me)
[21:43] <rbasak> that does suggest that it's working
[21:44] <hallyn> rbasak: no wait
[21:44] <rbasak> but my available space goes down by 300 megs each time
[21:44] <hallyn> no, no, no.  i don't know if i dropped a piece of smoser's work, but that is just not done right
[21:45] <hallyn> it creates teh btrfs subvolume, then still calls the lxc-ubuntu template to populate it with rsync
[21:45] <rbasak> that would explain what I'm seeing
[21:45] <hallyn> hm, though rsycn should see the same dates and leave the data alone
[21:46] <stgraber> hallyn: http://paste.ubuntu.com/851895/
[21:47] <hallyn> stgraber: thanks
[21:47] <rbasak> "For rapid provisioning, you may wish to customize a “canonical” container..." -- I like what you did there :-)
[21:47] <hallyn> i wanted to slap myself, but it's the right word to use there :)
[21:48] <stgraber> hallyn: I guess we can also mention the crazy architecture simulation stuff somewhere in there as an advanced feature ;)
[21:48] <hallyn> stgraber: hm, yeah
[21:48] <hallyn> at least, just mention --arch for creation
[21:49] <hallyn> hopefully it all works smoothly enough that we don't need to say any more about how it works
[21:49] <hallyn> (if we wanted to write a LISA paper, it'd be a cool section to add)
[21:49] <stgraber> hallyn: yeah, mentioning -a would be good and mentioning that if qemu-user-static is installed, then any supported architecture should work whatever your CPU actually is
[21:50] <hallyn> ok
[21:50] <rbasak> I haven't tried that feature yet but I think it'll be really useful for me. Thanks for that!
[21:50] <hallyn> at some point i'll add the other templates (and their limitations) as well.  but not this week probably
[21:54] <rbasak> hallyn: the documentation is great - thanks for that.
[21:55] <rbasak> hallyn: for networking, I'm interested in also connecting lxc to a bridge that I manage, and separately I wonder if I can get an lxc machine to share the same ethernet interface as the host? So that they can bind to different ports but on the same IP.
[21:56] <hallyn> (i've been typing so much on obsucre commands i rarely use that i forget what is actually in there, but) i do want how to use an existing bridge to be covered in teh guide...
[21:56] <rbasak> hallyn: lxc-start surprised me by grabbing my terminal; I expected it to daemonize (like virsh start does). This is a pain because I then had to look up the command to stop it in a separate terminal. Might be worth documenting
[21:56] <hallyn> different ports on same ip, that's not doable
[21:56] <rbasak> Understood :-(
[21:57] <hallyn> d'oh, yes, -d was on my list.  i must have deleted accidentally
[21:57] <hallyn> well, no, wait
[21:57] <hallyn> you *can* do that,
[21:57] <hallyn> just don't give the container a network definition at all, and it will share the host's network.
[21:57] <hallyn> i.e. leave out the 'lxc.network.type=veth' part of the config file
[21:57] <rbasak> oooh, awesome, thanks!
[21:57] <hallyn> belongs in there too, thanks :)
[21:58] <_KaszpiR_> hi
[21:58] <rbasak> that's my real use case - I have a VPS and IPv4 addresses are limited, but I would like separation between different things the server will do
[21:59] <rbasak> I'm curious about what lxcguest did, but I don't suppose that's particularly important.
[22:00] <rbasak> I hope that's useful feedback. The document really helped me though - I couldn't find this information concisely last week, and I think this will be really useful.
[22:00] <hallyn> it did things like send out fake net-device-up for lo, mount an empty /lib/init/fstab so that moutnall doesn't mess up the host, and start a getty on /dev/console
[22:00] <_KaszpiR_> I got a question, anyone got idea how to minimize server power usage? i'd like to make a home server that would not hod much pwoer if possible
[22:00] <_KaszpiR_> *hog
[22:00] <hallyn> (lxcguest ^)  but yeah i'm not sure i want that in the doc
[22:01] <hallyn> rbasak: great, thanks.  i sort of feel like it's too long to be useful, but obviously it's still incomplete.  thanksf or the comments.
[22:06] <rbasak> hallyn: for this kind of document, IMO long is not a problem, comprehensiveness is good - there should exist a document like that.
 makes sense
[22:44] <fluvvell> if anyone is familiar with dkim on postfix, can they suggest what "fatal: 8891:localhost: valid hostname or network address required" might mean in the mail log.
[23:11] <wmp> hello, anybody know how to disable access to PCI and lshw,lspci to user?