[00:03] <alaska13> I'm having trouble with setting up a l2tp/ipsec-psk server on 11.04, client connections seem to fail in phase 2 of ipsec connection.  Any suggestions?
[00:22] <samtheman> hi guys\
[00:23] <samtheman> just need some ideas please
[00:23] <samtheman> i wanted to add more ubuntu servers to my network
[00:23] <samtheman> i have monitoring and all that
[00:23] <its> hello
[00:24] <its> whats a good app for remote connections?
[00:24] <samtheman> any production servers you can suggest??
[00:24] <samtheman> rdesktop is a good and simple one
[00:24] <its> i have x11vnc installed right now...
[00:25] <samtheman> use it to remote to windows clients
[00:25] <samtheman> rdp client
[00:26] <its> what about mac?
[00:26] <samtheman> mmmm
[00:26] <samtheman> not sure about that one
[00:26] <SpamapS> samtheman: you mean what hardware do we recommend?
[00:26] <samtheman> never realy tried
[00:26] <samtheman> no
[00:26] <samtheman> like apps
[00:26] <samtheman> mmmm
[00:26] <SpamapS> its: what do you mean "remote connections" ?
[00:27] <its> like, screen sharing
[00:27] <samtheman> like i have nagios cacti
[00:27] <its> I want to be able to rack mount my server
[00:27] <its> and then remote into it using another machine
[00:27] <SpamapS> its: ssh is all you need. :)
[00:27] <samtheman> yeah ssh
[00:27] <its> yeah ssh
[00:28] <SpamapS> samtheman: ahh, well I think servers should probably use all that electricy they have going through them to solve problems .. what problems do you need to solve?
[00:28] <samtheman> i dont have any atm
[00:28] <samtheman> just trying to implement more ubuntu in the networ
[00:28] <samtheman> i was think instant messaging
[00:28] <its> more ubuntu? Thats odd.
[00:29] <samtheman> i dont like windows
[00:29] <samtheman> im working on network load balancing
[00:29] <samtheman> but that project is not due for ages
[00:30] <its> well i think linux is really good
[00:30] <samtheman> haproxy for that one ---- thats good one rite??
[00:30] <its> im super new to ubunter server edition
[00:30] <SpamapS> samtheman: again, thats like saying do you guys know where any nails are that I can hit with my hammer? .. more appropriate to find reasons to use nails. :)
[00:31] <samtheman> do you know how i can install webgui cms
[00:32] <samtheman> i thought that could be better than my knowldgetree document sharing
[00:33] <SpamapS> If you want a CMS, Drupal is the current powerhouse there
[00:35] <samtheman> it does intranet, document sharing stuff like that??
[00:35] <SpamapS> Its a CMS, so it... manages.. content. :)
[00:39] <patdk-lap> I always perfer the ones that self-manages :)
[00:42] <adam_g> RoAkSoAx: are you still around?
[00:47] <twb`> Will someone tell me why this watershed crap is still trying to run a vgchange -ay *12 hours* after a reboot?
[00:47] <twb`> http://paste.debian.net/129693/
[00:49] <SpamapS> twb`: thats run by /lib/udev/rules.d/85-lvm2.rules on my oneiric system..
[00:50] <twb`> Yeah
[00:50] <SpamapS> twb`: is it possible you have an intentionally not-activated pv somewhere?
[00:50] <twb`> I was going to blame that for locking my lvm layer, but apparently it's not his fault
[00:51] <twb`> root     20379  0.0  0.0  32208 15844 ?        D<Ls 08:18   0:00 lvcreate --quiet --snapshot --size 64G --name understudy /dev/omega/collectd
[00:51] <SpamapS> I don't recall exactly what sets ID_FS_TYPE .. but if its from the partition table.. maybe you have one thats marked inactive and has no defined vg
[00:51] <twb`> ^^ still stuck in D state
[00:51] <twb`> Stupid bot.
[00:51] <samtheman> sorry what are we talking about??
[00:51] <samtheman> watershed???
[00:52] <twb`> OK, the actual problem is that "lvs" and all other LVM commands act like this: http://paste.debian.net/129695/
[00:52] <twb`> This issue arose after deploying a backup solution that works by taking an LVM snapshot of each LV in turn, fscking the snapshot, mounting it, rsync the data off, then unmounting and lvremoving it.
[00:54] <SpamapS> twb`: I've used that scheme before and its pretty effective usually (as long as your daemons stop writing while you snapshot)
[00:54] <SpamapS> or all write atomically
[00:55] <twb`> I think the problem is collectd REALLY likes writing
[00:55] <SpamapS> twb`: so the lvcreate for thes napshot just sits there forever in disk wait?
[00:55] <twb`> And I didn't realize lvsnapshot needs the system to be quiescent
[00:55] <twb`> Yeah
[00:55] <SpamapS> yes thats true
[00:55] <twb`> 10:54 <twb> Evidence: http://paste.debian.net/129696/
[00:55] <twb`> On other systems I was calling "sync & sync & wait" before each lvcreate --snapshot IIRC
[00:56] <SpamapS> I've always found the handwaving people do about collectd's every-5-seconds-write-like-its-1999 behavior to be unconvincing :)
[00:56] <SpamapS> Its not that it has to be totally quiescent
[00:56] <twb`> SpamapS: do you use collectd yourself, and if so do you know how to make it be less write-happy?
[00:57] <twb`> SpamapS: I'm thinking either crank up the buffering options and/or deploy rrcached or whatever it's called
[00:57] <SpamapS> twb`: I've only evaluated it. I found the writing too intense for most light weight servers.
[00:57] <twb`> Yeah, I'm getting completely reamed here by its default config and about 20 servers sending it stats over UDP
[00:57] <SpamapS> the rrd caching daemon is a must
[00:57] <twb`> OK, I will make that a priority
[00:58] <twb`> (afk few minutes, talking to $boss)
[00:59] <SpamapS> twb`: I'd be curious to hear if your snapshots complete and how full they are right after they do complete. It shouldn't be as painful as you're describing, even with a tiny write every 5 seconds.
[01:03] <twb`> Let me show you an iostat
[01:03] <twb`> Hmm, that's actually not too bad
[01:03] <twb`> http://paste.debian.net/129697/
[01:04] <twb`> dm23 is /var/lib/collectd
[01:04] <twb`> I have seen that up around 6000 when everything else was more like 128.
[01:15] <twb`> SpamapS: OK explain this to me, I tried "lxc-freeze --name hugin" (the collectd container), and I also tried kill -STOP all the procs in that container, and the I/O write load is unchanged on that LV
[01:17] <twb`> OK what the shit, even after I stop the entire container and all its procs are gone, the I/O write load is still high
[01:20] <twb`> OK, *now* it's finally gone down
[01:20] <lifeless> dirty pages being flushed ?
[01:20] <twb`> The lvcreate --snapshot that was hung in D state is also gone, BUT I ran "lvs" and that's not in uninterruptable D state
[01:33] <RoAkSoAx> adam_g: im here
[01:33] <RoAkSoAx> what's up
[01:37] <twb`> I also noticed that I'm getting write amplification at the md layer (compare md1 with its underlying sda/b/c)
[01:37] <twb`> lucid's parted and kernel don't support --alignment optimal, but Debian Sid's does -- I'm wondering if I should have a flag day, turn everything off, and reprovision everything on top of partitions created from a sid live environment
[01:41] <samtheman> hi guys
[01:41] <samtheman> quick one
[01:41] <samtheman> theres a user here who is going to pay 5k for an app that changes pdf to flipbook
[01:42] <samtheman> is there a linux free app thats does the same??
[01:44] <twb`> samtheman: define "flipbook"
[01:45] <samtheman> so it looks like a book
[01:45] <samtheman> runs on flash i think\]
[01:45] <samtheman> u can turn pages
[02:00] <samtheman> anyone got any ideas please:)
[02:00] <Martyn> "Fried Oreo Sandwiches"
[02:00] <Martyn> There's a doozy of an idea for oyu
[02:01] <samtheman> thanx
[02:01] <Martyn> Also, here's a recycled one -- fried peanut butter, honey, and banana sandwiches
[02:01] <Martyn> "The Elvis"
[02:01] <Martyn> I think I'll convince the people at the Royal Caribe to make them this time :)
[02:05] <twb`> samtheman: sorry, I don't do flash.
[03:19] <T3CHKOMMIE> hey everyone, im trying to modify my grub menu. i made the changes i believe i needed. but the grub menu is not changing. any ideas?
[03:19] <T3CHKOMMIE> :(
[03:22] <ascheel> Can someone tell me why r8169 is still loading even though I've blacklisted it?  http://pastebin.com/5VRVwzci
[03:34] <stgraber> ascheel: try running "sudo update-initramfs -u" and then reboot
[03:43] <crass> hallyn: did you ever make a patch for https://bugs.launchpad.net/ubuntu/+source/libvirt/+bug/691590 ?
[03:46] <crass> hallyn: I see now that you made the patch, but did it ever get accepted upstream?
[04:02] <twb`> WTF?
[04:02] <twb`> "This filesystem will be automatically checked every 0 mounts or 0 days, whichever comes first.  Use tune2fs -c or -i to override."
[04:03] <twb`> Gee, thanks, mke2fs.
[04:03] <twb`> This is why we can't have nice things
[04:23] <jmarsden> twb`: That just means someone set the fs to not do mount checks.  setting to 0 or -1 disables them, IIRC.
[04:25] <twb`> I just made it
[04:25] <twb`> "mke2fs /dev/sdc"
[04:25] <twb`> I mean: I didn't ask it not to
[04:26] <jmarsden> Hmm... OK.  Did you accidentally ask it not to, in   /etc/mke2fs.conf ??
[04:54] <linocisco> hi
[04:54] <twb`> Don't think so
[04:54] <twb`> Hmm 	enable_periodic_fsck = 0
[04:54] <twb`> No idea how that got there tho
[04:55] <linocisco> After one unpropershutdown and reboot, I lost my ssh with putty to my ubuntu 10.04 LTS server. what would be the most likely problem ? If I use on server, it is fine. From terminal with ssh client is not fine
[04:55] <linocisco> ssh localhost is fine
[04:55] <twb`> etckeeper says it changed Sat Jul 23 01:47:19 EST 2011  twb
[04:56] <twb`> During the upgrade -e2fsprogs 1.41.12-4 +e2fsprogs 1.42~WIP-2011-07-02-1
[05:01] <jmarsden> twb`: Aha, I was (sort of) right.  So, a package upgrade changed the mke2fs defaults to doing no checks!  That's... bad.  Although the ~WIP... numbering scheme suggests it may not have been exactly the most refined/tested/stable version of e2fsprogs ever :)
[05:03] <twb`> Maybe I got it out of experimental or something -- this is a debian sid/experimental host btw
[05:04] <twb`> nope, that's sid
[05:04] <jmarsden> Ah, OK, that would make more sense... I was wondering what version of Ubuntu would use that, since oneiric is still only at 1.41.14-1ubuntu3
[05:05] <twb`> Sory about that
[05:07] <jmarsden> linocisco: Is the server at a static IP address?  If it uses DHCP, maybe the reboot gave it a new and different IP address?
[05:07] <linocisco> jmarsden: sure
[05:09] <linocisco> jmarsden: it is with static
[05:09] <linocisco> jmarsden: IP
[05:10] <jmarsden> linocisco: OK.  So, from the server console can you ssh to its network IP (not localhost) OK?
[05:14] <twb`> http://bugs.debian.org/632637
[05:26] <linocisco> jmarsden: yes. it is I can ssh to it's network ip if I ssh on server desk
[05:26] <linocisco> jmarsden: the thing is it can't be ssh from other machine
[05:35] <jmarsden> linocisco: OK, so do you have any firewalling or packet filtering on the server?  ufw, perhaps?
[05:36] <jmarsden> linocisco: You can do     sudo ufw status     to check if it is turned on, and    sudo ufw disable   to turn if off if you need to.
[05:37] <koolhead17> hi all
[05:46] <linocisco> jmarsden: status = inactive
[05:46] <linocisco> jmarsden: what do I do ?
[05:47] <twb`> Oh lame
[05:47] <twb`> debootstrap --limit-priority avoids downloading optional/extra packages, but it still LISTS them in its Packages file
[05:47] <jmarsden> linocisco: That means ufw is not the issue.
[05:48] <jmarsden> linocisco: Basically you need to troubleshoot this as a networking issue.  Can you ping from workstation to server, and from server to workstation?
[05:49] <linocisco> jmarsden: yes. it is fine vice versa
[05:50] <jmarsden> OK.  From the workstation if you    telnet 1.2.3.4 22     (where 1.2.3.4 is the IP of the server), what happens?
[05:53] <linocisco> jmarsden: not ok
[05:53] <jmarsden> linocisco: Be specific!  What error message do you see?
[05:55] <linocisco> jmarsden: I did telnet 1.2.3.4 22 from my windows XP.  no error message found. just cursor blinking and after one more keystroke. back to command prompt. I didn't see any telnet prompt or login prompt
[05:55] <linocisco> jmarsden: i have rebooted server 3 times.
[05:58] <jmarsden> linocisco: I need to go away from the computer soon... Windows telnet should (I think) either show you a connect message and the SSH version number in the banner, or it should (after a minute or so) give yu an error message.  So I am not sure what the issue is.
[05:59] <linocisco> ok
[05:59] <jmarsden> linocisco: wait... you did change the 1.2.3.4 into the real local IP of your server, right?
[05:59] <linocisco> jmarsden: what do you mean ?
[05:59] <jmarsden> Instead of 1.2.3.4 use the actual local IP address of your server.
[06:00] <linocisco> jmarsden:  on server, I tried both ssh localhost and ssh server-ip . not ok.
[06:00] <linocisco> jmarsden: then again from client XP, i tried telnet server IP 22 . not ok
[06:00] <linocisco> jmarsden: i didn't change server 's IP
[06:00] <jmarsden> OK.  I don't know what the issue is.  I need to go; hopefully someone else can help you some more.
[06:04] <linocisco> jmarsden: thanks anway
[06:04] <jmarsden> linocisco: You're welcome.
[06:45] <linocisco> fter one unpropershutdown and reboot, I lost my ssh with putty to my ubuntu 10.04 LTS server. what would be the most likely problem ? If I use on server, it is fine. From terminal with ssh client is not fine [11:25] <linocisco> ssh localhost is fine
[07:50] <dravekx> i need some help. I enabled the admin group for sudoers, and now I cant use my administrator user to sudo anything. If i try to sign in as root, it tells me bad password. anyone have a clue how to fix this?
[07:50] <sms_> I can't ftp into my server while logged out, (Response:	331 Please specify the password. Command:	PASS ******** Error:	Connection timed out, is as far as it goes) And rtorrent keeps seeding, but halts downloading in a detached screen session if I log out. What could this be?
[07:52] <SpamapS> sms_: possibly some kind of restriction in your pam rules.
[07:53] <sms_> What are pam rules?
[07:54] <SpamapS> heh.. not something you probably tweaked on your own
[07:55] <sms_> I installed rtorrent, then noticed it won't keep downloading, and then I couldn't ftp into it either. If I'm logged in locally/via ssh everything works
[07:57] <sms_> "File chunk write error no such file or directory" was on one of the torrents in rtorrent
[07:57] <sms_> but not always
[07:58] <dravekx> ugh. I think I'm screwed
[08:00] <sms_> The harddrive is connected via a sata pci card, if that could mean something
[08:15] <sms_> My home directory is encrypted too, and rtorrent was downloading into my home directory
[08:17] <ersi> You probably have some magic which locks your home directory when you log out then.
[08:18] <sms_> I hate magic :-/
[08:19] <ersi> It's not magic
[08:23] <sms_> Maybe I'll reinstall without encryption, I haven't gotten much configured yet
[08:24] <ersi> Or you could spend some time to unroll this and learn a little in the progress, if you got time (Meant in the most sincere way, realized it might sound snotty)
[08:25] <sms_> No I have time :), just that I'm completely lost
[08:33] <_GoRDoN_> sms_: Are you using full disc encryption or Ubuntus home-dir encryption? I also had some problems with home-dir encryption so I switched to full disc encryption.
[08:34] <sms_> home encryption
[08:37] <_GoRDoN_> sms_: Check your processor usage when you add a new torrent. I think the problem is that rtorrent tries to allocate new space for torrents and that takes too long so it fails.
[08:38] <_GoRDoN_> sms_: But there is no problems with full disc encryption
[08:39] <sms_> hmm ok
[08:39] <_GoRDoN_> sms_: And you can test if the encryption is problem by running rtorrent from /tmp, that should work without problems
[08:40] <Daviey> Good Day People!
[08:40] <Daviey> Everything looking dandy on the ship?
[08:43] <twb`> It's not that kind of ship, mate
[08:43] <ersi> _GoRDoN_: He did mention logging out.
[08:44]  * ersi sticks https://help.ubuntu.com/community/EncryptedHome#Caveats into the channel
[08:46] <sms_> hehe I love how the installer mentions "seamlessly"
[08:48] <ersi> It's seemless if you do not wander into the Caveats
[08:48] <sms_> Indeed
[08:50] <lynxman> Daviey: g'day o/
[08:54] <koolhead11> hey all
[08:55] <koolhead11> why the alpha release gives no information about the same in lsb-release file :(
[08:55] <koolhead11> am inserting it on my pressed to do it for me now
[08:56] <koolhead11> i want the reason for the same.
[08:57] <sms_> Yeah, _GoRDoN_, ersi, rtorrent kept going in /tmp. I don't exactly need encryption, can I remove it?
[08:57]  * koolhead11 looks at Daviey 
[08:59] <Daviey> hey lynxman
[08:59] <Daviey> and koolhead11
[09:01] <koolhead11> Daviey: what is a reason for not mentioning alpha/beta with release dates :(
[09:03] <koolhead11> it becomes troubling when we are testing things on daily build. :(
[09:04]  * koolhead11 wonders whom he should run after for the answer :P
[09:04] <Daviey> koolhead11: uh?  I don't follow
[09:05] <koolhead11> Daviey: i wanted to know why lsb-release file doesn`t make/give exact information about the distro, like alpha/beta with release date
[09:05] <koolhead11> :D
[09:05] <Daviey> koolhead11: Maybe you are looking for. https://wiki.ubuntu.com/OneiricReleaseSchedule :)
[09:05] <ersi> koolhead11: Uh, there's no field for date in /etc/lsb-release
[09:06] <koolhead11> ersi: indeed
[09:06] <ersi> Or do you mean that alpha/beta should preferebly be in DISTRIB_DESCRIPTION?
[09:06] <Daviey> koolhead11: part of the problem, is that we don't /exactly/ know what version of an archive snapshot will be the alpha / beta or even release
[09:06] <koolhead11> ersi: yeah it should be there with the date of built :P
[09:06] <Daviey> Once the archive has changed from what is on the released ISO, it's no longer THE alpha or beta.
[09:06] <ersi> koolhead11: uname -a :P
[09:07] <ersi> or uname -v
[09:08] <koolhead11> ersi: that does not solve my issue :D actully
[09:08] <sbeattie> koolhead11: /var/log/installer/media-info should have the specific information about what build it was installed from.
[09:20] <koolhead11> sbeattie: yeah, i found the date. :) but more detailed description needed
[09:20] <koolhead11> :D
[09:21] <linocisco> hi . i have WDmybookworld  NAS. I want to console log in to this. I dont want to reset. Do I know how I can know it's IP without knowing its IP range ?
[09:26] <ersi> lenios_: You could check the manual. Or do a broadcast ping. There's several methods. If you can log onto your router/switch, you could check the ARP tables/ports
[09:28] <sbeattie> koolhead11: the date + nearby milestone is all that there is for daily snapshots (it will include a respin revision if such things happen); what are you looking for or are trying to solve?
[09:28] <koolhead11> sbeattie: testing some apps on the daily built of oneric :D
[09:28] <rbasak> Does apt-get --download-only require a -y? I can't test right now as everything is up to date :-/
[09:29] <koolhead11> sbeattie: yeah, i think lsb-release should have more detailed description
[09:29] <sbeattie> alas, it would require updating the base-files package for each respin.
[09:33] <koolhead11> sbeattie: redhat on the other hand gives much detailed description on the same  i think
[09:36] <rbasak> koolhead11: what it sounds like you need is a list of all installed packages and their versions. Would that help? dpkg-query -W will give you that
[09:37] <koolhead11> rbasak: thanks. I got what i wanted. :D
[10:50] <RoyK> SpamapS: there's a -server and a -generic kernel - IIRC the -server kernel uses 100Hz and not preempt, both better for server stuff, while the -generic uses CONFIG_PREEMPT_VOLUNTARY
[10:50] <RoyK> that is, seems both are using 100Hz
[10:52] <euphor][a> hi guys, I'm having issues with install 10.04 on a Dell PowerEdge 2600 server. It seems the kernel version causing panics with the raid controller. Anyone come across this and/or got any suggestions please?
[10:53] <RoyK> euphor][a: never seen that - make sure you're using the latest release - 10.04.3
[10:58] <garymc> How do I make a launcher  load a page in firefox or google chrome. some sites i use I want to launch in chrome as they display better in my FAT clients
[10:58] <garymc> or should I ask elsewhere?
[11:00] <RoyK> garymc: I somewhat doubt that's server related...
[11:00] <garymc> yeah ok
[11:00] <garymc> im LTSP
[11:00] <euphor][a> RoyK: thanks, I have the latest release -- ironically an older release works as it has an older kernel
[11:01] <RoyK> euphor][a: then you've hit a bug
[11:01] <RoyK> euphor][a: what do you mean older? older 10.04 or older distro?
[11:01] <euphor][a> older distro, 8.04, having an older kernel
[11:02] <euphor][a> RoyK: yes, it does seem to be a bug
[11:04] <RoyK> euphor][a: there's a backported 2.6.35 available for Lucid as well, but I doubt there's an install CD with it
[11:09] <RoyK> euphor][a: you may try to boot without acpi or apic (press f6 in the initial grub menu)
[11:11] <euphor][a> RoyK: thanks, I'll give it a shot
[11:11] <euphor][a> RoyK: the bug persists up until 2.6.35 afaiaa, so it's 2.6.29 or older only I believe
[11:12] <euphor][a> (that I have to use(
[11:13] <RoyK> have you tried 10.10 or 11.04?
[11:13] <RoyK> not optimal if you want to run LTS, but then, 12.04 isn't that far off
[11:14] <euphor][a> looking at LTS only unfortunately
[11:15] <TheEvilPhoenix> well 10.04 is an LTS build.  You could theoretically try that
[11:15] <euphor][a> As opposed to 10.04.3?
[11:15] <TheEvilPhoenix> sorry, i interchange them.  10.04.3 is the LTS build
[11:16] <TheEvilPhoenix> 10.04 - 10.04.3 are all LTS
[11:16] <TheEvilPhoenix> but its in the 10.04 family :P
[11:16] <TheEvilPhoenix> regardless, its an LTS.
[11:17] <euphor][a> I have tried 10.04.3, it kernel panics randomly
[11:17] <euphor][a> I'll try 11.04 now, just to see
[11:17] <TheEvilPhoenix> well if 10.04.3 kernel panics randomly
[11:17] <TheEvilPhoenix> then its likely that 11.04 will as well
[11:17] <TheEvilPhoenix> is this, say, ancient system hardware you're on?
[11:18] <TheEvilPhoenix> >= 4 years of age?
[11:18] <euphor][a> yes
[11:18] <euphor][a> Dell PowerEdge 2600, EOL
[11:19] <TheEvilPhoenix> heh
[11:19] <euphor][a> I could run 8.04 LTS, but had trouble accessing EFI partitions
[11:23] <TheEvilPhoenix> EFI?  *loads google*
[11:23] <euphor][a> EFI GUID Partition support (NEW)
[11:23] <TheEvilPhoenix> EFI = ....?
[11:24] <euphor][a> to access 4Tb partitions
[11:25] <TheEvilPhoenix> you still havent explained what EFI is :P  *resorts to google*
[11:25] <TheEvilPhoenix> ahh
[11:32] <euphor][a> it's a partitions schema
[11:45] <CluelessPerson> Hello
[11:45] <CluelessPerson> How do you completely reinstall something when you've F**** up all the config files and etc?
[11:46] <CluelessPerson> and need it ALL completely reset?
[11:48] <CluelessPerson> ?
[12:12] <euphor][a> hi guys, I have a 4Tb external firewire drive that, when connected to 8.04, says 'sdb: unknown partition table'
[12:13] <euphor][a> It was created and worked fine on Gentoo, but doesn't mount on Ubuntu?
[12:33] <hallyn> crass: I don't recall what happened with that.
[12:33] <hallyn> somehow I think I thought I was waiting for an update from hggdh
[12:33] <hallyn> crass:  take it you see libvirt still doing that?
[12:37] <hallyn> hggdh: bug 691590, do you still see that behavior?
[12:39] <hggdh> hallyn: I *think* so -- but there is a setting on qemu.conf that disables it, as long as you are running libvirt as root
[12:39] <hggdh> which I have been doing since
[12:41] <hggdh> I can test it easily, though -- right now I am jumping off the fire to land in the frying pan (ISOs changing ownership and blowing out ISO update, or libvirt-as-root and blowing up security)
[12:42] <hallyn> hggdh: does that mean the bug is still worth fixing for your end?
[12:44] <hggdh> hallyn: ISOs should not change ownership, mostly should not have ownership set to a system account. Yes, if it still happens, certainly is worth fixing
[12:47] <zul> good morning
[12:47] <hggdh> morning zul
[12:48] <zul> hey hggdh
[12:48] <hallyn> hggdh: i'll follow up, thanks
[12:49] <hallyn> hggdh: well, to be clear - you're saying that running as root is giving you its own set of problems, right?
[12:49] <hallyn> (so i can prioritize)
[12:50] <hallyn> i'm forgetting what the core of the problem was.  a mirror updating the cdrom backing file seems problematic regardless of uid changes
[12:53] <hallyn> oh well.  patch -> list
[12:54] <hggdh> hallyn: sorry, wife called
[12:55] <hggdh> hallyn: we should not need to run libvirt as root. I am rerunning my scenario to see if it still is a problem
[13:01] <Daviey> hohoho
[13:02] <hallyn> hggdh: 'should not need to' where need is defined by special cases doesn't really help to prioritize :)
[13:03] <hggdh> hallyn: out default install does not run libvirt as root. Having to do so is a deviation
[13:03] <hallyn> IOW, the behavior you want can be argued either way.  So my question was whether the workaround does in fact cause you different problems
[13:04] <hggdh> hallyn: no problems, except running libvirt as root. I am personally confortable there, this is *my* machine
[13:04] <hallyn> hggdh: I honestly believe that having the user provide an iso which is specific for the VM makes more sense, than designing libvirt to handle isos being overwritten by a mirror
[13:05] <hallyn> but if the workaround caused problems, then it would be worth it
[13:05] <hallyn> However, it appears I really dropped the ball by neither marking the bug wontfix, nor forwarding the bug to upstream
[13:05] <hggdh> hallyn: OK. It is a question of usage and of interpretation. I accept it, and I have a workaround. I am cool, no matter what you decide
[13:07] <hallyn> hggdh: cool, thanks - so i'll fwd it upstream, but if they don't take it, then i'll mark it wontfix rather than patch in ubuntu.
[13:07] <hggdh> hallyn: certainly
[13:32] <kpettit> Good morning.
[13:33] <kpettit> What's considered the best VNC version latetly?   Wanting to put something on all the PC's here so I can remotly monitor/support and such.
[13:39] <cr3> what would be the recommended way to use cloud-init to deploy an application in ec2 using one shared configuration, revision controlled, and one private configuration, containing my access token and secret for launchpad?
[13:51] <zul> Daviey: where do you see this block-migration stuff in Openstack?
[14:09]  * pmatulis wonders why samba's testparm did not pick up 2 contradicting 'passdb' entries...
[14:22] <rbasak> So I'm looking at cobbler, and the docs are very RHEL-centric. If I just follow the instruction but with an Ubuntu ISO and preseed file with cobbler from oneiric, will it work, or are there other docs I can use?
[14:22] <jamespage> rbasak: i think there are some on help.ubuntu.com
[14:23] <jamespage> https://help.ubuntu.com/community/Cobbler
[14:23] <rbasak> jamespage: aha, thanks! I searched wiki.ubuntu.com; didn't realise that help.ubuntu.com was separate
[14:23] <jamespage> rbasak: np
[14:28] <jamespage> zul: have the fixes errant qemu-nbd processes landed in nova yet? and do you want me to test if so?
[14:28] <zul> jamespage: not yet...it landed in trunk...it should land in ubuntu this friday
[14:29] <jamespage> zul: ack - lemme know when
[14:29] <andreserl> rbasak, if you find something not working from the docs or anything just let me know
[14:29] <zul> jamespage: sure
[14:29] <RoAkSoAx> rbasak: errr if you find something not working from the docs or anything just let me know
[14:29] <RoAkSoAx> \lol
[14:31] <rbasak> ok, thanks :)
[15:04] <lajjr> kim0, are you online ??
[15:04] <kim0> lajjr: yep
[15:04] <lajjr> great...
[15:04] <kim0> what's up
[15:04] <lajjr> I have a ??
[15:04] <kim0> shoot :)
[15:04] <Daviey> Does anyone want to smoke test the openstack dashboard?
[15:05] <kim0> Daviey: do I have to install a full openstack env for that ?
[15:05] <lajjr> can I also take 15 - 16 ?!?
[15:05] <Daviey> zul: How is the migration looking?
[15:05] <zul> getting there
[15:05] <zul> now that i understand it
[15:05] <kim0> lajjr: heh sure! You can take whatever you want as long as you promise to complete it :)
[15:06] <Daviey> kim0: hmm, pretty much - but should be possible on an all in one node.
[15:06] <Daviey> (or VM)
[15:06] <lajjr> chapters that is.. I also am going to kill bugs for the mail part..
[15:06] <lajjr> heh of course..
[15:06] <kim0> lajjr: AWESOME .. please add yourself to http://pad.ubuntu.com/serverguide
[15:06] <al-maisan> hello there, how do I pass the "--master" flag to "/etc/init.d/mongodb start" ?
[15:07] <Daviey> zul: are you taking notes?
[15:07] <lajjr> OK ..also file server is done and didn't look at the bugs my bad... :(..
[15:07] <zul> Daviey: yep
[15:08] <lajjr> kim0, add, you mean to said chapter I mentioned right???
[15:08] <kim0> Daviey: Is there any doc on how to install this all in one openstack to test the dashboard ?
[15:08] <kim0> lajjr: I mean write your name on http://pad.ubuntu.com/serverguide besides those chapters (like the rest of the chapters)
[15:09] <lajjr> yep doing it right just making sure was in the right frame of mind LOL..
[15:09] <kim0> lajjr: hehe cool :)
[15:11] <lajjr> kim0, added...
[15:12] <lajjr> will try to complete ASAP.. bugs too...!.
[15:13] <lajjr> heh..
[15:15] <cr3> smoser: ^^^ what would be the recommended way to use cloud-init to deploy an application in ec2 using one shared configuration, revision controlled, and one private configuration, containing my access token and secret for launchpad?
[15:15] <Daviey> kim0: not off the top of my head, sorry
[15:17] <caribou> lajjr: I just saw that you signed in to update the chat chapters on the Server Guide
[15:17] <caribou> lajjr: I was about to do it myself (just set up the bzr branches). If you want, I can take care of the chat chapter. I'm not too good with version control
[15:20] <jamespage> smoser: any plans to add arch=arm support to cloud-utils
[15:20] <smoser> cr3, "access token" == "ec2 credentials" ?
[15:20] <jamespage> cloud-publish-image is reluctant todo it at the moment - "arch must be i386 or x86_64"
[15:20] <smoser> for ec2 credentials you most certainly should use 'iam' , and create some credentials that can only do what you want.
[15:21] <Daviey> smoser: for ec2 creds, or aws creds? :)
[15:22] <cr3> smoser: not ec2 credentials, just credentials specific to the application being deployed so that it can connect to launchpad without prompting for openid and oauth credentials
[15:22] <RoAkSoAx> is anyone free to try a PXE boot atm?
[15:22] <RoAkSoAx> I'm getting an endless loop
[15:22] <smoser> then, i would probably seed those (and your launchpad credentials) into the instance via '#include-once' that lynxman wrote.
[15:23] <lynxman> smoser: \o/
[15:23] <cr3> smoser: is #include-once supported in lucid running on the instance where I'm deploying?
[15:23] <smoser> jamespage, you want to open a quick bug ? and we can get it fixed.
[15:24] <smoser> include-once is not supported by lucid. :-(
[15:24] <jamespage> smoser: sure
[15:24] <smoser> cr3, so i would recommend doing some one-time url (or expiring) and maybe that would work with lucid...
[15:24] <smoser> i'm wondering what would happen on subsequent boots.
[15:24] <Daviey> jamespage: Yeah, i edited mine in-place
[15:25] <smoser> if that url is still there to be #included, but just with other data (not the secret from the first time) it would probably be OK.
[15:27] <smoser> jamespage, so euca-bundle-image wont complain about arch as arm ?
[15:27] <cr3> smoser: not sure how that would work: I could probably commit the url pointing to my credentials but couldn't people then access that url?
[15:27] <smoser> cr3, yes. so you need a one time use or expiring url
[15:27] <smoser> and you pass it a url like myhost.com/<some long secret here>
[15:28] <smoser> you can use s3 expiring urls.
[15:28] <jamespage> smoser: might do - Daviey might know the answer (I have not "fixed" mine locally"
[15:28] <smoser> let it live for 10 minutes
[15:29] <Daviey> i'm pretty sure i was able to bundle whatever i wanted.
[15:29] <Daviey> I might have 'fixed' it locally :)
[15:29] <smoser> so i guess i'll just ditch the 'arch' sanity check.
[15:30] <jamespage> smoser, Daviey: bug 849093
[15:30] <smoser> and, let it fail later if it does
[15:30] <Daviey> smoser: Hmm, for aws your logic still makes sense
[15:30] <smoser> yeah, but no way to really know.
[15:30] <smoser> that sucks
[15:30] <smoser> but what can you do
[15:31] <Daviey> smoser: for this release --force-arch=arm ?
[15:31] <smoser> nah.
[15:31] <cr3> smoser: the content of the url would be a cloud.init script, right? ideally, it would preseed the credentials I need, right?
[15:31] <Daviey> hmm
[15:31] <Daviey> that is crappy
[15:31] <smoser> its not like you're going to do that more than once.
[15:31] <smoser> cr3, right. it can be any cloud-init supported content.
[15:32] <jamespage> is it possible to have an openstack cloud running both arm and x86 hardware i.e. can it tell where to locate instances based on architecture?
[15:32] <Daviey> smoser: as long as you support commodore64, i am happy
[15:32] <jamespage> or is it one or the other
[15:32] <Daviey> jamespage: no
[15:33] <Daviey> jamespage: currently nova assumes everything can run on what it is thrown at.
[15:33] <cr3> smoser: I'm not seeing anything about preseeding on the wiki page, do I just make it a script and preseed as if I were on the command line? https://help.ubuntu.com/community/CloudInit
[15:33] <jamespage> hmm - that makes me a little sad
[15:33] <smoser> i just let it fail for the user.
[15:33] <Daviey> So, throwing an amd64 image at an i386 compute node will just fail.
[15:33] <smoser> it sucks, that the failure actually happens on 'register'
[15:33] <smoser> after you've potentially uploaded hundreds of megabytes of useless garbage
[15:33] <smoser> :)
[15:33] <smoser> i just tested ec2-bundle doesn't complain, but ec2-register does.
[15:33] <Daviey> jamespage / smoser: Did you see the arch type on this bug i raised? 831704
[15:34] <Daviey> bug 831704
[15:34] <jamespage> that makes me a little more sad
[15:34] <jamespage> nope
[15:34] <smoser> cr3, what do you mean ? preseed ?
[15:34] <smoser> oh..
[15:34] <smoser> i said 'seed'.
[15:34] <smoser> i just mean however you would insert data. there is no specific support in cloud-init for ec2 credentials
[15:35] <Daviey> jamespage: Yeah, i uploaded a bundle decalred as type commodore64 to our openstack deployment :)
[15:35] <jamespage> As there is no real ARM server hardware this is all a bit hypothetical anyway
[15:35] <jamespage> Daviey: nice...
[15:36] <Daviey> jamespage: Well i hoped to land the branch i had that would support arm via qemu arm emulater.
[15:36] <Daviey> But we've not had great success in starting arm images outside of openstack.. right utlemming / rbasak ?
[15:36] <cr3> smoser: preseed as in debian shove my own values into the configuration of a package at install time :)
[15:36] <jamespage> that might actually be OK - so long as it won't try to run it on a nova-compute node unless its the right arch.
[15:37] <rbasak> Daviey: I still need to sync with utlemming on that
[15:37] <rbasak> Not sure what the current status is
[15:37] <smoser> preseed as in debian, cloud-config has support for doing that.
[15:37] <Daviey> jamespage: No, it should matter of the arch type.  libvirt handles starting of kvm or qemu emulator.
[15:37] <cr3> smoser: a quick search shows that it's been postponed for natty though, probably got introduced in oneiric: https://launchpad.net/ubuntu/+spec/cloud-server-n-cloud-init
[15:38] <utlemming> Daviey: correct, everything that I am seeing is that qemu is broken for OMAP
[15:38] <Daviey> utlemming: rocking!
[15:38] <jamespage> Daviey: think libvirt is currently borked on ARM unless zul has fixed it - have not tried to test it yet
[15:38] <Daviey> utlemming / rbasak: you two probably need to grab hold of the arm team to try and find out what to do :/
[15:38] <smoser> cr3, http://bazaar.launchpad.net/~cloud-init-dev/cloud-init/trunk/view/head:/doc/examples/cloud-config.txt
[15:38] <smoser> see 'debconf_selections'
[15:38] <cr3> smoser: cheers!
[15:38] <jamespage> have been trying to reproduce the java task hang I've seen a few times
[15:39] <smoser> and that was backported to lucid
[15:39] <Daviey> jamespage: Hmm.. No, we are at cross-wires.
[15:39] <Daviey> jamespage: libivrt on amd64/i386, starting qemu to virtualise a fake arm cpu.
[15:39] <jamespage> Daviey: yeah - I got that
[15:40] <jamespage> I was really talking about lxc containers under openstack on ARM hardware
[15:40] <cr3> smoser: I was grep'ing around for preseed but debconf returns exactly what I was looking for, thanks again
[15:40] <smoser> yeah.
[15:40] <smoser> that sucks. i did the same.
[15:40] <smoser> i should add those strings to that doc.
[15:40] <smoser> debconf does hit, though, cr3
[16:34] <utlemming> rbasak: I think I may have a handle on booting qemu arm images
[16:34] <rbasak> cool
[16:34] <utlemming> rbasak: hopefully, I'll have something shortly
[16:34] <rbasak> OK I'll leave you to it then!
[17:36] <tvaldes> I need help with fstab (I believe). There are 3 drives that I swap in and out of a swap drive bay. I would set it up to auto mount sda1 or sdb1... but it's not always the same for some reason (I believe due to a new raid controller). So I have 3 seperate lines mounting the UUID's to the same point. The problem is if the first one isn't present, it gives an error on boot and never mounts. If I sudo mount -a, it works but shows 
[17:40] <koolhead17> hi all
[17:44] <defghanistan> Good morning, I am trying to mount an NFS share onto an ubuntu server box using /etc/fstab, via ip address (not dns). When I run mount -a it says that DNS resolution failed for //(ip address): Name or service not known. I have googled this to death, any suggestions?
[17:53] <just-a-visitor> defghanistan: Try to include the Ubuntu box's IP address into the NFS server's (providing the share)  /etc/hosts file.
[18:26] <its> hello
[18:26] <its> can someone help me screen share my ubuntu server from a mac laptop?
[18:27] <its> I have x11vnc installed and all but, I'm not sure how to get the screen sharing set up on the mac
[18:27] <its> if anyone can help me I'll give them a cookie
[18:37] <pmatulis> its: what kind?
[18:37] <its> what kind of what?
[18:38] <pmatulis> its: of cookie
[18:38] <its> chocolate chip
[18:43] <pmatulis> its: i'm sorry i can't help you
[18:44] <its> oatmeal raisin?
[18:44] <its> :/
[18:45] <Pici> its: #ubuntu would probably be a better place to ask, as the Server install doesn't even have a gui installed normally.
[18:45] <its> well
[18:45] <its> i installed the gui
[18:45] <its> ubuntu-desktop
[18:49] <dravekx> anyone running an IRC BOT on ubuntu server? I kinda want to, but not sure where to start.
[18:50] <genii-around> Maybe start by experimenting with the eggdrop bot
[18:51] <dravekx> Im reading about eggdrop bot, still a little confused. :S
[18:51] <Pici> I found supybot a lot simpler myself.
[18:55] <pmatulis> yes, i run supybot too
[18:57] <dravekx> I guess I can setup a user account on ubuntu for the bot and run it from the home directory?
[18:58] <dravekx> ahh found a tut
[19:00] <dravekx> i wonder if i should make a user account just for the bot.
[19:00] <dravekx> :S
[19:08] <dravekx> how do you clone download something?
[19:08] <Pici> dravekx: What do you mean?
[19:09] <dravekx> Pici, supybot instructions says "Now clone download Supybot from git"
[19:09] <dravekx> *after opening a terminal.
[19:10] <dravekx> ah I got it.
[19:12] <pmatulis> i usually just install the ubuntu package...
[19:21] <kirkland> soren: ping
[19:30] <jfb_h20> anyone have a pointer on setting up postfix to work as the mail server for mediawiki? It seems an 'apt-get install sendmail' worked, but postfix doesn't.
[19:38] <soren> kirkland: wazzup?
[19:38] <kirkland> soren: -> pm
[20:01] <j0nr> hey folks...I am having problems installing video driver so that I can just run xbmc in an X window session (I do not want to install a desktop environment on my server)
[20:02] <j0nr> but I do want to be able to run xbmc on my tv which is plugged into my server
[20:04] <hallyn> zul: is 811951 aimed at you?
[20:05] <zul> bug #811951
[20:05] <hallyn> zul: bc it doesn't make any sense to me
[20:05] <hallyn> (i.e the version # he's talking about)
[20:06] <zul> hallyn: kind of...i know what he is talking about though
[20:14] <hallyn> zul: ok, cool
[20:22] <j0nr> can anyone help me get video driver installed on my server so I can run graphical programs remotely
[20:24] <genii-around> !info xvfb
[20:38] <j0nr> genii-around: would that actually output via my vga port though? to a monitor?
[20:42] <warzauwynn> j0nr: no, it's a virtual frame buffer.  nothing is displayed on the monitor.  it's useful for remote graphics.
[20:42] <warzauwynn> j0nr: http://en.wikipedia.org/wiki/Xvfb
[20:42] <j0nr> ok when I try and launch xinit, I get a few errors, mainly saying no supported AMD display adaptors were found.
[20:42] <j0nr> I just have fglrx installed
[20:43] <j0nr> if I uninstall that and install xserver-xorg-video-radeon then I get errors when runngin xinit that it cant find fglrx components...
[20:43] <warzauwynn> j0nr: you said remote… do you need local too?
[20:43] <j0nr> I'm lost
[20:43] <j0nr> remote as in I will control it remotely.
[20:43] <warzauwynn> j0nr: so you don't need local?
[20:43] <j0nr> Very basically, it is a laptop with ubuntu server on it, sitting under my telly. I want to plug my telly into it and run xbmc
[20:44] <j0nr> I don't want a full on desktop env because it will slow the machine down
[20:44] <warzauwynn> j0nr: that's not remote, that's considered local because no network is involved.
[20:44] <j0nr> warzauwynn: ok, sorry for that
[20:44] <warzauwynn> no worries.  uninstall xvfb and just install the normal xorg stuff
[20:45] <j0nr> warzauwynn: i never installed xvfb.... would would the 'normal' xorg stuff be?
[20:45] <j0nr> I think all I am missing is the correct video driver
[20:45] <warzauwynn> you already have a graphical environment running?
[20:46] <j0nr> no, just ubuntu-server
[20:46] <warzauwynn> install xserver-xorg
[20:46] <j0nr> think I have done that
[20:46] <warzauwynn> well… actually there might be an approved way to do this for xbmc.
[20:46] <j0nr> yes I have
[20:46] <warzauwynn> xbmc might not need everything like the wm and stuff.
[20:46] <warzauwynn> did you install gnome / kde / xfce / or anything else?
[20:46] <j0nr> well it will need a working display adaptor at least
[20:47] <j0nr> warzauwynn: nope, nothing like that installed
[20:47] <j0nr> i've run graphical programs from a server before via ssh and tunneling... I see it as similar to that
[20:47] <warzauwynn> what verison of ubuntu?
[20:47] <j0nr> although it is all local, but it needs a working display adaptor right?
[20:48] <j0nr> warzauwynn: 11.04
[20:48] <warzauwynn> j0nr: nah, it's not like that.  those apps send the GUI information over the net, then your client renders them.  here we need the pc itself to render them.
[20:48] <j0nr> warzauwynn: ah ok
[20:49] <warzauwynn> j0nr: did you do this? http://forum.xbmc.org/showpost.php?p=786004&postcount=8
[20:49] <j0nr> warzauwynn: http://pastebin.com/hbcGAeej  current output when I run xinit
[20:50] <warzauwynn> reason i ask is that installing xbmc might auto-resolve all the dependencies to make it work.
[20:50] <warzauwynn> perhaps… not certain.
[20:50] <warzauwynn> really though, this is less of a server question and more of a desktop question.  you might ask in #ubuntu
[20:51] <j0nr> I used this: sudo add-apt-repository ppa:team-xbmc/unstable    which is meant to be what you should use for natty
[20:51] <warzauwynn> in fact, you'd probably have much better luck in there.  i can't remember the last time i installed xorg and graphics stuff.
[20:51] <j0nr> warzauwynn: oky dokes... thanks anyway!
[20:55] <aarcane> how can I force dpkg/apt-get/aptitude to accept that a package is properly configured when the configure script fails regularly.  it's failing to start a service, which I've configured not to start (and indeed successful starting would be disastrous) and everything else is properly configured.
[20:56] <drbobb> hello
[20:57] <drbobb> I was installing 11.04 on a 2TB pristine clean hard drive
[20:57] <drbobb> I created a LVM PV, and 2 LV's, for root fs and swap
[20:58] <drbobb> at the end of installation, boom: error installing grub
[20:58] <drbobb> unbootable system
[20:59] <drbobb> nothing works
[20:59] <aarcane> drbobb, you should probably try with a GPT partition table and make sure there's a bios-grub partition.  it sounds like grub probably couldn't install due to a bad partitioning scheme.  and yes, you can use LVM atop gpt seamlessly.
[20:59] <drbobb> yes after poking around a little I discovered the installer had used GPT on the drive
[21:00] <drbobb> and this definitely did not work, seamlessly or at all for that matter
[21:01] <drbobb> how do I make a bios-grub partition? the installer offered no such option
[21:02] <drbobb> and I definitely suspect that if I want to dual-boot windows, I should stay away from GPT
[21:03] <aarcane> dual boot windows works from gpt on 64 bit for both vista and 7.
[21:03] <drbobb> well for the moment booting ubuntu doesnt work for me, how do I fix that?
[21:04] <aarcane> also: you use gparted or parted from the installer.  the server installer excludes parted for some reason, as well as apt-get, so you need to use the desktop installer to partition your disks prior to installing ubuntu server.
[21:04] <drbobb> actually I worked around it by using grub-install --force, but I got some dire warnings
[21:04] <aarcane> can you pastebin your dire warnings?
[21:05] <drbobb> aarcane: I don't understand you
[21:05] <drbobb> I was definitely able to partition the drive in the server installer
[21:05] <drbobb> parted sees nothing wrong with those partitions
[21:05] <aarcane> drbobb, parted and gparted are utilities that give you detailed control over partitioning and allow you to manually inspect your partitions and ensure that a bios-grub partition exists.
[21:06] <drbobb> aarcane: I can't pastebin, the said pc is half-installed and in my office, and I'm at home
[21:06] <drbobb> but it said something about blocklists
[21:07] <aarcane> drbobb, these utilities are erroneously absent from the ubuntu server install, so you must use the desktop install to access them if you don't have access to a working environment.
[21:07] <drbobb> aarcane: I booted the server cd in rescue mode, and was able to access parted in that environment
[21:08] <aarcane> drbobb, alright then.  create a small partition at the beginning of the GPT volume and set the bios-grub flag and the boot flag.
[21:08] <drbobb> I mean when chrooted to the server root fs
[21:08] <aarcane> erm..  that might not work then.
[21:08] <drbobb> damn, why isn't that documented anywhere?
[21:09] <drbobb> I lost like 2 hours trying to figure this crap out
[21:09] <aarcane> drbobb, I don't know.  Most people don't use GPT partitions though, so it's uncommon.
[21:09] <drbobb> but I never asked for GPT, the installer applied it silently
[21:10] <drbobb> I guess I need to start over then, a couple of hours more down the drain
[21:10] <aarcane> it's not that bad.  now that you know what you're doing it'll go faster.
[21:11] <drbobb> but how do I ensure that the installer won't use GPT?
[21:11] <aarcane> on a 2TB drive you want (might need, must fact check) GPT partitions.
[21:11] <drbobb> I guess I should start by partitioning from a rescue environment
[21:12] <aarcane> just use the ubuntu desktop live cd.  it's very simple to use and provides gparted which will let you properly configure the bios-grub partition.
[21:13] <aarcane> only create the one partition, let the installer handle the rest of the procedure.  the bios-grub partition should be small (8MB should be enough) with the flags bios-grub and boot enabled near the beginning of the drive.
[21:14] <drbobb> but the desktop installer won't let me configure LVM iirc
[21:21] <drbobb> this doesn't make any sense to me, I go through the installer with no error or warning even, and end up with an unbootable system
[21:22] <warzauwynn> drbobb: what hardware and what software version?
[21:22] <drbobb> 11.04
[21:22] <drbobb> the hw is pretty common
[21:22] <warzauwynn> pretty common as in BIOS?
[21:22] <drbobb> nothing special about it
[21:23] <drbobb> yep
[21:23] <warzauwynn> what motherboard?
[21:23] <drbobb> pretty common as in a PC, not a blade or rack server
[21:23] <warzauwynn> only one disk, 2tb?
[21:23] <drbobb> right
[21:23] <warzauwynn> pc's have EFI.  i just built a mini-ITX with efi.
[21:23] <drbobb> a pc like a pc you get from a pc store
[21:24] <warzauwynn> like an asus motherboard?
[21:24] <drbobb> yeah might be an asus, can't check right now
[21:24] <warzauwynn> asus makes EFI motherboards.
[21:24] <drbobb> there is no EFI on that PC
[21:24] <warzauwynn> ok.
[21:24] <warzauwynn> disks?
[21:25] <drbobb> nothing remotely resembling EFI
[21:25] <warzauwynn> are you sure your install media burned correctly?
[21:25] <drbobb> a 2TB SATA drive, seagate or something
[21:25] <warzauwynn> i've done installs where the install media silently corrupted during burn, and i beat my head for hours over it.
[21:26] <warzauwynn> actually more than once, which is why i try not to use CDs anymore...
[21:26] <drbobb> I seriously doubt that is the issue
[21:26] <warzauwynn> =/
[21:26] <warzauwynn> ok.
[21:27] <warzauwynn> 64 or 32 bit OS?
[21:29] <drbobb> 64bit
[21:32] <adam_g> RoAkSoAx:  http://paste.ubuntu.com/688646/
[21:32] <adam_g> RoAkSoAx: cloud-init wont consume the hostname and domain name from there, will it? and it will consider no hostname from metadata, and update /etc/hosts with localdomain
[21:34] <RoAkSoAx> adam_g: that's the issue I found and we agreed with smoser that cloud-init should probably respect that
[21:34] <RoAkSoAx> adam_g: and tht's why he did the change on cloud-init and that's why hostname is setting correctly
[21:34] <RoAkSoAx> adam_g: because it used to set itself to ubuntuhost
[21:34] <RoAkSoAx> adam_g: so I think he'll probably have to do the same he did for hostname, for domain
[21:37] <adam_g> RoAkSoAx: so cloud-init was modified recently to look for hostname in the kernel paramters?
[21:37] <RoAkSoAx> adam_g: smoser what did you do to make cloud-init not set the hostname to ubuntuhost?
[21:37] <RoAkSoAx> smoser: remember we discussed that few days ago, because it was changing what was provided in the kernel parameters
[21:42] <drbobb> Is the use of LVM what's causing my install to fail? or is it something else?
[21:42] <warzauwynn> drbobb: which partitioning method did you use?
[21:42] <drbobb> manual
[21:43] <RoAkSoAx> adam_g: we just need to make sure that cloud-init is not overwriting that, and that --hostname when adding a system is set
[21:43] <drbobb> I was never aware that GPT would be used on the drive
[21:43] <RoAkSoAx> adam_g: and that should be enough to set the hostname
[21:43] <warzauwynn> drbobb: did you create a reserved bios boot area?
[21:43] <drbobb> and I have no idea of its (in)compatibilities or requirements
[21:44] <warzauwynn> drbobb: i'm playing with it now, it looks like gpt is the default and you need a 1mb reserved bios boot area at the beginning of the disk.
[21:44] <drbobb> there is not a single mention of that in the ubuntu install guide
[21:44] <drbobb> there is no warning in the installer, wither
[21:44] <drbobb> either
[21:45] <drbobb> to me that is seriously broken
[21:47] <warzauwynn> drbobb: incidentally, i can't find any mention of either MBR or GPT.
[21:47] <warzauwynn> however, the automatic installer does create a GPT with the necessary BIOS boot area.
[21:48] <crass> hallyn: thanks for looking at this again. I think my problem is more common than hggdh's, though philosophically we're on the same page.
[21:49] <drbobb> ok, so the solution would be to create this boot area with parted, and re-apply grub install?
[21:49] <drbobb> good, I'll give that a try
[21:49] <crass> hallyn: I just have an iso that will never be updated or sync'd to some mirror, but that was downloaded by the user and so is owned by the user and her default group
[21:51] <crass> hallyn: I don't think that running libvirt as root is a good solution, for security reasons, and there's no reason why libvirt shouldn't be able to read the iso if there's world read perms set on it. I'm not sure why it should ever try to change the ownership for any RO media.
[21:51] <warzauwynn> drbobb: you may want to play with it in a virtual environment where things go quicker and non-destructive.
[21:54] <drbobb> warzauwynn: I don't really need to, the PC is pristine and there is no data to lose
[21:54] <warzauwynn> ok.
[21:55] <drbobb> I just don't have the time to redo the install over and over
[21:55] <warzauwynn> yeah, i know what you mean. =/
[21:56] <drbobb> I must say I'm very disappointed with how the ubuntu installer behaves wrt this, though
[21:59] <hallyn> crass: just updated the bug with a link to my email to upstream.  we'll see waht they say
[22:02] <Xptical> I just installed Cherokee and I'd like some advice on how to manage my directories.  Should I move /var/www to /home/user/www?  Or should I give that user a symlink and permissions to /var/www?
[22:02] <Xptical> So that a user can FTP to his virtual host?