[00:22] <shauno> I read a forum recently, I don't recall where; someone pointed out a google query that'd bring up page after page of mysqladmin pages that were left completely open to the web
[00:22] <shauno> that's when these convenient applications become terrifying
[00:28] <SpamapS> shauno: yeah, backend web apps.. the scourge of the internet. ;)
[00:31] <shauno> http://www.forum.psoft.net/showthread.php?t=25834   that was the example I found.  someone dropped a table off a customer's db, and the autopsy revealed they came in just via google
[02:48] <shauno> not sure I'm understanding ufw correctly.  it seems if I enable it (the only rule I've added so far is allow ssh), I drop off IRC.  The connection appears active (if you can read this, I'm still connected ;), but ctcp ping doesn't seem to work
[02:49] <Italian_Plumber1> say I have two indentical files in different directories.  Can I use md5sum to verify they are identical, and do so in one line without writing the sums to a file?
[02:54] <Takyoji> Anyone know the practicality of applying grsecurity to an Ubuntu 10.04 (64-bit) installation?
[02:58] <erichammond> Italian_Plumber1: Why not use "diff -q" to see if the files are identical?  md5sum must read both files entirely, while diff may have the chance to stop at the point where they differ.
[02:59] <lifeless> erichammond: Italian_Plumber1 'cmp' is your friend.
[02:59] <twb> erichammond: why not use cmp, then?
[03:00]  * lifeless high fives twb
[03:00] <Italian_Plumber1> they're binary files, I only care that they differ
[03:00] <lifeless> yes
[03:00] <lifeless> cmp
[03:01] <twb> $ cmp <(head -1 /dev/urandom) <(head -1 /dev/urandom)
[03:01] <twb>  /dev/fd/63 /dev/fd/62 differ: byte 1, line 1
[03:04] <Italian_Plumber1> :) ok... cmp is cool..  now how do I compare the contents of two directories
[03:04] <erichammond> Italian_Plumber1: diff :)
[03:05] <Italian_Plumber1> hmmm.... all of these commands are producing no output. :)
[03:06] <erichammond> twb, lifeless: In my tests diff -q and cmp -s are performing equally well.
[03:06] <Italian_Plumber1> scratch that I'm comparing the same directory
[03:07] <Italian_Plumber1> ok cool.  thanks guys
[03:07] <Italian_Plumber1> and gals, if applicable.
[03:07] <erichammond> Italian_Plumber1: Send the output to /dev/null and use the exit value of the command to make decisions.  True means the files are the same.
[03:08] <Italian_Plumber1> I had copied a torrent from one place to another and I suspected there were some errors during the transfer.  Looks like there actually weren't any.
[03:08] <Italian_Plumber1> Thanks again...
[03:10] <lifeless> erichammond: diff -q yes, true.
[03:30] <Delemas> I'm trying to repackage my code for ubuntu. In the rpm world %config(noreplace) /etc/someconfigfile is frequently used. Is there an equivalent in Ubuntu?
[03:33] <shauno> ufw appears to be inhibiting outbound ipv6 connects as default?
[03:40] <twb> erichammond: what data are you testing on?
[03:41] <jdstrand> shauno: IPv6 is disabled by default and only packets on lo are allowed
[03:42] <jdstrand> shauno: to enable ipv6, you need to adjust /etc/default/ufw then disable and enable the firewall. see 'man ufw' for details.
[03:43] <twb> Interestingly, comparing a 1GB and a 2GB empty, sparse file is 7s for cmp and 0.005s for diff
[03:43] <twb> Presumably diff first checks the length, whereas cmp supports streams
[03:46] <twb> diff -q is equally fast for two 1GB sparse files, where one contains a byte at the 512M'th place.
[03:46] <twb> Perhaps diff -q looks at the same thing as du --apparent-size when comparing sizes.
[03:47] <shauno> jdstrand: thanks
[03:48] <Jordan_U> twb: I expect that it's just that reading a very sparse sparse file requires almost no IO, which is the main bottleneck.
[03:49] <Jordan_U> twb: Nvm, looking back I see you were talking about relative speed (not just that diff was fast).
[03:50] <twb> Probably someone suggested optimizing cmp in the same way and was told "you can't, it breaks <obscure posix compatibility requirement>"
[04:04] <AlexMax> Hi there, is there any way for me to take a look at the changelog for packages that apt wants to update before I update them?  I tried installing apt-listchanges which is supposed to run when running 'apt' but I never see it appear.
[04:04] <AlexMax> Does it not work with aptitude or something?
[04:12] <twb> !apt-listchanges
[04:12] <twb> Grr.
[04:12] <twb> The ops should just seed ubottu with the dpkg bot's database
[04:12] <twb> Oh, he left anyway
[04:15] <Jordan_U> twb: !info <package> works
[04:15] <twb> The answers he wanted were: use the C key in aptitude (installing libparse-debianchangelog-perl recommended), and apt-listchanges only runs AFTER apt (but before dpkg), and it only lists NEWS by default, not changelog -- most packages don't have NEWS files.
[04:16] <twb> Jordan_U: not quite the same thing though
[04:16] <twb> dpkg> well, apt-listchanges is a package that shows you the NEWS.Debian and/or the Debian changelogs for packages before upgrading them.  It is well worth it for stable users, to see what is happening to their machines as they apply security fixes.  It should be made _mandatory_ for users of testing and unstable.  Also ask me about <d-d-a>, <apt-listbugs>.
[04:16] <twb> versus
[04:16] <twb> dpkg> apt-listchanges: (package change history notification tool), section utils, is standard. Version: 2.85.3 (sid), Packaged size: 82 kB, Installed size: 280 kB
[04:17] <twb> (If you don't already know, you can /msg dpkg if you first /join #debian-bots.)
[06:20] <twb> So I have a prospective customer that needs some handholding for DNS for his cpanel-based VPS
[06:21] <twb> I haven't used cpanel before, but I expect it's basically like webmin, and that I shouldn't go behind its back
[06:35] <fluvvell> Any diff gurus who could tell me how to compare two lists (dpkg -get-selections lists) off two machines to see which packages arent in both lists?
[06:37] <twb> Never mind, he ran away
[06:39] <twb> fluvvell: just diff them
[06:39] <twb> fluvvell: you might need to sort them first
[06:40] <twb> You could also use comm(1) to emit the left-only, right-only or shared lines
[06:42] <SpamapS> twb: cpanel is WAY more controlling than webmin
[06:44] <fluvvell> twd, diff gives me a double list, I want a list of packages I can remove from the machine with more
[06:45] <fluvvell> shhesh, sorry i'm dislexic twb!
[06:45] <SpamapS> fluvvell: sort them, diff is great for sorted lists
[06:46] <twb> SpamapS: "webmin done by professional crack monkeys rather than amateur crack monkeys" :P
[06:46] <fluvvell> SpamapS, is there a setting that can pretty up the output without a whole lot of < > or hex numbers ?
[06:46] <fluvvell> I've sorted them
[06:46] <twb> comm -13 <(ssh foo dpkg --get-selections | sort) <(ssh bar dpkg --get-selections | sort)
[06:47] <twb> fluvvell: you should use diff -u unless you have a really good reason not to
[06:47] <twb> The default diff format is daft
[06:47] <SpamapS> twb: its just that cpanel dictates a lot of stuff, like there's only one mta, and one httpd ..
[06:47] <SpamapS> fluvvell: I like -u
[06:47] <fluvvell> Oh, comm !
[06:47] <fluvvell> so diff -u, are all the - signs the packages that don't exist in both ?
[06:48] <twb> SpamapS: so it's more like the bastard offspring of webmin and, say, zimbra
[06:48] <SpamapS> fluvvell: right, - will be everything that is in box1, but not in box2. + will be things in box2, but not box3
[06:48] <twb> fluvvell: -u lists deletions (-), insertions (+) and, by default, two lines either ide
[06:48] <twb> *side
[06:49] <fluvvell> yeah, comm -3 might do me
[06:49] <SpamapS> rright, comm .. comm is good for this too, I forgot
[06:49] <SpamapS> I always forget comm
[06:49] <fluvvell> cheers guys, thats been mega helpful
[06:50] <twb> fluvvell: #bash is good for generic scripting questions
[06:54] <intelliant> Hi!
[06:56] <intelliant> I have a few ubuntu VMs installed on KVM and they otherwise seem to work okay but randomly give this error - http://picpaste.com/ubuntu_vm-eGalwjCq.png
[06:56] <intelliant> How serious could this be and what could be the cause of the same?
[06:57] <intelliant> At times when I boot these VMs, the filesystem seems to be read-only. This is not necessarily after an improper shutdown.
[06:57] <intelliant> A clean reboot may also result in this.
[06:58] <intelliant> The host OS is openssue 11.3 .
[06:58] <qman__> on real hardware, that could indicate a failing disk or disk controller
[06:58] <qman__> but with a VM, lots of other factors come into play
[06:58] <intelliant> qman__: yes i agree
[06:58] <qman__> could just be a bug, could be failing hardware, could be a misconfiguration
[06:59] <twb> Where's TTL in dig's output?
[06:59] <intelliant> there are other openssue and centos VMs as well but this problem only shows up in ubuntu ones
[07:02] <intelliant> qman__: so i think it could not be a hardware issue
[07:02] <qman__> well, it's clearly failing with a DMA write to the VM's primary disk
[07:03] <qman__> does it happen periodically or only during boot?
[07:03] <intelliant> randomly
[07:04] <qman__> disabling DMA may eliminate the error, but may also impact performance
[07:04] <intelliant> the read-only problem only happens during boot and a reboot fixes it with pressing of F 'fix errors' prior to boot.
[07:04] <qman__> I don't know enough about how DMA performs in VMs
[07:04] <intelliant> the screen dump about DMA is random
[07:04] <intelliant> at times it has gone days without any such error, at times within a few hours
[07:06] <qman__> my best guess would be an issue between your virtual disk availability/performance and ubuntu's disk access performance tweaks
[07:06] <qman__> in that ubuntu is trying to write to the disk without waiting for it to become available
[07:06] <qman__> but that's only a guess
[07:07] <intelliant> qman__: so is there some other channel you suggest I post to ?
[07:07] <qman__> probably to do with upstart, which boots significantly differently from sysv-style rc
[07:07] <qman__> or a kernel setting that ubuntu has and the other VMs don't
[07:08] <qman__> it would take some more in depth investigation
[07:08] <intelliant> okay upstart is a quick-boot thingie in ubuntu, right?
[07:08] <intelliant> I am new to the ubuntu world
[07:08] <qman__> upstart replaces sysv-init
[07:08] <intelliant> right, thanks.
[07:08] <qman__> the way it works is, scripts have certain event triggers to start up
[07:08] <qman__> instead of loading in a pre-set order for given runlevels
[07:08] <intelliant> where do you suggest I continue digging
[07:09] <qman__> well, I'd first try disabling DMA as a kernel parameter, because it's easy to do and test
[07:09] <intelliant> but this problem doesn't seem to happen at boot only
[07:09] <intelliant> how do we do that?
[07:09] <qman__> hold shift while booting to get to grub, edit the boot line, and add the right parameter
[07:10] <qman__> I think it's "nodma" but I don't recall
[07:11] <qman__> looks like ide=nodma
[07:11] <qman__> this is a one-time option, just for testing
[07:11] <intelliant> qman__: http://ubuntuforums.org/showthread.php?t=1014723
[07:12] <qman__> other than that, I'd try to figure out what's writing to the disk and when, when that error comes up
[07:13] <intelliant> libata.dma=0 is the option to use as per this http://ubuntuforums.org/showthread.php?t=1396670
[07:13] <intelliant> that is very tough as it is very very random
[07:14] <intelliant> and the only way to figure that out is by having a constant watch at the virt-viewer
[07:14] <qman__> also, keep an eye on your disk files
[07:14] <qman__> if they're on a network share, see if they're disconnecting, or if they're local, see if something else is trying to lock them
[07:15] <qman__> check your disk performance, too
[07:15] <qman__> the fact that it's specific to ubuntu VMs in an environment where others work fine means that the VM side is more likely
[07:15] <qman__> but it's still possible
[07:15] <intelliant> yes the home is nfs mounted
[07:15] <intelliant> there are 4 test users but none of them have ever complained
[07:16] <qman__> also, check the dmesg on your other machines
[07:16] <qman__> they might be having the same problems but not telling you
[07:16] <intelliant> actually there seems to be no performance loss or disruption
[07:16] <intelliant> it is only the dump
[07:17] <intelliant> no they are well aware about the test env and report any kind of smallest niggles
[07:17] <intelliant> they are trained for that ;)
[07:17] <qman__> I mean the other OSes might not be reporting them to the console
[07:17] <qman__> depends on the defaults
[07:18] <intelliant> that could be
[07:18] <qman__> but the error would always show up in dmesg
[07:18] <qman__> regardless of console reporting
[07:19] <intelliant> but coincidentally i have tried this on 2 different hardwares (Intel and AMD) with the same set of VMs and the results for Ubuntu VMs remains consistent
[07:19] <qman__> if there is no data loss and no performance hit, you may just leave it, but failure to write to disk could be something more serious, it just requires more investigation
[07:19] <intelliant> hence I feel it is something to do with Ubuntu 10.04 installed as VMs only
[07:20] <qman__> well, you need to verify that it is just the ubuntu VMs, that would eliminate the host software
[07:20] <qman__> check the dmesg on the other VMs
[07:20] <intelliant> there is at times loss or corruption of some files such as package manager config or ldap.conf
[07:20] <intelliant> available file,etc
[07:20] <intelliant> but i have seen that is a common issue in ubuntu and deb package mgmt
[07:21] <intelliant> but ldap.conf being lost is something serious
[07:21] <qman__> I wouldn't call it a common issue
[07:21] <qman__> I've only run into it on failing hardware
[07:21] <intelliant> i simply keep a backup and restore it from there and all works again
[07:21] <intelliant> but deb pkg related issues are all over google
[07:21] <qman__> well, if it keeps happening post-boot, it's probably the kernel
[07:22] <intelliant> i have just updated the kernel this morning
[07:22] <qman__> try the no DMA setting, try playing with other performance settings relating to disks
[07:22] <qman__> maybe compare the ubuntu kernel with the kernels from your other OSes
[07:22] <intelliant> 2.6.32-24-generic #42-Ubuntu SMP Fri Aug 20 14:21:58 UTC 2010 x86_64 GNU/Linux
[07:22] <qman__> see what's set differently
[07:23] <intelliant> 2.6.34-12-desktop #1 SMP PREEMPT 2010-06-29 02:39:08 +0200 x86_64 x86_64 x86_64 GNU/Linux
[07:23] <qman__> not the versions, the settings
[07:24] <intelliant> okay
[07:24] <qman__> compile options, kopts
[07:24] <qman__> should be in the documentation or source packages for each distro
[07:24] <ttx> kirkland: I'm now
[07:24] <qman__> focus on disks
[07:24] <qman__> try to find something that sticks out
[07:25] <intelliant> you are right
[07:25] <qman__> if you can eliminate your environment as the problem, file a bug report too
[07:26] <intelliant> other OSes throw this up in dmesg
[07:26] <qman__> in that case, it's likely an issue with the NFS share
[07:27] <intelliant> http://pastebin.com/V2ZdBmxS
[07:27] <qman__> network congestion, hardware failure, host software issues, you name it
[07:27] <qman__> even NFS mount options
[07:27] <intelliant> but this VM has nothing to do with NFS
[07:27] <intelliant> oh it does
[07:27] <qman__> well, it depends on where the VM's disk is stored
[07:27] <intelliant> i will first work with the NFS options
[07:28] <qman__> that's what I'm getting at
[07:28] <qman__> the problem is between the running VM and its disk
[07:28] <intelliant> all are sparse
[07:28] <qman__> could be a number of things, the intermittency points at NFS if the disks are over NFS
[07:28] <intelliant> some are running from the local system and some from NFS shares
[07:28] <qman__> could be host machine getting bogged down, too
[07:29] <intelliant> but the 2 i am analysing now are from the same host system
[07:29] <qman__> too many iops
[07:29] <intelliant> so now it seems i need to look into mount options only
[07:30] <qman__> mount options, and start taking performance logs on the system where the disk images are stored
[07:30] <intelliant> or analyse a VM without any NFS mount
[07:30] <intelliant> performance logs - how?
[07:30] <qman__> if it has a failing disk or is struggling to keep up, that could be it
[07:31] <qman__> you probably want iostat
[07:33] <intelliant> qman__: will do some homework and revert, thanks for your prompt help
[07:34] <qman__> no problem
[07:34] <intelliant> should not be failing hw as it would be too much of a coincidence that both my servers are failing - one 3 yrs old and one a few weeks old
[07:39] <ttx> smoser: in case you're insomniac, ping
[07:56] <intelliant> qman__: exactly same problem - http://ubuntuforums.org/showthread.php?t=1450439
[07:56] <intelliant> similar issues - http://ubuntuforums.org/showthread.php?t=1489063
[07:57] <intelliant> http://tinyurl.com/37p7jj6
[08:01] <huats> morning
[08:45] <intelliant> qman__: looks like the actual demon lies here - https://bugs.launchpad.net/opensuse/+bug/584048
[08:45] <intelliant> this is getting very hot
[10:06] <zenmower> how do i write an iso to an empty partition
[10:13] <daxroc> zenmower: If you just want to output it as is use dd if='my.iso' of='/path/to/partition'  ( with out the quotes )
[10:14] <zenmower> yeah
[10:14] <zenmower> thanks
[10:14] <daxroc> actually not sure if that works
[10:14] <zenmower> well we'll see
[10:20] <daxroc> zenmower: any joy?
[10:22] <zenmower> no i gotta repartition
[10:22] <zenmower> so i havent tried yet
[10:49] <zenmower> ok it's doing somthing
[10:49] <zenmower> we shall see
[12:14] <zenmower> i think this happened to me
[12:14] <zenmower> http://linux.slashdot.org/story/10/08/28/2112208/Some-Windows-Apps-Make-GRUB-2-Unbootable
[12:26] <bt36> hello
[12:29] <bt36> Could somebody point me to a good step-by-step for setting up samba share for windows network on a non-graphical ubuntu server (8.04)
[12:29] <bt36> ??
[12:31] <bt36>  Could somebody point me to a good step-by-step for setting up samba share for windows network on a non-graphical ubuntu server (8.04)??
[12:36] <smoser> ttx, here now.
[12:37] <squidly> bt36: there are a couple of good ways to do that. 1. Read the examples. 2. use swat.
[12:37] <ttx> smoser: Do you think we should mention the results of sanscloud in the technicaloverview ?
[12:37] <squidly> also http://www.samba.org/samba/docs/man/Samba-HOWTO-Collection/install.html#id2551954
[12:37] <smoser> ttx, you have an example of past technical overview ?
[12:38] <zenmower> check ubuntuforums
[12:38] <ttx> smoser: if yes, please edit https://wiki.ubuntu.com/MaverickMeerkat/TechnicalOverview
[12:38] <ttx> smoser: see what's already there for "Ubuntu Server Cloud images"
[12:38] <ttx> smoser: I already mentioned cloud-init and kernel upgrades (for alpha2 and alpha3)
[12:39] <smoser> the kernel is the big thing.
[12:39] <smoser> i wouldn't mind putting the sans sclod info there.
[12:39] <ttx> smoser: please do !
[12:39] <smoser> as it is somethign people have asked about "Can I run this image on my  local system" ?
[12:39] <smoser> ok.
[12:40] <smoser> should I have a link to more info there also?
[12:40] <ttx> sure
[12:44] <Daviey> SpamapS, hey.. can you grab me when you are about please :)
[12:44] <ttx> smoser: also do you think any of the bugs reported against the EC2 images should be targeted to Maverick ?
[12:45] <smoser> ttx, http://paste.ubuntu.com/487194/
[12:45] <ttx> (the "ISO" testing bugs)
[12:45] <smoser> ttx, yes.
[12:45] <ttx> about your blurb: could use an external link for more info
[12:45] <ttx> about the bugs: which ones are "solvable" for release ?
[12:47] <smoser> hit reload on page and read.
[12:47] <smoser> ttx, let me put together a list of what is what.
[12:47] <smoser> a bug list.
[12:48] <ttx> smoser: hit reload; doesn't seem to include a link
[12:48] <smoser> There is more information available on the Ubuntu wiki.
[12:48] <smoser> "wiki" is a link
[12:48] <ttx> ah ,ok
[12:49] <ttx> smoser: works for me, thanks
[12:49] <mealstrom> hi
[12:51] <mealstrom> is that possible to separate ubuntu updates and ubuntu security-updates ?
[12:54] <ttx> mealstrom: yes
[12:54] <ttx> mealstrom: just enable to -security apt repository (and not the -updates one)
[12:56] <mealstrom> ttx: ahhh.. tnx a lot.
[12:57] <Daviey> Although.... I believe -security updates are tested against what is in -updates... Not a big deal, unless there is an ABI bump - which we have seen before.
[13:08] <matti> Hey DavidLevin
[13:08] <matti> Hey Daviey
[13:08] <Daviey> hey matti
[13:14] <DavidLevin> matti, Hi
[13:14] <smoser> ttx, http://paste.ubuntu.com/487207/
[13:14] <smoser> ttx, ping
[13:15] <ttx> smoser: yep
[13:15] <ttx> reading
[13:28] <\sh> maswan, great news...I tested lucid and maverick today on those 465g7 blade servers...and only the installer kernel modules udeb package is broken, means, the hp be2{net,scsi} drivers are not in those udebs...booting into the installed system everything works fine and doesn't give any headaches because of timeouts
[13:28] <\sh> maswan, for reference bug #628776
[13:37] <jdstrand> Daviey: just a very minor clarification-- we build security updates without -updates (so they are guaranteed to be installable for systems without -updates), but we do typically test with -security and -updates installed
[13:37] <jdstrand> Daviey: we also pull from -updates to create a -security update
[13:38] <jdstrand> Daviey: see https://wiki.ubuntu.com/SecurityTeam/UpdatePreparation#Packaging for details
[13:38] <Daviey> jdstrand, Interesting... so your work is based on that of -updates, but your main test base is the main release pocket?
[13:38] <jdstrand> Daviey: that isn't what I meant to say
[13:38] <Daviey> oh
[13:39] <jdstrand> Daviey: we use the latest source package (excepting -proposed) and build with only release and -security
[13:39] <jdstrand> Daviey: this is to support all Ubuntu users-- ones that only have -security enabled and ones who also have -updates enabled
[13:39] <jdstrand> Daviey: we test with -updates enabled
[13:40] <jdstrand> Daviey: since most people run with -updates
[13:40] <Daviey> jdstrand, Oh great.. How will that scale with the perception I have that we are producing more SRU's.. ie the delta between -updates and release is larger?
[13:41] <Daviey> jdstrand, I always knew you *could* run without -updates, but i just assumed it wasn't a supported model... :)
[13:42] <Daviey> jdstrand, So... If -security creates an ABI bump, meaning something statically built against it's rdepends; should the no change rebuild be put in -security or -updates?
[13:43] <jdstrand> Daviey: running with or without updates is supported
[13:44] <jdstrand> Daviey: we pull from -updates to reduce version skew within the archive
[13:44] <jdstrand> Daviey: if something from -updates needs a security update but also a bunch of other 'stuff', we will pull that other 'stuff' into the -security update
[13:45] <Daviey> jdstrand, thanks :)
[13:46] <jdstrand> Daviey: it happens less frequently than one might think, but we have most definitely been known to do 'no change rebuilds for security' for things that a security update pulled from -updates depends on
[13:47] <hallyn> that sounds painful
[13:48] <jdstrand> Daviey: and on a personal note, I always run -updates on desktops, and on servers it is more of a choice-- I've definitely been known to only run with -security on certain systems
[13:49] <Daviey> jdstrand, I remember discovering one.  doing the dance of raising a bug, creating a package, getting it into -proposed... only to see one uploaded into -security a few days later :)
[13:49]  * Daviey cursed :D
[13:49] <jdstrand> hallyn: it can be, but it is less painful than having to create and test up to 2 security updates per release
[13:49] <jdstrand> hallyn: ie, one for -updates and one for -security
[13:50] <jdstrand> Daviey: yes, we try to notice those and will mention in the bug that this will happen/has happened
[13:51] <Daviey> jdstrand, It was a universe one, that one of your chapies noticed later.
[13:51] <jdstrand> Daviey: but that goes both ways-- sometimes something pops up in -updates while we are preparing a -security update and on the day of publication we have to reroll and retest :)
[13:51] <Daviey> lovely :)
[13:52] <jdstrand> again, still less pain overall than having to test twice and potentially having significantly different versions of software to update
[13:52] <jdstrand> (and in the archive)
[13:52] <jdstrand> eg, gnome and kde sometimes get largish changes into -updates
[13:53] <jdstrand> (whole version bumps)
[13:53] <jdstrand> it is admittedly imperfect, but the least evil
[13:55] <jdstrand> we like to turn it around and say it is a feature for people running with only -security-- they get the benefit of getting particularly well tested bug fixes when running with only -security (ie, something goes to -proposed, goes to -updates and sits there until a -security update happens-- very low chance of regression typically)
[13:55] <jdstrand> the spin doctors that we are :)
[14:14] <maswan> \sh: excellent
[14:32] <benedikt> How can I add a route without being root? Since network-manager-openvpn is able to add the VPN routes without NM being root, I figure this is possible?
[14:33] <\sh> benedikt, I think NM is setting the routes via dbus magic which somehow has root privileges afaik
[14:34] <benedikt> \sh: ah. Thanks, now I'll figure out how to make dbus set routes
[14:36] <\sh> benedikt, na..openvpn sets the routes and this openvpn process is being started by nm-openvpn magic via dbus as root I'm not a specialist for NM...ask someone from plumbers frontier
[14:37] <benedikt> \sh: hm.. i was going to do a quick hack, but since dbus has been added to the equation this wouldn't be a quick hack any more.
[14:37] <benedikt> I think the quickest solution would be to allow the user to run sude ip route add without asking for password
[14:40] <ttx> mathiaz: do you plan to exercise part of your magic scripts on the beta ISo testing ? I did not see any results from you yet
[14:40] <mathiaz> ttx: working on it
[14:41] <mathiaz> ttx: any tests are missing for now
[14:41] <\sh> I really would like to read something about udevs magic how to number the different pci devices..is it from highest to lowest IRQ? so SmartArray on IRQ11 will become /dev/cciss/c0d0 and the smartarray on irq10 will become /dev/cciss/c1d0?
[14:41] <ttx> mathiaz: is that a question ?
[14:41] <mathiaz> ttx: I'm integrating my iso testing scripts with hudson
[14:41] <mathiaz> ttx: any tests are missing for now ?
[14:41] <ttx> i386 / Install (default + RAID1)
[14:41] <ttx> the ESX tests but we can't really do those
[14:42] <ttx> + some optional testcases, including the "UEC install (separate networks topology)"
[14:42] <mathiaz> ttx: ESX -> usualy done  by someone from QA
[14:42] <mathiaz> ttx: ok - I'll do the raid1
[14:42] <ttx> we may have a couple of hours left, fyi
[14:43] <ttx> but not much more.
[14:43] <ttx> Daviey, hggdh: did anyone test the separate networks topology (aka Topo3) ?
[14:45] <\sh> benedikt, I wouldn't allow any user to set ip routes at all...they could break things
[14:45] <Daviey> ttx, I have tested a 3 box topology a week ago.. Carlos has been mainly focusing on multiple
[14:46] <ttx> Daviey: would be good to register a result with the beta candidate on Topo3 in the tracker
[14:46] <Daviey> ttx, agreed
[14:46] <ttx> since complex topos tend to work better than simple ones those days
[14:47] <benedikt> \sh: this is just a quick script for me to route certain networks without going through the vpn router
[14:52] <Daviey> ttx, Would a preseeded test from hggdh suffice?
[14:54] <hggdh> ttx: I am going to test it now (topo3)
[14:54] <hggdh> Daviey: yes
[14:54] <Daviey> rockin'
[15:31] <SpamapS> Daviey: here
[15:31] <SpamapS> Daviey: though I'll need to step away briefly when the baby wakes. ;)
[15:32] <nxvl> zul: btw, i uploaded augeas to debian last week
[15:32] <Daviey> SpamapS, heh, ok ;)
[15:32] <zul> nxvl: is it worth asking for a FFE for it
[15:32] <ttx> SpamapS, zul: which one of you two plans to push the Lucid SRU on bug 564920 ?
[15:32] <nxvl> zul: let me check
[15:32] <zul> ttx: i can do it this afternoon
[15:33] <SpamapS> ttx: I submitted it already as a merge proposal, I can't push anymore than that. ;)
[15:33] <ttx> hmmm
[15:33] <nxvl> zul: seems so
[15:34] <zul> nxvl: k...remind me on monday
[15:34] <ttx> SpamapS: and i commented on your proposal
[15:34] <nxvl> zul: a couple of performance stuff, some bug fixing
[15:34] <ttx> SpamapS: I haven't seen those remarks taken into account yet :P
[15:35] <ttx> SpamapS: see my comment on https://code.launchpad.net/~clint-fewbar/ubuntu/lucid/php5/lucid-sru-lp564920/+merge/32803
[15:35] <nxvl> zul: http://paste.ubuntu.com/487274/
[15:36] <zul> nxvl: cool
[15:36] <ttx> SpamapS: maybe you can write the SRU report and zul can sponsor your branch ?
[15:37] <zul> ttx: that would be ideal ;)
[15:37] <SpamapS> ttx: oh I may have missed that.
[15:40] <X-Sleepy-X> ive plugged in a hdd into new hardware, m-board etc, and it has ubuntu server on it, think its 8.04. it doesnt find the network and i wonder if there's a neat little command to solve this problem?
[15:40] <patdk-wk> do-dist-upgrade? :)
[15:41] <X-Sleepy-X> patdk-wk: as i said, i dont have network connection...
[15:41] <patdk-wk> so?
[15:41] <patdk-wk> I can do it without a network connection
[15:42] <patdk-wk> :)
[15:42] <patdk-wk> what network chipset?
[15:45] <X-Sleepy-X> realtek semiconductor rtl-8139/8139c
[15:45] <patdk-wk> that is supported
[15:45] <X-Sleepy-X> and VIA Tecnologies vt8233
[15:45] <patdk-wk> dunno about that one
[15:45] <patdk-wk> are you sure it's just not showing up as eth1? eth2? eth3? ....
[15:46] <X-Sleepy-X> yes, but the installation used another motherboard so its like it hasnt detected the change in hardware
[15:46] <patdk-wk> ubuntu tracks network interfaces
[15:46] <X-Sleepy-X> yeah i see eth1 and 2 on iwconfig
[15:46] <patdk-wk> the id of the old motherboard eth0, will be eth0 forever
[15:46] <patdk-wk> new motherboard would use next available ethx
[15:46] <X-Sleepy-X> ok but it finds them as wireless cards
[15:47] <patdk-wk> unless you wipe out the persistant file that tracks it
[15:47] <X-Sleepy-X> ifconfig finds nothing and iwconfig finds them
[15:49] <ivoks> try not to use ifconfig, rather iproute
[15:49] <ivoks> (ip a would be the command)
[15:49] <X-Sleepy-X> k
[15:49] <ivoks> that's not the solution for the problem
[15:50] <ivoks> it's just general advice
[15:50] <X-Sleepy-X> hmm
[15:50] <ivoks> if you replaced you network cards with new ones
[15:51] <SpamapS> ttx: SRU info added
[15:51] <ivoks> delete /etc/udev/rules.d/70-persistent-net.rules file and reboot
[15:54] <X-Sleepy-X> ivoks: thanks, will try that if not changing the settings in /etc/network/interfaces work on this reboot
[15:54] <ivoks> ubuntu ties interface to hardware
[15:54] <ivoks> if you add/change new hardware, it will get new interface name
[15:55] <ivoks> those links are stored in /etc/udev/rules.d/70-persistent-net.rules
[16:03] <X-Sleepy-X> ivoks: didnt work
[16:03] <X-Sleepy-X> well
[16:03] <X-Sleepy-X> imean
[16:03] <X-Sleepy-X> it now detects the eth0 and 1
[16:03] <X-Sleepy-X> but still no connection to the internet
[16:04] <X-Sleepy-X> this is driving me nuts
[16:04] <X-Sleepy-X> shouldnt be that hard to fix
[16:05] <siretart> show the output of `ip a ls` (in a pastebot)
[16:08] <ttx> mathiaz: on http://iso.qa.ubuntu.com/qatracker/result/4442/287 you didn't encounter the same bugs as kirkland did in http://iso.qa.ubuntu.com/qatracker/result/4441/286 ?
[16:09] <ttx> mathiaz: ignore me, kirkland clarified on PM
[16:13] <zul> ttx: when i was testing samba yesterday i had both smbd and nmbd fyi
[16:16] <kirkland> mathiaz: however, i am curious
[16:17] <kirkland> mathiaz: can you look at the 3 bugs I reported against that test case (RAID) and tell me if any of those look familiar?  or how you worked around them?
[16:22] <Maletor> Why can I not grow my RAID array through disk utility. More information can be found here. http://ubuntuforums.org/showthread.php?p=9797001
[16:27] <mathiaz> kirkland: bug 628374
[16:27] <mathiaz> kirkland: ^^ didn't see it
[16:28] <mathiaz> kirkland: bug 628400
[16:28] <mathiaz> kirkland: didn't see it either
[16:28] <mathiaz> kirkland: I was prompted when the system was booting from a degraded array
[16:29] <kirkland> mathiaz: that's a regression since alpha3
[16:29] <kirkland> mathiaz: then
[16:29] <kirkland> ttx: I believe that one should be a regression and release blocker, and we need foundations to fix it by RC
[16:30] <mathiaz> kirkland: bug 628418
[16:30] <mathiaz> kirkland: ^^ haven't needed that
[16:30] <ttx> kirkland: I'm surprised it doesn't affect Lucid, after all plymouth didn't change that much in Maverick
[16:33] <garymc> Hi Guys, I setup my website off my office server sometime last year.... I have lost the chat files and documents I kept for adding websites to my server. I wonder if anyone can be of guidance?
[16:35] <SpamapS> garymc: thats a very open ended vague question. What exactly do you need help with?
[16:36] <garymc> OK I have a website hosted on my office server. mywebsite.com
[16:37] <garymc> I want to host mywebsite.com with a hosting company to free up my office internet connection
[16:38] <garymc> But I need to keep the Applyform.php on my office server. So I want to point mywebsite.co.uk at the office server instead. So when people are on my hosted site mywebsite.com and they click on Apply.php it links to the page on our office server
[16:38] <garymc> under mywebsite.co.uk
[16:39] <garymc> ok?
[16:39] <garymc> So right now I have pointed the .co.uk to my server ip address
[16:39] <SpamapS> garymc: sounds overly complex. Why can't you put Applyform.php on your hosted server too?
[16:39] <kirkland> mathiaz: ttx: can you boot maverick in a KVM and get to the grub menu?
[16:40] <kirkland> mathiaz: ttx: by pressing <shift>?  I can't at all
[16:40] <cyphermox> mathiaz, I'm testing JeOS on ESX for the beta and there's a number of things that don't match up with the test case... specifically, now uname -r shows "-virtual" for i386 and amd64 (like for KVM), and modules/full install take more space than they should
[16:40] <kirkland> mathiaz: ttx: this really sucks
[16:40] <garymc> Because it is using a mySQL server that is on our office machine. And the Terminals link to that machine. Thus when someone applies..
[16:40] <ttx> kirkland: i admit doing my ISO testing on real HW [tm]
[16:40] <patdk-wk> garymc heh, you are going have to use like a proxy for that file, to do that
[16:40] <mathiaz> cyphermox: uname -r is normal
[16:40] <mathiaz> cyphermox: it's has been corrected in the JeOS on KVM test case
[16:41] <mathiaz> cyphermox: install space is a known bug
[16:41] <cyphermox> mathiaz, cool, that I expected. bug #?
[16:41] <mathiaz> cyphermox: bug 621175
[16:41] <cyphermox> mathiaz, I'll fix the test case for uname -r
[16:41] <mathiaz> cyphermox: cool - thanks
[16:42] <SpamapS> garymc: that all sort of makes sense. So whats the question again?
[16:45] <garymc> brb
[16:47] <jdstrand> kirkland: if using libvirt, try something like this:
[16:47] <jdstrand> virsh start sec-maverick-i386 ; sleep 0.5 ; virt-viewer -c qemu:///system --wait sec-maverick-i386
[16:47] <jdstrand> kirkland: I have good success seeing the grub menu with that
[16:48] <kirkland> jdstrand: hmm
[16:48] <kirkland> jdstrand: okay
[16:48] <jdstrand> in fact, you could probably omit the sleep entirely
[16:53] <TeTeT> kirkland: unfortunately I could not test the virt PPA for bug 590929 today, as the firewall does not give me access to it. I will see if #IS opens it overnight and get it to it first thing next morning, if need be, d/l the files by hand
[16:54] <kirkland> TeTeT: okay, thanks
[16:58] <vmlintu> How does one set the limit for maximum number of open files in 10.04? In 8.04 one would edit /etc/security/limits.conf and add nofiles -setting there, but so far I'm not having any luck. slapd doesn't like the default 1024, so it needs to be made higher..
[17:14] <garymc> SpamapS : Ok was on important call where was i .... ?
[17:14] <garymc> I need Apache to recognize mysite.co.uk
[17:14] <garymc> what files do I need to alter
[17:16] <vmlintu> garymc: virtual host configurations are usually under /etc/apache2/sites-available/
[17:16] <garymc> ok and I just add mysite.co.uk in there like the others?
[17:17] <vmlintu> garymc: do you want to add a new site or new alias for an existing site?
[17:17] <garymc> Can I make the .co.uk use the same folders as the.com site?
[17:17] <garymc> new alias
[17:17] <vmlintu> do you have a file there for your site?
[17:17] <garymc> ?
[17:17] <vmlintu> under /etc/apache2/sites-available
[17:17] <garymc> I have a current file for 2 sites I have on our server
[17:17] <garymc> yes
[17:18] <vmlintu> Do you there ServerName directive?
[17:18] <garymc> ?
[17:19] <vmlintu> ServerName yoursite.com
[17:19] <vmlintu> in the configuration file under /etc/apache2/sites-available
[17:19] <garymc> I have file "hosted-site"
[17:20] <garymc> with details in there
[17:20] <vmlintu> Inside that file, do you have a line that has ServerName in it?
[17:20] <garymc> yes
[17:20] <vmlintu> After that line add a new line:
[17:20] <garymc> ok
[17:20] <vmlintu> ServerAlias yoursite.co.uk
[17:21] <vmlintu> That adds an alias for the existing site
[17:21] <garymc> ok
[17:21] <garymc> do i put the www
[17:22] <garymc> ok do i need to restart apache
[17:22] <vmlintu> apache recognises exactly the name you put in there, so if you want it to be www.yoursite.co.uk, you have to add www
[17:23] <vmlintu> Yes, apache needs to be restarted after the change
[17:23] <garymc> sorry how do i restart apache again. Long time since I last did it
[17:23] <garymc> apached restart?
[17:23] <vmlintu> /etc/init.d/apache2 restart
[17:24] <garymc> ok done
[17:24] <SpamapS> err
[17:24] <garymc> gonna test now
[17:24] <SpamapS> better to suggest people use 'service apache2 restart'
[17:25] <garymc> ok well my .co.uk is still not pointing at my server ?
[17:25] <vmlintu> SpamapS: true.. I'm just too used to using /etc/init.d/*.. old habits die slow..
[17:25] <garymc> or my server isnt allowing it
[17:25] <SpamapS> vmlintu: me too ;)
[17:26] <SpamapS> garymc: DNS will need to be setup first.
[17:26] <vmlintu> garymc: do you dns setup or /etc/hosts with the name?
[17:26] <garymc> I have it is pointed at my office server
[17:26] <garymc> but I only did that about 1 hour ago
[17:27] <garymc> maybe it takes a little time to go through
[17:28] <vmlintu> garymc: you can check with dig if dns gives you the right ip
[17:28] <vmlintu> dig www.yoursite.co.uk
[17:29] <vmlintu> And on the server you can check the logfiles to see if the request arrives there. The logfile is usually defined in the configuration file you just modified.
[17:30] <garymc> ok thanks
[17:31] <garymc> whats dig?
[17:32] <vmlintu> dig is a DNS lookup utility - I think it's in dnsutils package
[17:32] <vmlintu> nslookup works also
[17:32] <SpamapS> nslookup is oooll school
[17:32] <SpamapS> host ftw ;)
[17:33] <vmlintu> I feel really old..
[17:34] <SpamapS> garymc: your domain most likely has a TTL set very high. in the dig response, you should see the "TTL" of the response as the number right after the hostname
[17:34] <SpamapS> ;; ANSWER SECTION:
[17:34] <SpamapS> spamaps.org.		1800	IN	A	65.98.207.160
[17:34] <SpamapS> as an example
[17:34] <SpamapS> garymc: TTL is the number of seconds caches are allowed to hang on to the response.
[17:39] <ttx> yay complete ISO testing coverage for server !
[17:39] <ttx> Kudos everyone
[17:40] <ttx> -smoser who failed to validate some asia-pacific AMI, of course :P
[17:45] <SpamapS> ttx: ugh, sorry I wasn't able to contribute to the iso testing. Been chasing this libdbi stuff. :-P
[17:46] <garymc> ok my .co.uk is now pointing at my server, but if I type into the browser mywebsite.co.uk it displays my error 404 page
[17:46] <garymc> instead of the homepage
[17:47] <gholms> smoser: ping
[17:48] <garymc> what could cause this SpamapS ^^
[17:48] <garymc> ?
[17:48] <smoser> gholms, here.
[17:49] <SpamapS> garymc: enter the hostname, without any path, does it show you the "IT worked!" page?
[17:49] <gholms> smoser: I see version numbers on the cloud-init packages, but no tarballs on the LP site for it.  Is that by mistake, or do I need to be using checkouts to build tarballs all the time?
[17:50] <smoser> :)
[17:50] <smoser> i just haven't really had a reason to build them.
[17:51] <smoser> gholms, i usually do build the tarball and put it at http://smoser.brickies.net/cloud-init-dist/
[17:51] <gholms> Ah, that would work.
[17:51] <gholms> smoser: I'm looking into using most of those scripts for Fedora, so a more-or-less official spot to download release tarballs from would be most appreciated.
[17:52] <smoser> cloud-init is really almost a native-package.
[17:52] <smoser> gholms, i can start posting to launchpad
[17:52] <gholms> That would be really helpful.
[17:52] <smoser> and i would absolutely do that if your willing to help make it more portable :)
[17:52] <gholms> Obviously calls to apt and things wouldn't work, but I would hate to have to fork everything just because it doesn't quite work on RH-type distros.
[17:53] <smoser> yeah.
[17:53] <smoser> i'm definitely open to having distro specific stuff
[17:53] <smoser> and being more generic across the board.
[17:53] <smoser> ie, "apt_update" => "software_update"
[17:53] <smoser> or such
[17:53] <gholms> Sure
[17:54] <smoser> i'll post the most recent tarball on launchpad
[17:54] <smoser> and try to continue doing that.
[17:54] <gholms> That would be awesome.  I'll let you know how it seems to work out from the Fedora side.
[17:55] <garymc> why is my website mywebsite.co.uk working but if I do www.mywebsite.co.uk I get my error404 page?
[17:56] <garymc> fixed it
[18:01] <smoser> gholms, fwiw, you can get to the tarballs from launchpad, via the ubuntu builds.
[18:01] <smoser> ie: https://launchpad.net/ubuntu/maverick/+source/cloud-init
[18:01] <smoser> has cloud-init_0.5.14.orig.tar.gz
[18:01] <gholms> Yeah, that's what I initially grabbed.
[18:01] <smoser> i'm trying to figure out how to put a tarfile up there, though.
[18:02] <smoser> anyone know how i would do that?
[18:02] <gholms> It didn't seem "upstream" enough to satisfy the packaging committee.  :P
[18:03] <patdk-wk> heh, talk to salmon, they go upstream all the time
[18:03] <bjaanes> I
[18:06] <smoser> i'm seriously baffled.
[18:06] <smoser> i have no idea how to put a file up for download on launchpad
[18:06] <smoser> https://launchpad.net/cloud-init
[18:06] <bjaanes> I've just set up my DNS & DHCP server. Everything is pretty basic (from Ubuntu Server book and "official" documentation and such.) but the server keeps getting an DHCP address from itself. I have configured the interfaces file with static IP - but it doesnt seem to bother with it. So now I have to restart the networking service every time want to have the right IP. Anyone got a clue about this?
[18:06] <patdk-wk> heh?
[18:07] <patdk-wk> what is in /etc/network/interfaces?
[18:07] <bjaanes> one moment
[18:08] <bjaanes> auto eth0
[18:08] <bjaanes> iface eth0 inet static
[18:08] <bjaanes>         address 192.168.192.11
[18:08] <bjaanes>         network 192.168.192.0
[18:08] <bjaanes>         netmask 255.255.255.0
[18:08] <bjaanes>         broadcast 192.168.192.255
[18:08] <bjaanes>         gateway 192.168.192.1
[18:08] <bjaanes> and it works when I go "sudo /etc/init.d/networking restart"
[18:09] <bjaanes> but withing a minute or two it gets an IP like 192.168.192.52
[18:10] <bjaanes> Could even my dhcp server override static stuff like that?
[18:10] <patdk-wk> dhcp servers don't override
[18:10] <bjaanes> what i thought
[18:10] <patdk-wk> but if you are running a dhcp client, it would
[18:10] <bjaanes> okey?
[18:11] <bjaanes> Why would a dhcp client be running?
[18:11] <patdk-wk> dunno :)
[18:11] <patdk-wk> ps ax | grep dhc
[18:11] <patdk-wk> might show if one is
[18:11] <bjaanes>   816 ?        Ss     0:00 dhclient3 -e IF_METRIC=100 -pf /var/run/dhclient.eth0.pid -lf /var/lib/dhcp3/dhclient.eth0.leases eth0
[18:11] <bjaanes>  2902 ?        Ss     0:00 /usr/sbin/dhcpd3 -q -pf /var/run/dhcp3-server/dhcpd.pid -cf /etc/dhcp3/dhcpd.conf eth0
[18:11] <bjaanes>  3781 pts/0    S+     0:00 grep --color=auto dhcp
[18:12] <patdk-wk> yep, dhclient is running
[18:12] <bjaanes> Awe. Why does it do that? :( My last server didn
[18:13] <RoAkSoAx> kirkland: when's next ISO testing?
[18:13] <patdk-wk> I dunno :)
[18:13] <patdk-wk> unless it was set to, iface eth0 inet dhcp
[18:13] <patdk-wk> and you just changed it to static
[18:13] <patdk-wk> did you try rebooting?
[18:13] <patdk-wk> and see if dhclient is still running then?
[18:13] <patdk-wk> cause it shouldn't
[18:13] <patdk-wk> and I dunno what you did to make it do it
[18:14] <bjaanes> It helped to reboot. I can
[18:14] <bjaanes> Sorry to have bothered you with something I should have thought of trying =/
[18:14] <bjaanes> But very much thanks :D
[18:16] <smoser> gholms, whoowhoo. https://launchpad.net/cloud-init/trunk now has a download.
[18:16] <smoser> that was a lot of work
[18:16] <RoAkSoAx> smoser: when's the next ISO testing?
[18:16] <gholms> smoser: Awesome
[18:16] <smoser> RoAkSoAx, ?
[18:16] <smoser> iso testing occurs ~ 3 days before release.
[18:17] <smoser> RC release info athttps://wiki.ubuntu.com/MaverickReleaseSchedule
[18:18] <RoAkSoAx> smoser: Ok, I just read in the meeting minutes that soon will be ISO testing. ANyways, how soon do you need the  'testdrive -u' thing for UEC images?
[18:18] <gholms> smoser: The first thing I think would help is factoring out package management code.  I can definitely try to help figure that out.
[18:18] <smoser> RoAkSoAx, what is 'testdrive -u' ? what is working now ?
[18:19] <smoser> at this point its probably very difficult to get any new features into ubuntu.
[18:20] <smoser> gholms, yeah, that'd be great. its probably a good time for it too, as the maverick code is roughly frozen (other than some bug fixes).
[18:20] <RoAkSoAx> smoser: 'testdrive -u' support to test drive local uec images
[18:20] <gholms> smoser: If all else fails, 10.10 can release without the generic code while Fedora picks up the newer stuff in the mean time.
[18:21] <smoser> so right now there is generally functional use of uec images in the archive, RoAkSoAx  ?
[18:21] <smoser> gholms, yeah, 10.10 is releasing with what is there. which is fine.
[18:21] <RoAkSoAx> smoser: currently it only supports testing daily images from uec-images.u.c
[18:21] <smoser> RoAkSoAx, thats fine then.
[18:22] <smoser> so, i dont' see a real reason to push to get '-u' into the archive.
[18:22] <smoser> for maverick.
[18:22] <RoAkSoAx> none pf this is on maverick
[18:22] <smoser> so, i'd say, just leave that on your own schedule, and lets make sure we have it in natty
[18:22] <ssureshot> I cant seem get ubuntu to boot using a software raid on my server... Are there any workarounds for this? If I boot to the cd and choose to boot fronm hard disk it works fine... Drive is set with the boot flag...
[18:22] <smoser> (Narwhals Narwhals swimming in the ocean)
[18:23] <smoser> RoAkSoAx, "none pf" ? i'm confused now.
[18:23] <RoAkSoAx> smoser: i'm gonna backport when the changes are ready, but first I want to add suport for testing local UEC images, given that you were requesting that :)
[18:23] <RoAkSoAx> smoser: "none of"*
[18:23] <smoser> oh. i thought that you had gotten part of it into maverick.
[18:24] <RoAkSoAx> smoser: nope, got busy with coming back to the US so I couldn't push it to maverick,a nd I've been sick the past few days
[18:24] <smoser> oh well.
[18:25] <hggdh> ScottK: do you have a bit of time for a chat?
[18:26] <ScottK> hggdh: My latency might be a bit high, but sure.
[18:26] <RoAkSoAx> smoser: I just wanna know for when would you like the functionality of testdriving local UEC images.
[18:31] <smoser> RoAkSoAx, well, at this point i'm not in a big hurry.
[18:31] <smoser> so a couple weeks is fine. it missed maverick, so i'd like it in natty.
[18:35] <RoAkSoAx> smoser: I thought you just wanted the functionality, and not specifically in maverick :). That's why I actually only added functionality for uec-images.u.c but didn't upload. Anyways, testing local uec images will be ready soon, push to PPA, upload to Natty eventually, and backport to Maverick
[18:35] <smoser> i opened a feature freeze bug
[18:36] <smoser> and i thought i said that we'd need it uploaded before beta freeze.
[18:36] <smoser> sorry i wasn't more clear.
[18:39] <RoAkSoAx> smoser: It's my bad :(, since I thought you only wanted support for uec-images.u.c, but when you told me local uec image too, I wanted to have both before uploading, but anyways, we can always backport :)
[19:19] <pfifo> hi
[19:20] <pfifo> can just anyone use the public cloud that is shown on the site, and dose it cost anything?
[19:25] <patdk-wk> heh?
[19:25] <patdk-wk> shown on what site?
[19:25] <patdk-wk> only public cloud I know of is ec2
[19:25] <patdk-wk> I believe rackspaces cloud is internal only
[19:25] <pfifo> yes amazon ec2
[19:26] <patdk-wk> ya, ec2 costs
[19:26] <patdk-wk> your basically, renting a server from them
[19:26] <pfifo> i figured it would
[19:26] <patdk-wk> now, that you can rent it for just a few hours, vs per month, makes it cheaper, overall
[19:28] <pfifo> heres the sitrep, i just lost my harddrive last night, i had some really important data on it that is now lost in the void. Im searching around for something that will allow me to store files remotely. Looked into gmailfs but that appears to be slow and broken, i looked at ubuntu one and came across the ec2 thing. Any other places I can snag several gigs of data storage from?
[19:29] <pfifo> it dosent nesscarrily have to be reliable
[19:32] <gholms> If storage is what you're after what you really want is S3 as far as Amazon goes.
[19:32] <gholms> There isn't really much for free.  JungleDisk, which runs on top of S3, is designed for exactly what you want, but you have to pay for that as well.
[19:32] <_ruben> pfifo: a couple of usb flash disks containing multiple copies would probably be cheapest and also fastest
[19:32] <gholms> ^ this
[19:33] <gholms> ...
[19:33] <pfifo> if i was going to pay, i would do 10 a month to ubuntu
[19:34] <smoser> pfifo, so you're looking for a place to stash lots of data temporarily.
[19:34] <smoser> realistically, i say you either pay ubuntu $10 (thats nice)
[19:34] <smoser> or use s3.
[19:34] <gholms> Doesn't Ubuntu use S3 anyway?
[19:35] <smoser> if you put up 100G into s3, the storage would only cost you $10 for the first month.
[19:35] <pfifo> smoser: longterm, but i dont care if it gets deleted due to something like gmail tos violation cause if its free i can obivously make mutiple copies
[19:35] <smoser> you only need this for short time ?
[19:35] <pfifo> smoser: LONG term
[19:35] <smoser> oh.
[19:35] <smoser> sorry.
[19:36] <smoser> i thought you were just trying to migrate data.
[19:36] <smoser> actually, google offers a file storage option too
[19:36] <pfifo> smoser: no what i lost was my cross-tool chain and my code. (and some other crap)
[19:38] <smoser> http://picasa.google.com/support/bin/answer.py?hl=en&answer=39567
[19:38] <pfifo> the trick is though, it needs to be mountable or somehow (rsync) on my computer, none of this web interface crap, rapidshare would work for that
[19:38] <smoser> yeah.
[19:38] <smoser> maybe picasa interface
[19:38] <smoser> you just have to call them "pictures"
[19:38] <smoser> :)
[19:39] <pfifo> ever tried to upload an exe to docs? google throws a fit
[19:39] <smoser> http://code.google.com/apis/storage/docs/faq.html#pricing
[19:39] <gholms> smoser: Thanks for getting that posted.
[19:40] <smoser> there is that. its free !
[19:40] <pfifo> which they should how many users really make their own exes
[19:40] <smoser> for a while.
[19:40] <smoser> that is googles "cloud storage" (s3). they're beta and developers only stuff right now. so probably not a lot of tools available at the moment.
[19:41] <pfifo> google is cheap though
[19:42] <pfifo> is googles extra storage pricing compatable with google code too?
[19:42] <pfifo> cause that would be perfect, git would sync everything perfectly
[19:49] <gholms> smoser: I'm looking at DataSourceEc2.py here; are things like us-west-1b.ec2.archive.ubuntu.com supposed to be cnames that resolve to mirrors in the right region, or...?
[19:50] <smoser> pfifo, i dont know. somehow i doubt it.
[19:50] <smoser> yeah. it tries to lookup mirror based on availability zone
[19:50] <gholms> Is it only resolvable within EC2 or something?
[19:51] <harpette> hi there. Anyone has an idea why a "vmbuilder --mac=00:16:3e:10:34:31" builds a VM with <mac address='52:54:00:7f:93:28'/> ?
[19:51] <smoser> gholms, availability_zone[:-1]
[19:51] <smoser> us-west-1.ec2.archive.ubuntu.com , not us-west-1b.ec2.archive.ubuntu.com
[19:51] <gholms> Doh!
[19:51] <smoser> but they are only *reachable* inside ec2
[19:53] <gholms> Do they have a fixed set of IP blocks for each region?
[19:53] <pfifo> i hope i didnt get you guys sidetracked lol
[19:54] <gholms> Crap, looks like they don't.
[19:55] <pfifo> im out, thanks for the info guys
[20:00] <Datz> Hi, when I log on to the system, after recent updates, system statistics are not show if load is over 1. How can I change this threshold?
[20:03] <gholms> smoser: Instead of (or in addition to) adding "built-in" config data do you suppose cloud-init could read a distro-provided config file in /usr/share for defaults?
[20:04] <gholms> Sorry if I'm missing something here; I'm more or less asking questions as I go along.
[20:04] <smoser> "built in config data" ? youmean from /etc/cloud/cloud-init.cfg ?
[20:05] <smoser> err..
[20:05] <smoser> cloud.cfg
[20:05] <gholms> __init__.py in the cloudinit directory
[20:05] <gholms> That and DataSource.py have a bunch of default configuration info embedded into the source.
[20:06] <gholms> cloud.cfg goes in /etc, so it might be a bad place to specify, say, mirror locations since people can overwrite it completely.
[20:07] <smoser> i would not be opposed to it.
[20:07] <smoser> one thing you may realize is that its kind of messy...
[20:07] <gholms> ?
[20:07] <smoser> cloud.cfg is semi "cloud-init config" and "default #cloud-config"
[20:08] <smoser> that make sense ?
[20:08] <gholms> I'm not sure what the second part means.  :-\
[20:08] <smoser> i think there is one or two tihng that you can change in /etc/init./cloud.cfg that make no sense to "cloud config".
[20:09] <smoser> cloud_type for example.
[20:09] <smoser> so, cloud-config is the "easy config".
[20:09] <smoser> but cloud-init supports other types of input.
[20:09] <smoser> and largely, cloud.cfg just serves as the default "cloud-config"
[20:10] <smoser> there isn't really a file to configure cloud-init
[20:10] <smoser> make sense ? or did i make it worse.
[20:10] <gholms> More or less, yeah.
[20:10] <smoser> either way, i'm all for moving hard coded values out of __init__.py
[20:10] <smoser> :)
[20:11] <gholms> There needs to be a way to pick which modules should run on a given system.  cc_byobu.py makes no sense on Fedora, for instance.
[20:12] <gholms> So that might be the purview of such a config file.
[20:12]  * gholms hopes he's making sense
[20:17] <gholms> Whoops, there's no regular init script.
[20:17] <ideaman> Who can tell me whats the best way to figure out where my 100 megs of UDP usage a day is going on my server?
[20:17] <ideaman> Ubuntu 10.04
[20:18] <smoser> that does make sense, yes.
[20:18] <smoser> cc_byobu would makes sense on fedora (it is packaged there)
[20:19] <smoser> gholms, yeah, its upstart only
[20:19] <\sh> ideaman, tcpdump ?
[20:19]  * gholms might need to get a packaging committee exception
[20:19] <ideaman> will try that
[20:20] <ideaman> was trying bandwidthd and doing w3m, but not very comprensive
[20:20] <gholms> smoser: Would you be opposed to someone's submitting systemd configs to go along with the upstart ones?
[20:20] <smoser> gholms, no.
[20:21] <kirkland> Daviey: are you gone for the day?
[20:22] <Daviey> kirkland, sadly not
[20:29] <zul> kirkland: he is having fun with me
[20:52] <gholms> smoser: How do you guys get packages from the master package mirror to the S3 mirrors?
[20:52] <smoser> they're not s3
[20:52] <gholms> Or for that matter, how does Ubuntu's mirror infrastructure work inside EC2?
[20:52] <smoser> canonical runs ec2 instances in each region providing a apt mirror.
[20:53] <smoser> so it basically runs as any other mirorr would run.
[20:53] <gholms> Oof, that means an awful lot of inter-zone transfer fees.
[20:54] <gholms> I presume Canonical uses elastic addresses for those and just eats the data transfer fees?
[20:55] <ahasenack> hey, do you guys know if in UEC one instance can connect to another one using the external ip?
[20:55] <ahasenack> instead of the internal one, I mean
[20:55] <gholms> ahasenack: They can, but then you get charged transfer fees.
[20:55] <ahasenack> I managed to connect using the internal one and after having the security group authorizing the other group as the source, but not using the external ip
[20:56] <ahasenack> gholms: it works in ec2, right
[20:56] <gholms> Err, right.  I guess I shouldn't be speaking for UEC.
[20:56] <ahasenack> gholms: I'm wondering if it should work in UEC too or if it's a network setup problem in this particular UEC I'm using
[20:56]  * gholms shuts up and lets someone knowledgeable about it answer
[20:56] <ahasenack> it would be much simpler for my setup if it worked
[20:57] <ahasenack> I'm using SSL and I have the external hostname as the commonName of the cert...
[20:57] <ahasenack> and gnutls didn't like me putting other names in subjectAltName
[20:58] <patdk-wk> heh? if you use subjectaltname, commonname is ignored, so you have to duplicate the commonname in subjectaltname
[21:03] <RoyK> Mein Führer - I can walk!
[21:06] <_ng> :O)
[21:06] <RoyK> Dr. Strangelove ftgw
[21:06] <RoyK> Dr. Strangelove ftw
[21:10] <Datz> Hi, when I log on to the system, after recent updates, system statistics are not show if load is over 1. How can I change this threshold?
[21:12] <ScottK> Datz: It's a recent change that was on purpose (if perhaps misguided).  See the recent history of the ubuntu-server mailing list for details.
[21:16] <gholms> smoser: Why not import boto instead of using boto_utils.py?
[21:16] <smoser> it should import boto now.
[21:16] <smoser> the reason it has the staic copy is that lucid at one point was about to revert to boto 1.8
[21:17] <smoser> which didn't have the function that i needed.
[21:17] <gholms> Is that obsolete now, then?
[21:17] <smoser> yes
[21:17] <gholms> Cool
[21:17] <smoser> well, we would need to remove it
[21:17] <smoser> the boto_utils.py is used now.
[21:17] <smoser> but there is no reason for that any more.
[21:18] <gholms> Oof, there's nowhere to report bugs for that.
[21:19] <gm1959> can anyone please tell me how to get 'upstart' to not run the X server / gnome stuff?  There's no inittab to edit it out of?
[21:23] <|rt|> gm1959: upstart still reads the sysvinit rc runlevel files so you can still use the normal update-rc.d script to add and remove things from the various runlevels
[21:24] <gholms> smoser: Where should bug reports and patches and things for cloud-init go?
[21:25] <smoser> launchpad.
[21:25] <smoser> open agains the cloud-init project.
[21:26]  * gholms thinks https://launchpad.net/cloud-init needs to be set up the rest of the way since "Report a bug" doesn't work there
[21:27] <smoser> well that stinks.
[21:27] <smoser> let me look
[21:30] <smoser> gholms, fixed. now you can "Report a bug"
[21:30] <smoser> :)
[21:30] <gholms> :D
[21:31] <gholms> Does python << 2.7 mean that this stuff doesn't work with python 2.7?
[21:37] <gholms> (Because that would be a problem for me)
[21:44] <Datz> ScottK: thanks.. I'll see if I can hunt down that list
[21:45] <ScottK> Datz: archives are on lists.ubuntu.com.
[21:46] <Datz> great, thanks ScottK
[21:58] <Datz> humm.. ScottK do you perhaps know a specific range of days that the change might have taken place?
[21:59] <ScottK> Datz: This week.
[21:59] <ScottK> Today or yesterday.  They run together.
[22:06] <Datz> thanks
[22:12]  * Datz doesn't know what it would solve even if he could find it. :P
[23:45] <Madwill>  hey what do you guys use for os level virtualisation ?
[23:45] <Madwill> my clients use OpenVZ and i would like home to go for something more recent then ubuntu 8.04
[23:45] <Madwill> what it used now as lightweight container
[23:46] <SpamapS> LXC seems to be popular
[23:46] <SpamapS> And I believe is going into, or has been accepted into, the upstream kernel
[23:47] <Madwill> nice
[23:47] <Madwill> starting to read about it
[23:47] <Madwill> the poor dev's team made deprecated choices..
[23:48] <Madwill> i mean the poor men' dev team
[23:55] <soren> Yeah, LXC has been upstream for a while now.