[00:46] <phillw> hi, anyone here for 12.10?
[00:47] <phillw> dns-nameservers is no longer an inbuilt command
[01:18] <disown> I have to say I don't like the new design of ubuntu that you have to search for a terminal and your applications like a calculator or gedit ,...what was there target audience for this one for... And linus thinks linux will never over take microsoft or get as big as microsoft ... well not by this crappy design ... I will definitely uses a different desktop manager or at least I will have to look more deeply into if you can config
[01:18] <disown> ure gnome or the stupid sidebars away... Temp fix just add the newer repo's to the source.lst because the os's is fine the way it was... Another way I could do it is just make my own or uses a different distro other then ubuntu at this point each is pretty easy. Though the more important thing ubuntu let me down :(
[01:19] <disown> after all the time I prefered it over the rest :(
[01:20] <disown> I guess I will go back to knoppix or fedora or suse,...etc which is ok but I always had a crush on 10.4 lucid :(
[01:26] <disown> even the scroll bar by default is tiny looks ugly well I suppose I can keep using the newest kernel with the older gnome or kde windows managers / desktop managers. In theory I don't see a problem with that... its essentially runing an old userland program on a new kernel
[01:27] <lunaphyte> hi.  i have a computer with two network interfaces.  i'm using one for the os on the computer itself, with a traditional static ip address, and the other for kvm guests, so it doesn't have an ip address, as far as the host os is concerned.  i'd like to bring up this interface, such that it's "up", but with no ip address, but i'm not quite sure what the appropriate method is to do this.
[02:16] <yeats> lunaphyte: you'd probably do well to read this: https://help.ubuntu.com/community/KVM/Networking/
[02:17] <lunaphyte> thanks, i have, actually.  i'm using macvtap though, not traditional bridging - so i just need to interface "up" - but without an address.
[02:18] <lunaphyte> ifconfig eth1 0.0.0.0 seems to do this, but i don't know if that's really the best way of going about that.
[02:19] <yeats> I've only done KVM with bridged networking, so I'm not familiar with what you're attempting to do
[02:19] <lunaphyte> well, tbh, it's not really got much to do with kvm.
[02:20] <mattcen> Hi all. I've downloaded a new LSI MegaRAID driver (http://www.lsi.com/downloads/Public/MegaRAID%20Common%20Files/Ubuntu_10.04_LTS_05.30.zip) and as far as I can tell, have installed it via DKMS, but can't get it to load as a priority over the Ubuntu-supplied (out-dated) module.  Anybody know how I can diagnose this?
[02:20] <lunaphyte> really just a question about bringing up an interface without an address, regardless of what it might ultimately be use for.
[02:20] <lunaphyte> *used for
[02:20] <yeats> lunaphyte: as far as I know, any interface has to have an IP
[02:20] <mattcen> lunaphyte: If you give me some more context, I may be able to help.
[02:21] <mattcen> yeats: incorrect. Set "iface eth0 inet manual" in your /etc/network/interfaces, and you can ifup without adding *any* IP/subnet info.
[02:21] <mattcen> (Or you can bring it up manually without ifupdown using 'ip link set dev eth0 up')
[02:21] <disown> O so they switch over to the crappy stuff in versions 11 and 12 so oneiric had this that I am using now around when the kernels went from the 2.6's into the 3 darn .
[02:21] <yeats> mattcen: thanks for the info
[02:21] <disown> And another thing is the lsb-release command doesn't even work well you shouldn't even needed it since cat /etc/lsb-release does fine by itself
[02:21] <disown> They should really just incorporate this information with a switch in uname or something but ok
[02:22] <lunaphyte> here's sort of an illustration: http://dpaste.com/785921/
[02:22] <lunaphyte> you can see that by doing ifconfig eth1 0.0.0.0 it brings the interface up, but without an ip address.
[02:22] <ScottK> SpamapS: I'd appreciate it if you could take a look at http://tanghus.net/2012/03/yet-another-mysql-vs-apparmor-barf/#comment-299 and see if there's an issue there.
[02:23] <lunaphyte> that's what i want, but i'm not certain if ifconfig [...] 0.0.0.0 is really the right way to do it, and i'm not certain how to have the system do it automatically at boot.
[02:23] <mattcen> yeats: np. A bit more info: http://www.cyber.com.au/~mattcen/Preferences/doc/networking/interfaces See the br0 interface. That stanza will work fine for a standard eth* iface if you remove the bridge_ lines.
[02:24] <mattcen> lunaphyte: See that link I just posted.  Use the 'manual' stanza, with (probably) no other options
[02:25]  * mattcen has become distracted from his original question. Any input on that would be greatly appreciated.
[02:25] <lunaphyte> that's what i have now: http://dpaste.com/785923/
[02:25] <lunaphyte> but the system doesn't bring up eth1
[02:25] <mattcen> O rly? Interesting.
[02:26]  * mattcen tests.
[02:26] <mattcen> lunaphyte: What Ubuntu release are you running?
[02:26] <lunaphyte> 12.04
[02:27] <mattcen> lunaphyte: Ah!
[02:27] <mattcen> I know what you need
[02:27] <mattcen>   up   ip link    set   dev $IFACE up
[02:27] <mattcen>   down ip link    set   dev $IFACE down
[02:27] <mattcen> (See my link)
[02:28] <lunaphyte> ah, i see.  i'll give that a shot, thanks.
[02:28] <mattcen> np.
[02:29] <lunaphyte> is $IFACE an actual variable i can use, or is that just for illustrative purposes in your example?
[02:30] <lunaphyte> i guess i can answer that question myself. :)
[02:32] <mattcen> lunaphyte: Ah sorry. No, $IFACE is treated specially by interfaces(5)
[02:32] <mattcen> It is replaced by the name of the interface inside which the stanza is found (e.g. eth0)
[02:33] <mattcen> lunaphyte: That entire file that I posted is perfectly valid in every way (though parts of it have dependencies on things like my wpa_supplicant.conf and wvdial.conf)
[03:53] <mattcen> Nobody have ideas about my DKMS issue?
[04:00] <KM0201> !scp
[06:09] <SpamapS> ScottK: ACK. There was a dh_apparmor bug and I'm not 100% sure its been fixed.
[06:51] <SpamapS> ScottK: that was probably bug 986892 btw
[09:40] <michelelv> ciao
[09:40] <michelelv> !list
[11:21] <jamespage> zul: around? just looking at the FTBFS for requests - not sure there is much point running the test suite
[11:22] <koolhead11> jamespage: can i get some help with his https://bugs.launchpad.net/juju/+bug/897645
[11:24] <jibel> todays quantal server with LVM fails to reboot after installation
[11:24] <jibel> bug 1036612
[11:24] <xnox> jibel: interesting. there was another ~ report in #ubuntu-arm reported ogra_
[11:25] <jamespage> koolhead11, what help are you looking for?
[11:26] <jamespage> jibel, nice
[11:26] <jamespage> xnox, any ideas?
[11:27] <jibel> I haven't found any recent upload to Quantal that could cause this excepted the kernel itself
[11:27] <jibel> linux 3.5.0-10.10 was uploaded last night
[11:28] <xnox> same thoughts. I didn't touch lvm, nor those parts of installer, and I don't think cjwatson touched anything user space either.
[11:28] <xnox> jibel: poke kernel people? =)
[11:29] <jibel> xnox, yep, I'm comparing with alternate, which doesn't fail and runs the same test.
[11:31] <xnox> jibel: it would be nice to have a diff of packages version numbers between last-good and failure runs.
[11:31] <xnox> that kind of narrows down the search ;-)
[11:31] <jamespage> xnox, I think you can pull that from the ISO images - pitti had something that did that from memory
[11:32] <jamespage> jibel, does the alternate image use the squashfs/or whatever based approach that server does now?
[11:33] <Daviey> jamespage: no
[11:33] <Daviey> (afaik)
[11:35] <jibel> jamespage, no
[11:36] <jibel> I'm diffing the manifests
[11:36] <jamespage> jibel, might be something in the image
[11:41] <jibel> hm, it boots with a 3.5.0-9 but the initrd generated during install is for 3.5.0-10
[11:42] <jamespage> jibel, is it some sort of sync issue between the squashfs image and the initrd?
[11:43] <jamespage> zul: disabled the tests - 10/11 of them require network connectivity so break in the buildds
[11:44] <jibel> jamespage, it could be, but I don't know enough to be sure.
[11:48] <Malcor|Work> Hello
[11:52] <Malcor|Work> Could I have some help please?
[11:53] <koolhead11> jamespage: is there a way out to define http_proxy during bootstrap
[11:53] <jamespage> koolhead11, other that what mgz outlined in #juju no
[11:54] <koolhead11> jamespage: yes catching up on hos suggestion. let me see
[11:54] <jamespage> Malcor|Work, just ask your question...
[11:59] <Malcor|Work> I am trying to set up a raid on the old server at work, the previous version was ubuntu 9 I think. I am trying to delete the MD Devices and restart the RAID configuration but I am not sure how to unmount the RAID devices. I am in the Ubuntu server installer at the moment. 10.04 btw
[12:01] <xnox> Malcor|Work: go into manual partitionair, entel setup raid devices, it will have an option to remove them.
[12:02] <Malcor|Work> xnox, I have tried that but I get "There was an error deleting the software RAID device. It may be in use."
[12:03] <xnox> are there any raid devices listed in the manual partitionaire?
[12:03] <xnox> enter each one and mark to not use
[12:05] <Malcor|Work> The device itself I cannot enter but nested ti them I can geter the harddrives and they are set to do not use. Do I have to do that to the devices in the SCSI?
[12:12] <Malcor|Work> there are two raid devices in the manual partitionaire. All of them are FREE SPACE
[12:12] <zul> jamespage: agreed
[12:22] <Malcor|Work> I have taken a picture, just uploading.
[12:24] <Malcor|Work> https://dl.dropbox.com/u/4777114/IMG_20120814_131750.jpg
[12:25] <Malcor|Work> oops, sorry still uplkoading ;-(
[12:25] <Malcor|Work> There link works now
[12:27] <Malcor|Work> Done it
[12:27] <Malcor|Work> I just undo changes and it worked. Thanks anyways
[12:34] <hallyn> jjohansen: hey, did you ever get a chance to look into the apparmor+no_new_privs bug?
[12:35] <xnox> Malcor|Work: you could boot into e.g. livecd and zero-out superblocks with $ mdadm --zero-superblock
[12:35] <xnox> on them
[12:35] <xnox> danger, you will loose all data
[12:37] <Malcor|Work> I have sorted for now xnox but I will remember that for a few months time when I break another raid controler lol
[12:37] <xnox> Malcor|Work: ok cool.
[12:37] <Malcor|Work> Thank you
[13:08] <ScottK> SpamapS: Thanks for looking into it.
[13:49] <Daviey> hallyn: hey, do you have thoughts on bug 985489?
[13:56] <hallyn> Daviey: no - i've seen no bugs report about virsh list hanging (without all of libvirtd being hung), so maybe it's a python binding bug in libvirt, or a usage bug in nova...
[14:00] <Daviey> hallyn: Have any ideas on how we can progress it?
[14:03] <hallyn> code review of the relevant nova code; set up a test grid with debugging enabled (both in /etc/libvirt/libvirtd.conf and installing libvirt and nova debug symbols), and separately (on a single host) trying to pound libvirt with virsh list's looking for a hang
[14:06] <hallyn> Daviey: i think we need an idea of the size and traffic required to reproduce this, but let me look through the libvirt git changelog for relevant commits
[14:08] <Daviey> hallyn: appreciated !
[14:09] <stgraber> hallyn: are you planning on getting the include change into quantal's lxc this week?
[14:09] <stgraber> hallyn: I think it's the last bit we want before rebasing the API branch on quantal and getting that uploaded before feature freeze
[14:15] <krneki> is there posiblle to boot from raid5? or you have to do raid1 boot partition
[14:18] <patdk-wk> you can boot raid5
[14:20] <krneki> patdk-wk: so i set up raid 5 with lvm root swap and home. this is fresh installation. so you are saying that this should work? or is there bug like in the older grub version that it couldnt recognize raid5?
[14:21] <_ruben> /boot ideally should be not on top of lvm
[14:21] <_ruben> / can be tho
[14:21] <xnox> krneki: you should have a separate boot partition.
[14:22] <krneki> by separate you mean put on raid 1 /boot? :)
[14:22] <patdk-wk> there is no *bug* about older grub not *supporting* raid5
[14:22] <patdk-wk> that would be a feature :)
[14:23] <RoyK> patdk-wk: iirc the latest grub2 supports it, but I don't think it's trivial to setup
[14:23] <patdk-wk> ya, never really tried to set it up
[14:24] <patdk-wk> just seems like, begging for trouble, to me
[14:24] <hallyn> stgraber: oh did i not do that yet?  i thought i had
[14:24] <RoyK> krneki: for my home server, I've just put the root on a separate disk. for something with higher demands for uptime, I'd use a mirror for the root, boot, swap etc
[14:25] <hallyn> stgraber: sure, i can push both that and the lxc-start-ephemeral rootfs workaround
[14:25] <hallyn> the include file patch is so minimal after all :)
[14:25] <krneki> yeah but i m using raid 5 for i while.. and i was happy to see that now it supportes raid5. now when i read some more i m not happy anymore :P
[14:26] <stgraber> hallyn: I don't remember seeing an upload, but I might just have missed it. I'm not using quantal ;)
[14:26] <krneki> i will just do raid 1 for system and then storage raid 5
[14:26] <RoyK> krneki: what has stopped supporting raid5?
[14:26] <RoyK> krneki: afaik grub never did support it, at least not until recently
[14:27] <krneki> RoyK: before i was installing boot on raid 1 and system on raid 5
[14:27] <RoyK> krneki: that's still supported
[14:27] <RoyK> krneki: just can't boot off raid-5, that's all
[14:28]  * RoyK still prefers a dedicated system drive or mirror and separate data area on raid-[56]
[14:29] <hallyn> stgraber: no, i guess it's not in there.  oops.
[14:29] <hallyn> i'd like to get a userns kernel built today though, so i'll do lxc package tomorrow, unless there's urgency?  when is FF?
[14:32] <stgraber> hallyn: FF is on Thursday next week, so as long as we get the include change in quantal this week, I should have enough time to do the rebase and some testing
[14:41] <hallyn> stgraber: oh, sorry, right you need to rebase
[14:41] <hallyn> stgraber: i'll get it pushed today.  ttyl
[15:19] <hallyn> stgraber: http://people.canonical.com/~serge/lxc-include.debdiff tests fine for me.  i'll push in a bit.
[15:19] <hallyn> bleh, suppose lxc.conf(5) should get an update
[15:20] <stgraber> hallyn: diff looks good. I guess we'll wnat to SRU the ephemeral part of it (with whatever else is currently pending SRU, I seem to remember us having a few of these)
[15:21] <hallyn> yeah probably should
[15:25] <jamespage> utlemming, please can you add an appropriate comment to bug 1014864
[15:26] <ssvss> Hello, any idea on how ubuntu server works in a mac mini ?
[15:26] <ssvss> are there any issues with power management
[15:30] <hallyn> stgraber: updated http://people.canonical.com/~serge/lxc-include.debdiff to update the manpage.  (i do notice there is no entry for lxc.seccomp, but that's more of a tech preview thing ...
[15:37] <hallyn> pushed
[16:11] <jamespage> hggdh: are you likely to implement multi-nic and usb features in the current ISO testing framework still?  sounds unlikely to me....
[16:11] <smartboyhw> jamespage: Is it that you're in charge of testing?
[16:12] <jamespage> smartboyhw, well coordinating at least for server
[16:13] <smartboyhw> OK, I wanna help
[16:13] <jamespage> smartboyhw, great!
[16:13] <jamespage> hang on in the meeting - more info later on
[16:14] <smartboyhw> What? I'm going to sleep!
[16:14] <jamespage> about 15mins probably
[16:15] <smartboyhw> I'm Ubuntu Studio member in testing, also helped chairing QA meetings
[16:15] <jamespage> smartboyhw, how would you like to help?  we are due to migrate to a new automated testing framework this cycle; expanding both the depth and breath of tests will be important
[16:15] <hggdh> jamespage: I dont think it will be an easy change either. but we could add multi-nic to the preseed/libvirt config
[16:15] <jamespage> hggdh, I just don't really see the value considering its officially deprecated....
[16:15] <smartboyhw> Testing!
[16:16] <hggdh> jamespage: we can discuss that later, I just got nuclearbob to add multi-disk support for VMs on UTAH
[16:17] <hggdh> (needed for kernel testing)
[16:17] <jamespage> hggdh, also needed for RAID testing :-)
[16:17] <hggdh> heh, well, yes, there is that also
[16:20] <smartboyhw> jamespage: I really can't stay for 15 minutes
[16:20] <jamespage> smartboyhw, sure - I'll catch you tomorrow then instead - go to sleep!
[16:21] <utlemming> smoser: I can do #1028503...that is kind-of important for Azure
[16:21] <smartboyhw> jamespage: tmr 14:00GMT..I will be there
[16:22] <smoser> bug 1028503
[16:22] <smoser> why is it important for azure (as opposed to anywhere else)? I'm just curious.
[16:23] <smoser> but feel free to work on it.
[16:23] <utlemming> smoser: well, its important for azure in a backwards way. Right now we have to do some massaging of the images for Azure, and one is to delete the ubuntu user per the cloud vendor's rules. So if we have cloud-init do this work, then it means that we can cut out a couple steps
[16:23] <smoser> please start with wriging out what you think the config looks like.
[16:24] <smoser> i'd like to support adding multiple users, there is a request to support 'None' also.
[16:24] <utlemming> smoser: ack, and I want to shoe in some stuff for sudo too
[16:51] <www2> hi can some one tell me how i can config apache that each virtual host get his own php.ini?
[16:52] <www2> note i have use suphp but it don't work for my current server config.
[16:56] <jamespage> adam_g, I just uploaded that openssl fix to -proposed so that it can be discussed further - but I think it will probably make it still
[17:12] <adam_g> jamespage: ya, i saw. thanks
[17:13] <jamespage> np
[17:17] <stgraber> hallyn: just saw the lxc uploaded to -proposed, what did you bundle in there?
[17:18] <hallyn> stgraber: fix that was staged (for 'stop lxc-net' possibly breaking) and the new one for lxc-start-ephemeral
[17:18] <hallyn> (nothing else was in lp:ubuntu/precise-proposed/lxc)
[17:19] <stgraber> hallyn: ok, I guess we could have cherry picked the lxc-start-ephemeral echo fix, but that's very low priority, can do that in the next upload
[17:19] <hallyn> what is taht?
[17:19] <hallyn> i diff'd the lxc-start-ephemeral against quantl's and saw only a whitespace diff
[17:19] <stgraber> the echo 'bla $something' that should have been echo "bla $something" in lxc-start-ephemeral
[17:20] <stgraber> that was fixed in  0.8.0~rc1-4ubuntu22
[17:20] <stgraber> though maybe 12.04 wasn't affected or we SRUed it already without marking it in the changelog
[17:21] <hallyn> oh.  now i see
[17:21] <hallyn> mayhaps i should change my font.   i couldn't spot the ' vs " in the diff
[17:23] <hallyn> stgraber: do you want to kick that package from the queue?
[17:24] <stgraber> hallyn: nah, it's really a cosmetic fix, we'll fix it in the next upload
[17:24] <hallyn> ok
[18:31] <ssvss> Hi, any suggestions on how I can create a snapshot of my ubuntu server. I am using this as a development server, and would like to easily start from fresh every once in a while. and need a easy way to do it.
[18:42] <andol> ssvss: physical server or a virtual one?
[18:48] <RoyK> ssvss: for a physical server, you need a filesystem with snapshotting, or you can use lvm. The only current somewhat supported filesystem with snapshop capabilities, is btrfs, and lvm slows down the system if you have many snapshots. zfs may do as well, but not for system things, and under fuse (which is default), it's rather on the slow side on writes
[18:50] <ssvss> physical server
[18:51] <andol> ssvss: Hmm, if you want to have convenient snapshots/restores, a virtual devel server would really make things so much easier.
[18:51] <RoyK> then probably vlm snaps should do
[18:51] <RoyK> lvm even
[18:52] <RoyK> or as andol says, use kvm to create a vm on which you do the development
[18:54] <RoyK> btw, kvm snapshots aren't very flexible either, at least not on lucid, guess I'll have to check that better
[18:55] <ssvss> I prefer the physical server. I run some performance tests in this machine, so would like it to be as close as possible to the production server. I will checkout the lvm snapshot.
[18:57] <RoyK> lvm snapshots are true "copy on write", the old way, meaning new writes are sent to lvm, and the old data moved away before the new data replaces them. with 1 snapshot, this doubles the number of I/O operations to the lv, with two, it triples, etc
[18:57] <RoyK> btrfs is designed to do this the "new" way, meaning only pointers are updated and the data is stored in new blocks. btrfs is not, however, flagged as stable
[18:58] <rbasak> I'm not sure snapshotting makes sense if you really want to test performance. Whatever snapshotting you do will affect IO speeds.
[18:58] <RoyK> rbasak: with the btrfs or zfs way, it works well, with lvm it doesn't
[18:58] <xnox> RoyK: ssvss: with lvm snapshots you can cheat =) boot into the snapshot. Then writes have no penalty. You can easily drop them, or commit to preserve.
[18:59] <rbasak> RoyK: with btrfs and zfs you're changing the filesystem will will affect IO performance too!
[18:59] <xnox> ssvss: you can use rsnapshot backups onto a different drive / locations. and restore if need be.
[19:00] <RoyK> rbasak: even with ext2, you're changing the filesystem whenever you update a file :þ
[19:00] <xnox> ssvss: with backups you have 'production' performance
[19:00] <rbasak> RoyK: use dump/restore if you really care about that :-P
[19:01] <RoyK> xnox: with zfs, you have 'production' performance all the time ;)
[19:01] <RoyK> too bad the native zfs port isn't ready (yet)
[19:06] <ssvss> changing the file system is not a option, since this has to be close to production system. I will look in to rsnapshot or lvm booting into snapshot cheat. I think rsnapshot backup will suit better for my need. Thanks for the suggestion guys !!
[19:07] <RoyK> rnsapshot is a cool hack
[19:22] <xnox> RoyK: zfs latest development is closed source in oracle, and everyone who did zfs/btrfs left oracle....
[19:23] <RoyK> xnox: well, latest zfs development is in illumos ;)
[19:25] <nandersson> Hi, I am doing unattended installations and I had a look at how Ubuntus metapackages are build up. It looks like the packages in ubuntu-minimal aren't part of ubuntu-standard, and neither ubuntu-minimal nor ubuntu-standard are part in ubuntu-desktop. Wouldn't that mean that if I use config "tasksel tasksel/first   multiselect ubuntu-desktop" I wouldn't get the packages in ubuntu-minimal or ubuntu-standard? I am confused.
[19:27] <nandersson> For example "iptables" is part of the "ubuntu-standard" meta-package. I haven't tested, but I think iptables is available in the system even though I write "tasksel tasksel/first   multiselect ubuntu-desktop"
[22:18] <three18ti> is there a way to do a nqa install from usb?
[22:22] <three18ti> nqa = no questions asked.  I know you can from cobbler, but I'd like to make a usb stick I can just boot and walk away.
[22:36] <three18ti> preseeding is the term
[22:54] <xnox> three18ti: presseding and unattended install
[22:55] <xnox> check either ubiquity preseeding or debian-install preseeding depending on the image you use
[22:55] <UndiFineD> hello, has anything been changed to apache2 in 12.04 ? I used to have my local mirror available but now it shows me somewhat near ubuntu marked up 404 page
[22:57] <three18ti> xnox, cool thanks.
[23:31] <adam_g> zul: still around?
[23:33] <zul> adam_g: almost
[23:33] <zul> adam_g: whats up?
[23:33] <adam_g> zul: first time trying this, but on quantal: python-cinderclient : Depends: python2.7-cinderclient but it is not installable
[23:33] <adam_g> zul: any hint?
[23:34] <zul> adam_g: hmm....have a look at python-cliff
[23:39] <adam_g> zul: cool, thanks
[23:53] <kermit> how is it that the pid, even with -p, is missing from a LISTEN socket in netstat?