[01:01] <eross> i installed ubuntu 64bit server on a VM and I can't access the internet although it should work. i think i bjorked my network settings, how do i reset them to 'default' ?
[02:23] <aot2002> I'm trying to activate curl with php but it just won't acknowledge the extension is installed.
[02:24] <twb`> Maybe ask #php
[02:27] <aot2002> it's currently invite only
[02:29] <twb`> That probably means you need to register your nick with freenode
[02:37] <aot2002> twb`, no it means that you cannot join because the channel is locked
[02:37] <aot2002> hence invite only
[02:39] <twb`> Plonk.
[02:39] <twb`> (Oh, it's ##php, you might be hitting the funky redirect misfeature.)
[02:40] <twb`> (Namely if you're already in the channel you're redirected to, you get an error.)
[02:40] <twb`> (Certainly, ##php is not invite only for registered users.)
[03:00] <patdk-lap> #php ##php :Forwarding to another channel
[03:00] <patdk-lap> heh, stupid
[03:00] <patdk-lap> and I'm registered
[03:00] <twb`> patdk-lap: yes, but he wasn't
[03:01] <patdk-lap> this channel redirection stuff on freenode really pisses me off though
[03:01] <twb`> I don't really see what the alternative is, though
[03:02] <twb`> Register #php and set /topic to "join ##php" ?
[03:02] <patdk-lap> if I wanted to join #php, I want to join #php not ##php
[03:02] <patdk-lap> if I can't join #php, ok, tell me, don't force me into ##php
[03:02] <twb`> I guess
[03:03] <twb`> FYI, it's ##foo not #foo because the latter is reserved for official channels
[03:03] <patdk-lap> same goes for #ubuntu+1 :)
[03:03] <patdk-lap> if I wanted #ubuntu, I would join #ubuntu :)
[03:04] <twb`> What, #ubuntu+1 redirects now?
[03:04] <twb`> That's lame
[03:04] <patdk-lap> it will, oct 31st
[03:04] <twb`> Why?
[03:04] <twb`> +1 will still be +1 after a release
[03:05] <patdk-lap> cause 12.04 won't exist in the build yet
[03:05] <patdk-lap> once the package system is done, it's unredirected
[03:05] <patdk-lap> happens every release cycle
[03:05] <patdk-lap> #ubuntu+1 closes for about a month
[03:05] <twb`> I guess given that ubuntu is mostly used by noobs, I can understand doing that to flush them out
[03:06] <patdk-lap> but, I hardly talk in ubuntu+1 anymore, moved more to ubuntu-testing now
[03:06] <patdk-lap> oh ya, I was suppost to file bugs tonight
[03:06] <patdk-lap> but stupid hp firmware died on me, and had emergancy downtime :(
[03:06] <twb`> Ubuntu doesn't have a testing alias :P
[03:07] <patdk-lap> #ubuntu-testing: Currently testing Oneiric Beta 1 candidates | http://qa.ubuntu.com/testing/iso-testing/ | http://iso.qa.ubuntu.com
[03:07] <twb`> I know I dropped #debian for #debian-next to reduce noise.
[03:07] <twb`> Obviously the right name should be #foo′ :p
[03:36] <[1]VampsDaBeast> hey guys. i have a question bout setting up a server to share external hard drives thru my network, how difficult is that to do
[03:42] <idlemind> [1]VampsDaBeast it depends. do you want to manage file sharing permissions to lock areas down do you want to mix in authentication? what os's are you working with?
[03:45] <[1]VampsDaBeast> idlemind, sorry bout that. net went out on me
[03:49] <idlemind> vampsdabeast np so you are running linux and windows boxes?
[03:52] <VampsDaBeast> idlemind, yes.. my network is a mix of both win and ubuntu.... i want to set up a box to share 3 external HD's
[03:52] <idlemind> i'd roll with samba
[03:53] <VampsDaBeast> can 10.04 setup with a LAMP/ISPConfig be able to handle that
[03:54] <idlemind> not sure what ISPConfig is but yes you can throw SAMBA on a server that has a LAMP stack installed
[03:54] <VampsDaBeast> ISPconfig = web hosting system
[03:54] <idlemind> if that LAMP box is connected to the Internet you may be opening up your data to any attack that gets into your server
[08:04] <jamespage> morning all
[08:12] <jamespage> Daviey: would you like me to pickup that extra euca security fix into the MP I have already raised for the SSLv3 compatibility?
[08:14] <Daviey> jamespage: that sounds like a great idea.
[08:16] <Daviey> jamespage: wow, this debdiff for asterisk is huge.
[08:16] <jamespage> yeah - makes be kinda nervous TBH
[08:17] <nigelb> woah.
[08:17] <nigelb> mysql installation just threw gibberish at my terminal
[08:17] <nigelb> Anyone ever seen something like that before?
[08:17] <koolhead11> hi all
[08:18] <koolhead11> nigelb: gibberish ? pastebin/screenshot
[08:18] <Daviey> nigelb: no, can you pastebin?
[08:18] <Daviey> hey koolhead11
[08:18] <koolhead11> hey Daviey :)
[08:18] <Daviey> nigelb: did you get your cron issue sorted btw?
[08:20] <Daviey> jamespage: $ wc -l *.debdiff
[08:20] <Daviey> 92404 asterisk_1.8.4.4~dfsg-2ubuntu1.debdiff
[08:20] <nigelb> Daviey: yep
[08:20] <Daviey> 10MB!
[08:20] <jamespage> yikes
[08:20] <jamespage> even though it is just bug fixes it feels like that really needs a FFE for sanity
[08:21] <twb> Grumble, should be using .debian.tar.gz by now
[08:21] <Daviey> Well that is how i'm treating it.
[08:21] <nigelb> koolhead11 / Daviey - http://people.ubuntu.com/~nigelbabu/mysql-wtf.png
[08:21] <Daviey> twb: uh?
[08:21] <nigelb> the dbconf prompt for root password had something similar too.
[08:22] <Daviey> nigelb: do you hav a funky locale set?
[08:22] <twb> Daviey: re .debdiff.  Except I guess that's a diff between debian and ubuntu or something, not upstream and ubuntu
[08:22] <twb> nigelb: something has asked to move from G0 to G1 mode
[08:22] <Daviey> twb: diff between ubuntu old and ubuntu new, bringing in debian changes.
[08:23] <twb> nigelb: you need to run "reset" (type it blind) to get back to normal
[08:23] <twb> Daviey: ok nm then
[08:23] <nigelb> Daviey: Nope, its en_US; twb - thanks!
[08:23] <nigelb> Oh hell.
[08:23] <twb> As to *why*; I don't know
[08:23] <nigelb> I managed to lock dpkg
[08:23] <twb> Usually it's when you cat a binary file
[08:24] <twb> I suppose it might be possible to do it by mashing your head on the keyboard, too
[08:24] <Daviey> nigelb is well known for head smashing the keyboard.
[08:24] <nigelb> :D
[08:25] <Daviey> nigelb: what was the instance ami id?
[08:25] <nigelb> I maintain summit! Its understood! :P
[08:27] <Daviey> nigelb: /etc/cloud/build.info
[08:27] <twb> Perl programmer hey
[08:28] <koolhead11> zul: hey
[08:28] <nigelb> Daviey:
[08:28] <nigelb> build_name: server
[08:28] <nigelb> serial: 20110719
[08:29] <twb> What happens if you do >1 build in a day?
[08:30] <Daviey> twb: .X
[08:30] <Daviey> For Kubuntu Oneiric it reached .8 yesterday. :)
[08:30] <Daviey> hah, .9, http://cdimage.ubuntu.com/kubuntu/daily/
[08:30] <twb> Fair enough
[08:31] <twb> I was gonna suggest following the DNS convention
[08:31] <Daviey> twb: zone files?
[08:31] <twb> yes
[08:31] <twb> YYYYMMDDNN
[08:32] <nigelb> YAY! All restored :-0
[08:32] <Daviey> twb: Yeah, i'm not a fan of having .'s in names.. but meh, it works.
[08:32] <Daviey> nigelb: Would be interesting to find out what caused it
[08:33] <twb> The .N scheme means you have to stop at 9 or break LC_COLLATE=C
[08:33] <nigelb> Daviey: I'm been pro yaking shaving this week :-)
[08:33] <Daviey> I've never seen that before following an apt install of a package
[08:33] <nigelb> Neither have I.
[08:33] <nigelb> I'm guessing I've screwed up mysql totally.
[08:34] <nigelb> I've been setting it up since 9 in the morning. Its 2 pm now.
[08:34] <twb> Is it a mysql package you just built?
[08:34] <Daviey> twb: TBH, if we roll >9 images in a day, we are doing it wrong™
[08:34] <twb> Daviey: optimist
[08:34] <nigelb> twb: I don't build packages for production :)
[08:34] <twb> nigelb: so how did YOU screw it up?
[08:34] <nigelb> twb: Actually, I think he means given the time it takes, there may not be enough time :)
[08:34] <nigelb> twb: chaging datadir
[08:34] <nigelb> its always a mess.
[08:34] <nigelb> Thank you apparmor.
[08:35] <Daviey> twb: The reason there were 9 yesterday for Kubuntu, was due to archive inconsistencies on the build archive.  Ie, a bug.
[08:35] <Daviey> so it was restarted multiple times.
[08:48] <Daviey> jamespage: Just gone through the changelog, and it does seem to all be bug fixes, High/Medium and making stuff DTRT.
[08:48] <jamespage> Daviey: yes - that was my take as well
[08:48] <jamespage> but there are **alot** of them
[08:57] <Daviey> jamespage: i'm going to try the package myself for a bit.. to see if it WFM.
[08:57] <Daviey> https://launchpad.net/~davewalker/+archive/junk/+sourcepub/1949848/+listing-archive-extra
[09:04] <nigelb> oh, WIN.
[09:04] <nigelb> mysql doesn't start up anymore \m/
[09:06] <nigelb> Anyone knows which package has my.cnf?
[09:07] <RoyK> mysql-server
[09:07] <nigelb> thanks!
[09:08]  * popey hands nigelb a "dpkg -S /etc/mysql/my.cnf"
[09:08] <nigelb> popey: I don't have the file :D
[09:09] <nigelb> popey: Right. I could just run it on the laptop. Thanks!
[09:09] <twb> Even if you deleted it, if it's a connfile it'll still be listed by dpkg -S / dlocate
[09:09] <popey> indeed
[09:09] <popey> also
[09:09] <nigelb> AHH!
[09:09] <twb> However if it's e.g. ucf'd, then you'd need to read /var/lib/dpg/info/...
[09:09]  * popey hands nigelb http://packages.ubuntu.com/search?searchon=contents&keywords=my.cnf&mode=exactfilename&suite=natty&arch=any
[09:09] <nigelb> <-- sucks as a sysadmin.
[09:09] <twb> nigelb: you are not alone
[09:10] <twb> Much of my job is cleaning up after that kind of sysadmin.
[09:11]  * koolhead11 thinks going to be tough day figuring out keystone 
[09:40] <jamespage> Daviey: MP for euca updated with the clock difference patch
[09:40] <jamespage> just testing it now
[09:47] <Daviey> jamespage: rocking
[09:49] <jamespage> i love cobbler - makes tearing down and rebuilding systems for testing so much quicker
[09:51] <koolhead11> jamespage: totally agree,  although am yet to find way to provide custom partitioning
[09:51] <koolhead11> :(
[09:52] <hkraal> Hi, I'm having a problem with a fresh installed Ubuntu 10.04.3 LTS server with MySQL 5.1.41-3ubuntu12.10. I've imported 1500 databases, 10 GB, (including the mysql DB) and copied the /etc/mysql/debian.cnf from the old (Ubuntu 8.04.4 LTS, MySQL 5.0.51a-3ubuntu5.7)  to the new server. All went fine and everything is up and running. The only problem I've is debian-start is running 16 times pushing the load to 17 and keeps doing this continuesly. Is anyone
[09:52] <hkraal> fammiliar with this kind of behaviour?
[09:53] <koolhead11> zul: there is no keystone configuration file with pkg keystone?
[09:55] <hkraal> some log and ps aux information can be found here: http://pastie.org/private/ivf3wb7jzuorq3mdlazfya
[10:06] <koolhead11> Daviey: ping
[10:33] <Daviey> hey koolhead11
[10:33] <lambda_x> I have one 8.10 ubuntu-server, is it possible to upgrade this system to LTS?
[10:33] <lambda_x> (8.10 repositories are currently unavailable)
[10:36] <koolhead11> Daviey: am uninstalling the keystone package as well and trying the pkg from github. I have no luck in running keystone yet :(
[10:44] <Daviey> koolhead11: Yeah, keystone is erm, interesting at the moment
[10:45] <Daviey> lambda_x: yes
[10:45] <Daviey> !oldreleases | lambda_x
[10:45] <koolhead11> Daviey: whole openstack is running around keystone :D
[10:52] <Daviey> koolhead11: well yes.. did you see the meeting last night?
[10:56] <koolhead11> Daviey: read the final outcome of the meetup. :D
[10:56] <koolhead11> long mail with plans
[10:56]  * koolhead11 wants to see dashboard replaced from django to something simpler :P
[11:00] <Daviey> koolhead11: nah, django is a good choice.
[11:02] <koolhead11> Daviey: i would love to see http://flask.pocoo.org/ instead :D
[11:02] <koolhead11> i find django jumbo app :D
[11:03] <AdamBlack> im trying to figure out a way to set up an incremental backup, to backup .tar files from server1 to server2 over a WAN.. would RSYNC work or is there something better
[11:04] <zul> koolhead11: son of a....can you open up a bug about that
[11:04] <ersi_> rsync is freggin' awesome. It's up for the task.
[11:04] <Daviey> AdamBlack: you need tarballs?
[11:05] <Daviey> rsync is not good for doing incremental tarballs, is it ersi_ ?
[11:05] <koolhead11> zul: hehe. i thought i should not. :D
[11:05] <koolhead11> zul: i will replace ... to 'god' :D
[11:06] <ersi_> Daviey: True, not to my knowledge atleast. I was more on track about that rsync is great at transfering. The data to be transferred should be fixed prior to rsync use ;p
[11:07] <AdamBlack> Daviey: Server1 hosts websites and uses DirectAdmin, DA stores the site backups in tarball and nobody knows where the sites are stored on the server(just joined this place) its not usuall www
[11:08] <Daviey> AdamBlack: DirectAdmin.. :(
[11:08] <Daviey> Can't easily help with that myself.
[11:09] <AdamBlack> worst case scenereo i can backup the whole server1 and scp it to server2 :P
[11:10] <rbasak> AdamBlack: Try my tool, www.synctus.com/ddar. It'll de-duplicate what it transmits over the wire and store it in an archive at the remote end.
[11:11] <rbasak> AdamBlack: it'll work for tarballs - letting you store/trasmit multiple ones that are similar efficiently
[11:11] <Daviey> rbasak: interesting!
[11:11] <AdamBlack> rbasak: having a look now :D
[11:13] <koolhead11> rbasak: hey.
[11:13] <rbasak> hey koolhead11!
[11:14] <koolhead11> zul: okey, am reporting it sir. :D
[11:15] <Daviey> koolhead11: less bugs, more patches :P
[11:16] <koolhead11> Daviey: :)
[11:30] <eagles0513875> hey guys im running into an issue where i cannot set my locale when running tzdata anyone else experiencing this issue
[11:30] <eagles0513875> im on 10.04.3
[11:32] <eagles0513875> this is the output im getting
[11:32] <eagles0513875> http://pastebin.com/byk0TbNT
[11:39] <pmatulis> eagles0513875: do you want me to google that for you?
[11:40] <eagles0513875> no need just the first time im seeing that
[11:40] <eagles0513875> pmatulis: i ran dpkg-reconfigure tzdata and set it to my right time zone yet it still brings what i pasted up
[11:45] <eagles0513875> pmatulis: why are these msg's popping up about lc_ctype cannot change and warning about setlocale lc_ctype cannot change locals when i do a tab completion
[12:19] <Jeeves_> Anyone familiar with Symfony (PHP) here?
[12:19] <koolhead11> https://bugs.launchpad.net/openstack-dashboard/+bug/855467
[12:19] <koolhead11> Jeeves_: ?
[12:20] <Jeeves_> koolhead11: If anyone is familiar with it
[12:21] <koolhead11> Jeeves_: you can ask the question if its related to ubuntu server i suppose :)
[12:23] <Error404NotFound> i am about to launch 2 webservers. Both of these will host same app code. I want both server's nginx, php conf and app code to be in sync. Whats the best way? NFS? cron'ed rsync? something else?
[12:51] <jamespage> Daviey: euca appears to still work with both of those patches applied
[12:51] <zul> good morning
[12:52] <jamespage> morning zul
[12:52] <jamespage> good if you want to sponsor - https://code.launchpad.net/~james-page/ubuntu/oneiric/eucalyptus/fix-sslv3-compat/+merge/76258
[12:52] <zul> but i would be tainted
[12:57] <Daviey> jamespage: you've missed the fact that you can upload that yourself
[12:57] <Daviey> :)
[12:57] <jamespage> I don't think so
[12:57] <jamespage> I used to be able to
[12:58] <Daviey> ah
[12:58] <jamespage> nope - I can't :-)
[12:58] <Daviey> Yes, sorry.. you can for SRU, oddly :)
[12:58] <Daviey> jamespage: you used to be trusted, but not anymore :)
[12:58] <jamespage> yes - but thats because its generated from the seed
[12:59] <jamespage> and the seed if difference pre oneiric
[12:59] <Daviey> wel yes
[13:00] <jamespage> zul: I saw that cgroups/libvirt issue again last night - bug 842845
[13:01] <zul> jamespage: bah
[13:01] <jamespage> but I was only trying to start one instance this time :-(
[13:01] <zul> jamespage: did you switch between like kvm to lxc?
[13:01] <jamespage> nope - it was a fresh install with just lxc only
[13:01] <Daviey> In other news, turns out rice cooked in a microwave is less than good.
[13:02] <zul> Daviey: duh..
[13:02] <jamespage> I started up a 4 OK (in a single reservation) - left it overnight and then tried to start one this morning - quack quack oooppps
[13:02] <zul> jamespage: can i get acccess to the machine?
[13:02]  * koolhead11 kicks himself
[13:02] <jamespage> zul: nope - I had to scrub it to test the euca fixes :-(
[13:03] <zul> jamespage: doh
[13:03] <jamespage> but I'll set it up again ~ 20 mins
[13:03] <zul> jamespage: ok lemme know when you can give me access
[13:03] <jamespage> sure
[13:05] <Daviey> zul: whilst you are waiting, you could setup ipv6 to connect to jamespage's network. :)
[13:05] <jamespage> that you could
[13:05] <zul> jamespage: how can i do that?
[13:06] <jamespage> I'm IPv6 enabled so all my servers get globally accessible IP addresses
[13:06] <zul> okies...meaning :)
[13:06] <jamespage> I'll have to permit you access through my firewall
[13:07] <jamespage> zul: you don't have IPv6?
[13:07] <zul> jamespage: ok
[13:07] <Daviey> zul: forwarding you a mail
[13:07] <zul> jamespage: no
[13:07] <jamespage> :-(
[13:07] <zul> still old school
[13:07] <Daviey> zul: chuck.shortk right?
[13:07] <zul> Daviey: eh?
[13:08] <Daviey> zul: you have mail.
[13:08] <zul> oh goody i love email
[13:08] <jamespage> this might be a stupid question
[13:08] <jamespage> but how do I easily see what preseed configuration parameters a package has?
[13:09] <rbasak> What's the preferred way of getting a nova test environment going on oneiric? I'm trying https://github.com/vishvananda/novascript/blob/master/nova.sh but it doesn't seem to want to work with lp:ubuntu/nova
[13:10] <jamespage> rbasak, all-in-one install try this - https://wiki.ubuntu.com/ServerTeam/Oneiric/OpenStackTestPlan
[13:10] <Daviey> jamespage: something like, $ debconf-get-selections  | grep mysql
[13:10] <Daviey> ?
[13:11] <rbasak> jamespage: thanks, but can I run out of my own nova source tree?
[13:12] <Daviey> rbasak: if using lp:nova, i would use nova.sh
[13:12] <rbasak> Daviey: nova.sh works with lp:nova but evidently not with lp:ubuntu/nova
[13:13] <Daviey> rbasak: if using lp:ubuntu/nova, i'd probably generate a deb TBH
[13:13] <rbasak> I have a patch that may or may not work but I can't get a sensible environment to test it in
[13:24] <lynxman> morning everyone
[13:35] <aubre> does anyone know how to make iscsi connection logins persistent across reboots on the initiator ? using 10.04 server LTS
[13:45] <koolhead11> jamespage: i don`t see mysql there :P
[13:45] <jamespage> it does not need it
[13:46] <jamespage> for all-in-one anyway
[13:48] <soren> Daviey: Is there any documentation at all for the orchestra stuff?
[13:48] <soren> Daviey: I've installed ubuntu-orchestra-provisioning-server. Now what?
[13:48] <RoAkSoAx> soren: in progress
[13:48] <soren> RoAkSoAx: Does that mean "no"?
[13:48] <RoAkSoAx> soren: there is
[13:48] <soren> Great! :)
[13:48] <soren> Where?
[13:49] <RoAkSoAx> soren: but im in the mprocess of writing it on help.u.c
[13:49] <RoAkSoAx> soren: starting to
[13:49] <lynxman> jcastro: ping
[13:49] <soren> RoAkSoAx: Where is it now?
[13:50] <RoAkSoAx> soren: private
[13:50] <koolhead11> jamespage: earlier we used mysql for all in one on natty and were thinking the same for oneiric. :d
[13:50] <soren> I see.
[13:50] <koolhead11> hey soren
[13:50] <soren> The problem is that from what I can see, it's completely and utterly broken, but the docs might tell me that this is how it's supposed to (read: not) work.
[13:50] <hkraal> Hi, I'm having a problem with a fresh installed Ubuntu 10.04.3 LTS server with MySQL 5.1.41-3ubuntu12.10. I've imported 1500 databases, 10 GB, (including the mysql DB) and copied the /etc/mysql/debian.cnf from the old (Ubuntu 8.04.4 LTS, MySQL 5.0.51a-3ubuntu5.7)  to the new server. All went fine and everything is up and running. The only problem I've is debian-start is running 16 times pushing the load to 17 and keeps doing this continuesly. Is anyone
[13:50] <hkraal> fammiliar with this kind of behaviour? some log and ps aux information can be found here: http://pastie.org/private/ivf3wb7jzuorq3mdlazfya
[13:51] <Daviey> soren: what specific issues are you seeing?
[13:51] <soren> Daviey: I don't know. What is it supposed to do?
[13:52] <soren> Almost every conceivable answer I've been able to come up with for that answer, it doesn't do.
[13:52] <Daviey> soren: I was expecting you to reply with package conflicts, or a traceback or something!
[13:52] <soren> So maybe I'm not veyr good at guesssing what it's meant to do.
[13:52] <soren> I install the package. It goes... err... fine, I guess. It asks a bunch of questions. I provide answers. When it's done, I haven't a clue what to do.
[13:52] <soren> There's a new file in /etc/apache2/sites-available.
[13:53] <soren> (even thought it's not a site)
[13:53] <soren> ...I enable that..
[13:53] <soren> and try to access the URL it "suggests".
[13:53] <soren> and I get a 403.
[13:53] <soren> It points to an empty dir, so it's not very interesting anyways.
[13:53] <RoAkSoAx> soren: what dir?
[13:53] <soren> So maybe I'm doing this *all* wrong.
[13:53] <soren> /var/lib/orchestra/www/
[13:53] <soren> RoAkSoAx: ^
[13:54] <RoAkSoAx> soren: that might be just leftovers
[13:54] <soren> From what? I *just* installed it.
[13:54] <soren> It's not been half an hour.
[13:54] <RoAkSoAx> soren: from previous orchestra version 1.x
[13:54] <RoAkSoAx> soren: hehe just hold on for a while and I'll have some documentation ready
[13:54] <soren> In the packages or on my system?
[13:54] <RoAkSoAx> soren: packages
[13:54] <soren> Ah.
[13:55] <soren> Am I installing the wrong pacakge, perhaps?
[13:55] <soren> If I want the thing that can provision new servers, what should I have installed?
[13:55] <smoser> soren, i think you're not in RoAkSoAx's target audience
[13:55] <smoser> :)
[13:55] <RoAkSoAx> soren: ubuntu-orchestra-provisioning-server
[13:55] <soren> smoser: I'm sure I'm not.
[13:55] <soren> RoAkSoAx: YEah, that's what I got.
[13:55] <soren> SHould I be using a PPA instead?
[13:55] <RoAkSoAx> soren: so that should have pulled cobbler
[13:56] <RoAkSoAx> soren: then imported ISO's automatically to cobbler
[13:56] <RoAkSoAx> soren: now you need to add systems into cobbler
[13:56] <soren> I'd be *happy* to report these as bugs, but I don't know if they are. I might be doing this *all wrong* and it's behaving exactly like it's meant to.
[13:56] <smoser> you should have a cobbler web ui at http://127.0.0.1/cobbler_web
[13:56] <soren> cobbler*_web*.
[13:56] <soren> Of course.
[13:56] <soren> Why didn't I guess that?
[13:56] <soren> Great.
[13:57] <soren> What's the username? I was only asked for a password during install.
[13:57] <smoser> cobbler
[13:57] <smoser> as root, you can run 'cobbler' commands
[13:57] <smoser> and it has useful-ish help
[13:57] <smoser> but i agree that the http:// is completely un-discoverable
[13:57] <andreserl> soren, cobbler
[13:58] <andreserl> soren, it tells you that the username is cobbler and you need a password for it xD
[13:58] <Daviey> smoser / soren: think it requies a debconf note?
[13:58] <smoser> i dont know. i think soren should have just known to go to /cobbler_web
[13:58] <smoser> seriously
[13:58] <soren> Daviey: debconf note, mention in package description, in /usr/share/doc/<whatever>/README, on help.ubuntu.com. ANywhere, really. Just *somewhere*.
[13:58] <soren> Not just in the logs from this channel :)
[13:59] <Daviey> soren: google will parse this channel shortly.
[13:59] <smoser> fwiw, i only knew cobbler_web because it is in my browswer history
[14:00] <zul> ok next upload will ship configuration files for keystone
[14:00] <Daviey> zul: er, for what part?
[14:02] <zul> Daviey: it doesnt ship configuration files in the first place so it doesnt know about the sqlite database that it uses
[14:04] <Daviey> zul: ah, ok - you are not coupling anything to keystone, right?
[14:04] <zul> Daviey:  hell no im just making it usuable
[14:04] <jamespage> smoser, utlemming: beta-2 ec2 tests look OK - a couple of failures where ec2 failed to provision anything
[14:05] <jamespage> and one where the network did not come up on first boot - http://tinyurl.com/6dxsazg
[14:06] <Daviey> zul: hmm.. we don't want stuff easy to use, do we? :)
[14:06] <koolhead11> RoAkSoAx: supp?
[14:07] <zul> Daviey: of course not :)
[14:08] <RoAkSoAx> koolhead11: nothing much, you?
[14:09] <koolhead11> RoAkSoAx: hitting head against keystone. planning to finish my cobbler left over work. Am yet to able to succeed in automated custom partitioning. :(
[14:11] <lynxman> Daviey: easy of use? pft
[14:13] <koolhead11> hey lynxman
[14:13] <lynxman> hey koolhead11 o/
[14:18] <zul> jamespage: can you try applying this patch? people.canonical.com/~chucks/nova-get-info-exception.patch
[14:18] <jamespage> zul: sure
[14:32] <koolhead11> rbasak: have you documented the cobbler VM install steps somewhere?
[14:35] <rbasak> koolhead11: I ended up using http://pad.ubuntu.com/arm-server-netboot (not by me)
[14:41] <jamespage> zul - testing that now
[14:42] <Error404NotFound> i am about to launch 2 webservers. Both of these will host same app code. I want both server's nginx, php conf and app code to be in sync. Whats the best way? NFS? cron'ed rsync? something else?
[14:46] <koolhead11> Error404NotFound: rsync
[14:46] <koolhead11> i have herad of people using for the same
[14:46] <zul> rsync
[14:46] <koolhead11> *hear :D
[14:47] <Error404NotFound> what about the part which would involve creating new vhosts? i would also have to do a graceful web server restart for those rsync'ed files.
[14:50] <koolhead11> Error404NotFound: i am not very sure of your application but what you can do is check for nginx as load balancer and rsyc to sync up data on both the systems.
[14:51] <Error404NotFound> hmmm, i went another way around, my scripts test for /etc/nginx.reload and if it exists, reload nginx, deletes that from from source as well as destionation.
[14:51] <Error404NotFound> destination*
[14:51] <Error404NotFound> though still one problem : single point of failure.
[14:51] <koolhead11> rbasak: https://gist.github.com/1205825 is not accessible to me. am i the only one in world facing that ?
[14:52] <Error404NotFound> koolhead11: yes
[14:53] <rbasak> koolhead11: wfm
[14:53] <hkraal> http://www.downforeveryoneorjustme.com/gist.github.com/1205825
[14:55] <koolhead11> rbasak: hkraal seems like websensce is culprit
[14:55] <Daviey> smoser: you were asking why nova hasn't built?
[14:55] <Daviey> Since the last nova built, it was promoted to main.
[14:55] <smoser> i was, how did you know, Daviey
[14:55] <Daviey> glance isn't yet in main, and is a build dep
[14:56] <smoser> ok, so lets pretend for a moment that i dont care why
[14:56] <smoser> and i want to be able to install nova
[14:56] <smoser> just say someone actually wanted to try that
[14:56] <Daviey> jd-strand is reviewing glance as we speak, once that is good, we are GOLD>
[14:56] <smoser> ok.
[14:57] <Daviey> smoser: then the prior built binary will be installed.
[14:57] <Daviey> Until a binary is produced, it will not be superseeded.
[14:57] <smoser> complete with crippling bugs
[14:57] <smoser> :)
[14:57]  * Daviey cripples smoser 
[14:57] <smoser> knowledge of this came from agy telling me that there were no nova updates for Canonistack.
[14:59] <koolhead11> Error404NotFound: http://wiki.nginx.org/LoadBalanceExample  this is what i meant
[14:59] <Error404NotFound> hkraal: isup.me
[15:02] <Error404NotFound> koolhead11: even with this, configs would have to be copied between both backends and also single point of failure with respect to copying configs.
[15:04] <Daviey> smoser: http://pb.daviey.com/qdVg/
[15:05] <smoser> suck
[15:06] <smoser> open a bug on this.
[15:18] <Daviey> smoser: i think you need to, $ ubuntu-bug ubuntu-bug
[15:19] <smoser> no
[15:19] <smoser> i think i discovered the problem
[15:19] <smoser> and i feared this would happen
[15:19] <smoser> :)
[15:19] <smoser> i think its because 'telnet 169.254.169.254 80' hangs
[15:20] <Daviey> smoser: uh?
[15:24] <Daviey> smoser: can you explain?
[15:25] <smoser> apport checks to see if it is in ec2
[15:25] <smoser> by checking for the metadata service
[15:26] <smoser> Daviey, ok. any idea why its not a genuine package?
[15:27] <smoser> so i veriifed that the 169.254.169.254 was the issue
[15:27] <smoser> i did:
[15:27] <smoser> route add -host 169.254.169.254 reject
[15:27] <smoser> which can then be removed by
[15:27] <smoser> route del -host 169.254.169.254 reject
[15:30] <CluelessPerson> h5
[15:30] <CluelessPerson> '
[15:30] <CluelessPerson> I'm getting an error whenever I boot up my server.
[15:31] <CluelessPerson> init: ureadahead-other main process (712) terminated with status 4
[15:31] <CluelessPerson> [   103.509461] [drm:pch_irq_handler] *ERROR* PCH poison interrupt
[15:32] <CluelessPerson> What does this mean?  google is unhelpful.  although I do think it does have something to do with checking the disks.
[15:32] <Daviey> CluelessPerson: What ubuntu release?
[15:33] <CluelessPerson> 11.04
[15:33] <CluelessPerson> before that error, a couple lines above it speaks about forcing a check on the drives, it being mounted 38 times without a check
[15:34] <Daviey> CluelessPerson: Looks like a kernel bug, and known.
[15:34] <Daviey> It would help i you raise a bug, or continue the discssuion in #ubuntu-kernel
[15:35] <CluelessPerson> I did see a post about a bug related, but didn't realize that this was specifcially it
[15:35] <CluelessPerson> I'll look in to i
[15:35] <Daviey> it might not be, but the kernel folks will be able to help more.
[15:35] <CluelessPerson> thanks
[15:37] <smoser> Daviey, bug 855651 opened for the apport issue we saw.
[15:39] <Daviey> cool
[15:42] <smoser> bug 855660 for the DescribeInstances bug that we were seeing which now seems gone.
[15:43] <Daviey> wron
[15:43] <Daviey> g
[15:55] <Daviey> adam_g: around?
[15:58] <adam_g> Daviey: hey
[16:01] <Daviey> adam_g: rocking
[16:01] <lynxman> adam_g: morning ;)
[16:12] <Daviey> adam_g: can you see my message?
[16:24] <smoser> is/has anyone (utlemming, jamespage) looked at results at https://jenkins.qa.ubuntu.com/job/oneiric-server-ec2/
[16:25] <jamespage> smoser: I took a run through earlier
[16:26] <smoser> did we opne bugs ?
[16:26] <smoser> or mark failures as a particular bug ?
[16:26] <smoser> is there any sane way to do that in jenkins ?
[16:27] <jamespage> there are only three test failures
[16:27] <jamespage> two looked like ec2 rather than Ubuntu errors - machines where never provisioned
[16:27] <jamespage> and one was lack of network on first boot
[16:28] <zul> jamespage: shazbutt
[16:28] <jamespage> smoser: http://tinyurl.com/6zfumu2
[16:28] <jamespage> as utlemming just asked me to re-run against the most recent images I stopped filing results
[16:29] <smoser> oh. those are against yesterdays images ?
[16:29] <smoser> we really need to improve our test suite.
[16:29] <smoser> as openssl was broken for https traffic yesterday :)
[16:29] <jamespage> smoser: 20110920
[16:30] <smoser> jaso i ask a general question....
[16:30] <smoser> we're going to be running these tests more often (i hope)
[16:30] <smoser> is there a way that we can attach bug numbers and or data to the results ?
[16:32] <jamespage> smoser: yes
[16:32] <jamespage> https://jenkins.qa.ubuntu.com/job/oneiric-server-ec2/ARCH=i386,REGION=ap-southeast-1,STORAGE=instance-store,TEST=simple-user-data,label=ubuntu-server-ec2-testing/
[16:33] <jamespage> for example
[16:36] <smoser> ah. ok. that is useful somewhat.  that comment there i son the parituclar test ?
[16:36] <smoser> that is nice.
[16:37] <Daviey> smoser: great! instances refuse to die
[16:38] <Daviey> ah, just slow to die for some reason
[16:38] <smoser> i think they're dead now
[16:39] <Daviey> yeah
[16:39] <smoser> that is know behavior
[16:39] <smoser> terminate-instances; list
[16:39] <smoser> will show terminated instnace as 'running'
[16:40] <adam_g> i think i saw that yesterday
[16:40] <adam_g> instance is dead, /var/lib/nova/instances/instance-* is deleted, etc. but updating host status in db seems delayed
[16:41] <adam_g> or instance status
[16:45] <lynxman> I had that several times myself, I had to delete the objects from the DB manually at the end
[16:53] <zul> heh....im dont dead im just resting
[16:53] <zul> sorry i watched monty python last night
[16:57] <lynxman> zul: bring out yeh deaaad
[17:04] <Daviey> smoser: do you have a euca-terminate-all helper?
[17:05] <Daviey> I made one a while ago, but wonderd if you had anything clean?
[17:07] <smoser> i can make one
[17:08] <Daviey> smoser: mine was euca-describe | grep running | awk | -> for i in that ; terinate $i
[17:09] <Daviey> smoser: see win #4
[17:19] <Daviey> 011-09-21 13:18:53,070 WARNING nova.scheduler.manager [-] Driver Method schedule_run_instance missing: 'ChanceScheduler' object has no attribute 'schedule_run_instance'.Reverting to schedule()
[17:19] <Daviey> nice
[17:21] <adam_g> oh
[17:21] <adam_g> hold on
[17:21] <Daviey> adam_g: do you have an idea why only the first node is working?
[17:22] <adam_g> yes
[17:41] <kirkland> RoAkSoAx: yo!
[17:41] <kirkland> RoAkSoAx: mtaylor has a some ideas about the pxe preseed opts in Orchestra
[17:41] <kirkland> RoAkSoAx: you around?
[17:42] <lynxman> kirkland: that sounds interesting :)
[17:42] <kirkland> mtaylor: RoAkSoAx: I'm nibbling on lunch, so make sure you highlight my nick when you speak :-P
[17:42] <mtaylor> RoAkSoAx: so - I was checking in with kirkland about whether or not orchestra had done anything about preseed's inability to properly deal with multiple nics. (last I checked, I needed to add "netcfg/choose_interface=eth0" to the kopts even if I had that in the preseed file)
[17:44] <mtaylor> RoAkSoAx: so I guess the inquiry is in two parts - is there a way to get preseed itself to do what I'm asking it to do without having to pass kopts, and if not, is there a way to make orchestra scan through stuff and do the right thing so that folks don't get screwed by doing a preseed config that still asks menu questions
[17:47] <RoAkSoAx> mtaylor kirkland im on way to lunch too can we talk in ~1hr? im from cell atm
[17:47] <smoser> mathiaz, meet utlemming .
[17:47] <kirkland> mtaylor: what would your desired behavior be?
[17:48] <mtaylor> RoAkSoAx: sure
[17:48] <smoser> utlemming, is doing much of the ec2 work now. you can ping him if i'm not available for sure.
[17:49] <kirkland> mtaylor: ie, you don't want it to be an orchestra-wide setting, do you?
[17:49] <kirkland> mtaylor: different machines might have different nic counts, right?
[17:49] <mathiaz> smoser: great -thanks!
[17:50] <kirkland> matttthiiiiiiiiiaaaaaaaazzzzzzzzzzzzzzzzzzzzzzzz
[17:50] <mathiaz> kirkland: :)
[17:50] <kirkland> :-)
[17:50] <mathiaz> kirkland: how are you doing/
[17:50] <mathiaz> ?
[17:50] <kirkland> mathiaz: well, man :-)
[17:50] <kirkland> mathiaz: et toi?
[17:50] <Daviey> mathiaz: I seem to remember you saying, "I won't be a stranger"
[17:51] <smoser> yes, its good to see the ubuntu hero mathiaz around.
[17:51] <smoser> we were just swearing^H^H^H^H^H^H talking about some of your work yesterday
[17:52] <mathiaz> smoser: ah nice! Hopefully some of it is still used :)
[17:52] <mathiaz> kirkland: doing well here
[17:52] <smoser> hoping to get some of the euca-stress tests pointing at openstack
[17:55] <mtaylor> kirkland: well, thing number one is that if I put netcfg/choose_interface=eth0 into a preseed file, I'd love for that to work
[17:56] <DanaG> weird... looking at my management tool's results for chkdsk eventlogs, I see some computers where it did this:  Check C:, fix C:.  Reboot.  Check D:, fix D:.  Check D:. No errors.
[17:56] <DanaG> Note that the "C:" is one event, and the "D:" pair is another event... yet the TimeGenerated field is the same on both.
[17:56] <mtaylor> kirkland: BUT - if that doesn't work, it would be a nice hack if orchestra could look in the preseed file for netcfg/choose_interface and if it's there toss something into the kopts
[17:56] <zul> mathiaz: hey!
[17:58] <DanaG> oops, wrong channel.
[17:59] <lynxman> DanaG: :D
[18:01] <SpamapS> mtaylor: is this a well known bug in netcfg or something?
[18:02] <mtaylor> SpamapS: afaik - I learned about the kopt workaround from a blog somewhere a couple of months ago
[18:02] <mtaylor> SpamapS: so, I think so
[18:06] <mtaylor> SpamapS: related (marginally) ... how does one connect a node in orchestra to a particular set of juju charms (that's never going to get easy to say)
[18:08] <SpamapS> mtaylor: juju wants to be the thing that tells orchestra to turn a machine on / configure it..
[18:08] <SpamapS> mtaylor: so you use the orchestra provider
[18:08] <mtaylor> SpamapS: AH
[18:08] <kirkland> mtaylor: orchestra sort of becomes your EC2 or OpenStack, so to speak
[18:08] <SpamapS> mtaylor: I believe adam_g is working on a blog post for Juju+Orchestra == OpenStack
[18:13] <lynxman> darn, somebody released mcollective on debian after our package, now we have delta :(
[18:13] <zul> lynxman: how big of a delta?
[18:14] <lynxman> zul: afaict biggest thing being the recommend activemq instead of rabbitmq
[18:14] <lynxman> zul: and a couple small details
[18:14] <zul> lynxman: do you know why that is?
[18:14] <lynxman> zul: I have a friend who's a maintainer in debian, he recommends me to join the maintenance team for the package :)
[18:14] <zul> lynxman: good idea
[18:14] <lynxman> zul: because upstream recommends activemq
[18:15] <mtaylor> maybe because activemq doesn't require erlang?
[18:15]  * mtaylor hides
[18:15] <zul> erlang is off the hook
[18:16] <lynxman> mtaylor: well player sir, well played
[18:16] <lynxman> s/player/played/
[18:16] <lynxman> zul: I think they'll eventually switch to rabbitmq tbh
[18:17] <lynxman> zul: anyhow I need to get involved :)
[18:23] <kpettit> any good opensource database browser/query tools for talking with M$ databases?
[18:26] <lynxman> kpettit: odbc I reckon should do the job, that's how I do it normally
[18:27] <kpettit> I'm not that familiar with dealing with ODBC.  I just need to be able to talk to it so I can run sql queries and such.
[18:27] <kpettit> is there good ODBC query/browser tool?
[18:33] <SpamapS> I think the only reason upstream recommends activemq is that its what they started with.
[18:33]  * SpamapS said referring to mcollective
[18:34] <SpamapS> kpettit: I believe libreoffice-base can use any odbc driver
[18:34] <kpettit> worth a try.  thanks
[18:35] <lynxman> SpamapS: yeah actually mcollective upstream recommends activemq as well, but I'm in close contact with the author and he blesses our use of rabbitmq :)
[18:35] <lynxman> SpamapS: he's thinking about switching to rabbitmq at some point
[18:39] <SpamapS> lynxman: right, thats what I mean.. AMQ is just what he used first, so he can confidently recommend it.
[18:39] <SpamapS> my experience with ActiveMQ was that it was dog slow, crashy, and frustrating to configure.
[18:39] <lynxman> SpamapS: deffo but it's a big delta with our package :)
[18:40] <lynxman> SpamapS: and it's not even packaged for Ubuntu
[18:40] <SpamapS> libactivemq-java - ActiveMQ message broker core Java libraries
[18:40] <SpamapS> not that?
[18:40] <lynxman> SpamapS: those are just the java libraries, not the server
[18:41] <SpamapS> Oh is there some reason the server isn't available?
[18:42] <SpamapS> I'm sure somebody might like to use it. :p
[18:42] <lynxman> SpamapS: well... I think I can take it from debian, would be a good exercise :)
[18:42] <lynxman> SpamapS: considering that packaging activemq is a b**tch
[18:42] <lynxman> SpamapS: so if the work has already been done... :)
[18:44]  * lynxman tries to have fun while at puppetconf
[18:46] <SpamapS> lynxman: what do you mean "take it" from debian? it should get synced automatically.
[18:46] <lynxman> SpamapS: it's in unstable
[18:47] <SpamapS> for how long?
[18:47] <SpamapS> Ahh its just arrived.
[18:47] <lynxman> SpamapS: http://packages.debian.org/sid/activemq
[18:47] <lynxman> SpamapS: yeah, it's new and fresh
[18:47] <SpamapS> lynxman: you should file a sync request
[18:47] <lynxman> SpamapS: oh, do explain :)
[18:47]  * lynxman learns something new
[18:48] <SpamapS> lynxman: 'requestsync' on any ubuntu system with ubuntu-dev-tools installed
[18:48] <lynxman> SpamapS: cool, ty
[18:48] <SpamapS> lynxman: its a bit late, but I think its worth putting it in 11.10
[18:49] <lynxman> SpamapS: maybe it can get in universe, but yeah I think it's late for 11.10
[18:50] <smoser> hallyn, so you were asking me to sponsor ifupdown for bug 850309, right ?
[18:50] <SpamapS> lynxman: everything starts in universe :)
[18:51] <CluelessPerson> I get an error while booting my server now....
[18:51] <CluelessPerson> init: ureadahead-other main process (712) terminated with status 4
[18:51] <hallyn> smoser: for starters just a sanity check.  But if it looks good to you then sure.
[18:51] <lynxman> SpamapS: as it should be ;)
[18:51] <hallyn> smoser: it'll need the teensiest of libvirt update as well :)
[18:51] <CluelessPerson> while doing it's check mount/disk thing after however many boots, I'm getting that and this other line and several others
[18:51] <SpamapS> CluelessPerson: do you have a separate /var partition ?
[18:52] <CluelessPerson> [   103.509461] [drm:pch_irq_handler] *ERROR* PCH poison interrupt
[18:52] <smoser> hallyn, i generally think its sane.
[18:52] <CluelessPerson> SpamapS, No, or, not that I know of, then again I'm a new.  JNot that I set up.
[18:52]  * mtaylor is currently battling libvirt. swings arms aimless in the air. collapses sobbing
[18:52] <SpamapS> CluelessPerson: ahh ok.. that used to cause a similar error. ureadahead is just an optimizer, so that message is probably just an annoyance.
[18:53] <smoser> and i even generally think it is non-intrusive
[18:53] <SpamapS> mtaylor: use fire
[18:53] <mtaylor> SpamapS: trying man
[18:55] <smoser> hallyn, i guess i dont have anything terribly against it.  i can't see how it would regress anything, its just an upstart event that nothing would ever listen for (until you hadded libvirt, where youd' ahve to depend on that version or greater of ifupdown)
[18:55] <SpamapS> smoser: +1 from me.. not having the 'wait forever for dhclient' part of static-network-up is disappointing, but it makes it so people will be setup for a better experience on upgrade.
[18:55] <hallyn> great, thanks guys.
[18:55] <hallyn> objections to doing the same for lucid?
[18:56] <SpamapS> I think it should go all the way back to lucid
[18:56]  * mtaylor thinks SpamapS should go all the way back to lucid
[18:56] <smoser> i have no objections.. but see what the patch looks like. if it actually is as trivial and non-intrusive as it is here.
[18:56] <SpamapS> There will be plenty of people who will appreciate that.
[18:56]  * SpamapS thinks mtaylor should stop being so hoary
[18:57] <mtaylor> SpamapS: but the breeze is so karmic...
[19:02] <smoser> hallyn, one comment, if you're merging changelog entries, you might as well merge them all i think
[19:03] <smoser> rather than confusing *\n[Scott Moser]\n...[Clint]\n...[Scott ]\n...
[19:03] <lynxman> SpamapS: btw you're packaging juju right? Let me know when I can sync my port :)
[19:07] <SpamapS> lynxman: right, we're waiting for all the eureka features to land then will upload it.
[19:07] <lynxman> SpamapS: schweet
[19:11] <CluelessPerson> SpamapS, It's keeping my server from booting
[19:14] <SpamapS> CluelessPerson: it may be a symptom of whats keeping your server from booting, but its probably not the root cause. What version of Ubuntu?
[19:21] <CluelessPerson> 11.04
[19:21] <CluelessPerson> SpamapS, ^
[19:23] <soren> Phew. Finally got a server off of its 2.6.24-19 kernel.
[19:24] <soren> Tough nut to crack, that one. It was infected with evms, which held my boot record hostage.
[19:25] <soren> /boot was an lvm volume that lived on top of a software raid1 set.
[19:26] <soren> Because evms had been installed at some point (this is a rather old server, let me tell you), it was in the initrd.
[19:26] <soren> ...this meant that instead of having md0 (the raid1 set holding my lvm stuff) be built up of sda1 and sdb1, it was built up of dm-0 and dm-1 which were in turn evms crackpot aliases for sda1 and sdb1.
[19:27] <soren> ...and lilo refused to install an mbr on a devmapper device.
[19:27] <soren> ...but I couldn't get rid of evms, because that would require installing an initrd that didn't have evms in it.
[19:28] <soren> I finally had the epiphany that I could use kexec to load a new kernel and initrd.
[19:29] <soren> and rid myself of this ancient kernel. I can't imagine how many security bugs have been fixed between 2.6.24-19 and 2.6.24-29.
[19:34] <j0nr> \leave
[19:43] <SpamapS> CluelessPerson: maybe try editting the command line in grub and remove the 'quiet' keyword so you can see what the system is doing before it fails.
[19:46] <Daviey> CluelessPerson: did you speak with kernel?
[19:53] <Dulcin> I'm getting this error when starting mysql: ERROR 2002 (HY000): Can't connect to local MySQL server through socket '/var/lib/mysql/mysql.sock' (2)
[19:53] <Dulcin> I've tried like the first 20 google results, but can't find a solution
[19:54] <Dulcin> I'm not sure what caused it, it used to work...
[19:56] <jeblair> SpamapS: do you have any pointers to docs on how to get started using orchestra with juju?  ie, how to have certain charms run on a (physical) server bootstrapped with orchestra?
[19:57] <SpamapS> jeblair: docs are being worked on right now.. its only just started working reliably in the last week or two. ;)
[19:58] <SpamapS> jeblair: Assigning services to specific machine types is one of those "oops we need to do a better job of that" problems. I believe the way right now is to change environments.yaml when deploying so that a certain mgmtclass of machines is chosen.
[19:59] <SpamapS> jeblair: its equally painful on EC2 .. so its a juju-wide problem. ;)
[20:00] <jeblair> SpamapS: is there a sample environments.yaml that uses orchestra?
[20:00] <jeblair> (or cobbler)
[20:02] <jeblair> ah, i think i see what i need in the source.  i'll try to follow that. :)
[20:07] <lynxman> SpamapS: juju-wide? :)
[20:15] <ScatterBrain> I just updated my laptop with the newest Kernel - came in as a software update.  When I rebooted, EVERYTHING is slower now.  Booting, Starting Gnome everything except for typing is slower.  Anyone know of a problem with the newest kernel?
[20:16] <GRMrGecko> What would you guys recommend using for Dell PowerEdge T710? Ubuntu Server or Debian?
[20:21] <lynxman> GRMrGecko: I think it'd work properly yeah, T710 has pretty standard components :)
[20:22] <RoyK> GRMrGecko: just try ubuntu 10.04 on it - if it works, well, you'll have a good server
[20:23] <RoyK> if lucid fails, try something newer
[20:23]  * RoyK prefers LTS for servers
[20:24] <lynxman> RoyK: definitely LTS for server good choice
[20:25]  * RoyK is managing about 40 ubuntu servers and doesn't want to upgrade them every 6 months
[20:25] <StevenR> upgrading is dull. it's much easier to do that every few years
[20:26] <RoyK> I still have a few 8.04 boxes running
[20:26] <StevenR> GRMrGecko: we have a load of R710 (I guess the rack version of the T710) running Lucid with no problems
[20:26] <RoyK> no problems with those
[20:26] <GRMrGecko> ok
[20:26] <GRMrGecko> thanks
[20:27] <RoyK> GRMrGecko: use cutting edge software when you need the latest and hottest. use LTS if you need something stable. use the former LTS release if you want something rock stable
[20:28] <RoyK> 8.04 is the most stable ubuntu release I've seen so far
[20:28] <RoyK> but then, it's old, and may lack a few drivers etc
[20:37] <RoAkSoAx> mtaylor: sorry for the delay got stuck in traffic, and phone run out of battery so no 3g
[20:37] <mtaylor> RoAkSoAx: that sounds terrible!
[20:38] <RoAkSoAx> mtaylor: anyway, as far as I know, there's been many people trying to configure many nics
[20:38] <RoAkSoAx> in a preseed
[20:38] <RoAkSoAx> and they have been unsucessful
[20:39] <mtaylor> RoAkSoAx: yes - well, I don't even care about configuring all of them, as much as just selecting the default one
[20:39] <mtaylor> but it, indeed, does not even come close to working
[20:40] <mtaylor> which would make orchestra/cobbler/pxe/preseed kind of useless in most server environments if it weren't for being able to pass netcfg stuff in to the kernel as boot options
[20:41] <kirkland> mtaylor: would gpxe help?
[20:41] <kirkland> mtaylor: i've heard its more configurable...though I've not used it myself
[20:41] <mtaylor> kirkland: I don't think it's a pxe problem as much as it's a preseed problem
[20:41] <mtaylor> everything _does_ work if you pass the extra kopt
[20:42] <mtaylor> it's just why preseed doesn't honor the value in the preseed file for default nic is beyond me
[20:42] <RoAkSoAx> mtaylor: there's mtaylor right, but how are you thinking to select the default network interfacE?
[20:42] <RoAkSoAx> mtaylor: we could do some tricks to get things passed "automatically"
[20:43] <mtaylor> RoAkSoAx: well, I'm fine with even coding the default interface into the preseed file in a normal cobbler setup (if that worked)
[20:43] <kirkland> RoAkSoAx: worst case, each machine's kopts can be edited in the web interface, right?
[20:43] <kirkland> RoAkSoAx: and on the cli, too, right?
[20:43] <mtaylor> yes. which is what I've been doing
[20:44] <kirkland> mtaylor: was that "yes" to me?  for the web ui or the cli?
[20:44] <mtaylor> well, cli - but it's the same overall effect - I have to configure extra kopts, which is very counter-intuitive
[20:44] <mtaylor> given that preseed HAS the option listed
[20:46] <mtaylor> in any case- I would think that a) fixing preseed would be stellar - but b) if preseed can't be fixed for some reason, it would be a nice user-friendly hack to have orchestra scan the preseed file it's using for that machine and if it finds the appropriate netcfg line to add it to kopts
[20:46] <mtaylor> (since orchestra seems to be doing a great job of providing sensible default setups for cobbler)
[20:48] <RoAkSoAx> kirkland: yes we can
[20:49] <RoAkSoAx> mtaylor: but for example, if we modify network values on a system, those network values do not reflect in the preseed as we arent actually editing the preseed for that
[20:49] <kirkland> RoAkSoAx: I think mtaylor's request is reasonable, though landing this in 11.10 might be hard, with final freeze next thursday
[20:49] <RoAkSoAx> or cobblre is not automatically editting the preseed to introduce network values
[20:50] <kirkland> mtaylor: would you mind filing a bug against orchestra so that we can document this as a bug/feature-request?
[20:50] <mtaylor> kirkland: sure! although I tihnk I should probably also file a bug against preseed... any idea where the right place to file that one is?
[20:50] <kirkland> RoAkSoAx: cobbler has support for some variables to get passed through to the preseeds;  smoser did some work around that
[20:51] <kirkland> mtaylor: file it against debian-installer
[20:51] <RoAkSoAx> kirkland: yeah we can pass variables and stuff
[20:51] <kirkland> mtaylor: you can click "also affects" orchestra
[20:51] <RoAkSoAx> kirkland: we can even obtain the info for eth0 ip, mac, etc etc and put it in the preseed
[20:51] <kirkland> mtaylor: so that it shows up in both lists
[20:51] <mtaylor> kirkland: sweet
[20:51] <RoAkSoAx> kirkland: however, in cobbler system we cannot say which interface is default
[20:51] <kirkland> mtaylor: cjwatson will move around the "debian-installer" bit to the right component
[20:51] <kirkland> RoAkSoAx: understood;  i think i told mtaylor as much
[20:51] <mtaylor> RoAkSoAx: would it perhaps be reasonable to file a bug against cobbler for a way to express default interface?
[20:52] <mtaylor> I mean - it's rare to find a machine _without_ 4 NICs these days
[20:52] <kirkland> RoAkSoAx: mtaylor: orchestra assumes that eth0 is the default interface;  machines where that's not true are handled on a one-off basis
[20:52] <RoAkSoAx> mtaylor: I guess we could but that would be a new feature
[20:52] <smoser> what is "default interface" ?
[20:52] <kirkland> RoAkSoAx: new features are okay for 12.04, just not 11.10 at this point
[20:53] <kirkland> smoser: i think mtaylor means "the netboot" interface, right?
[20:53] <mtaylor> kirkland: I'm fine with that assumption - the main thing I want to make sure is that booting the machine doesn't hang on a menu option
[20:53] <kirkland> mtaylor: +1, agreed
[20:53] <kirkland> mtaylor: so when you file your bug, express it in that manner
[20:54] <mtaylor> I shall attempt to do so
[20:54] <kirkland> mtaylor: "the bad thing that happens here is that headless, orchestra-installed machines are hanging on a menu option...etc."
[20:54] <mtaylor> yes. which is - you know - not usually what one is trying to accomplish
[20:55] <kirkland> mtaylor: :-)
[20:55] <smoser> why would a machine hang on the menu entry ?
[20:55] <smoser> the mnu is supposed to time out and fall back to locak disk.
[20:55] <smoser> no?
[20:55] <smoser> i'm clearly missing something. so i'll be quite.
[20:55] <smoser> quite quiet even.
[20:59] <RoAkSoAx> kirkland: yeah we could use the netboot interface as the default interface
[20:59] <RoAkSoAx> kirkland: an we could fill kopts or the preseed with that
[21:00] <smoser> there is an 'early_command' (similar to 'late_command')
[21:01] <mtaylor> smoser: so, outside of the orchestra context - if you pxeboot an image and hand it a preseed file, and in that preseed file you have netcfg/choose_interface=eth0
[21:01] <mtaylor> smoser: to tell the machine on what interface to listen for dhcp
[21:01] <mtaylor> smoser: it doesn't work
[21:01] <mtaylor> and it pops up a menu and asks you what interface you want as your default interface
[21:01] <mtaylor> if, however, you pass netcfg/choose_interface=eth0 to the kernel as a boot option - it works as expected
[21:02] <kirkland> mtaylor: okay, if that's the issue, that's a bug; and a big one
[21:02] <kirkland> mtaylor: i'm sorry if i missed that part directly, from the start
[21:02] <kirkland> mtaylor: that's a debian-installer issue
[21:02] <mtaylor> kirkland: yes. sorry - I may not have been clear earlier
[21:02] <kirkland> mtaylor: i probably just read right past that, and assumed it was an orchestra bug :-)
[21:03] <mtaylor> kirkland: totally - I was just thinking since it seems to be a known issue, perhaps orchestra could help work around it
[21:03] <kirkland> mtaylor: can you pastebin all or part of your preseed?
[21:03] <kirkland> mtaylor: any chance its just grammar/syntax?
[21:03] <kirkland> mtaylor: http://www.debian.org/releases/lenny/example-preseed.txt
[21:03] <kirkland> # To pick a particular interface instead:
[21:03] <kirkland> #d-i netcfg/choose_interface select eth1
[21:04] <RoAkSoAx> mtaylor: yeah that's totally a bug in the preseed
[21:04] <kirkland> RoAkSoAx: have you seen this before?  reproduced it?
[21:04] <mtaylor> kirkland: d-i netcfg/choose_interface select eth0
[21:04] <RoAkSoAx> kirkland: i have never personally tried but adam_g encountered the problem and used kopts to chose default interface
[21:05] <RoAkSoAx> kirkland: so yeah we encountered the same issue but was work arounded the same
[21:05] <mtaylor> kirkland: I ran in to this 2 months ago in a data center - and then I found a blog post about using kopts
[21:05] <kirkland> RoAkSoAx: hmm, okay, we need a pretty urgent bug filed and assigned to cjwatson, then
[21:05] <RoAkSoAx> kirkland: i'll test it again and see what I find though
[21:05] <mtaylor> speak of the devil
[21:05] <RoAkSoAx> kirkland: the thing is withing cobbler we could actually just add an option to select the default interface to place it in the preseed
[21:06] <RoAkSoAx> kirkland: i'll look into that though
[21:06] <RoAkSoAx> kirkland: will investigate a bit more how was that done in rhel
[21:06] <mtaylor> of course, in _my_ personal usecase, I'd also need it backported to d-i of lucid, maverick and natty ... so I might just need to keep doing kopts for those. but still
[21:07] <kirkland> hmm, the closest bugs I see are:
[21:07] <kirkland> https://bugs.launchpad.net/ubuntu/+source/netcfg/+bug/713385
[21:07] <kirkland> and https://bugs.launchpad.net/ubuntu/+source/netcfg/+bug/56679
[21:07] <smoser> mtaylor, thank youfor explanation.
[21:08] <mtaylor> smoser: sure! thanks for asking the question - it seems to have caused me to re-explain in a more useful way :)
[21:09] <mtaylor> kirkland: https://bugs.launchpad.net/ubuntu/+source/debian-installer/+bug/855921
[21:10] <kirkland> mtaylor: this is Oneiric?
[21:10] <kirkland> mtaylor: one thing you might want to edit in your bug description ....
 kirkland: d-i netcfg/choose_interface select eth0
[21:11] <mtaylor> kirkland: tbh, I haven't tried oneiric - it wasn't booting last time I was setting up cobbler
[21:11] <kirkland> mtaylor: whereas you say: "in that preseed file you have netcfg/choose_interface=eth0"
[21:11] <mtaylor> kirkland: good point
[21:11] <kirkland> mtaylor: okay, so this is natty?
[21:11] <kirkland> mtaylor: or something else?
[21:11] <kirkland> mtaylor: i'd like to note which version you're seeing this one
[21:12] <mtaylor> kirkland: it was either natty or maverick
[21:12] <mtaylor> kirkland: I _believe_ both
[21:12] <kirkland> mtaylor: k
[21:12] <mtaylor> kirkland: I'll work on re-reproducing it
[21:12]  * mtaylor just got 5 machines in a lab for cobbler/install use
[21:12] <kirkland> i wonder if I can reproduce this in a vm with multiple nics ....
[21:13]  * kirkland tries
[21:17]  * RoAkSoAx waaaaaaaaaaaaaaaaaa compiz acting up grrr
[21:18] <kirkland> RoAkSoAx: unity2d, baby
[21:18] <RoAkSoAx> kirkland: hehe yeah I'm gonna have to change
[21:18] <kirkland> RoAkSoAx: for bonus, you'll get 3 extra hours of battery, too!
[21:19] <RoAkSoAx> kirkland: really? heh good to know
[21:19] <RoAkSoAx> brbr
[21:19] <RoAkSoAx> lol
[21:19] <RoAkSoAx> exit
[21:20] <kirkland> RoAkSoAx: um
[21:20] <kirkland> RoAkSoAx: how does cobbler feed the preseed to the machine being installed?
[21:20] <kirkland> RoAkSoAx: over http?  or pxe?
[21:21] <RoAkSoAx> kirkland: http
[21:21] <kirkland> mtaylor: hmm
[21:21] <kirkland> mtaylor: we need the network to be "up", to retrieve the preseed
[21:22] <kirkland> hello chicken, my name is egg!
[21:22] <segv> so, i installed mongrel-clustesr and it has a sites-enabled/sites-available dir, but no examples or anything really, is there a stock page or package documentation for the repo version so I don't have to modify anything
[21:22] <RoAkSoAx> kirkland: basically, cooobler just creates a pxe file that lists auto url==http://blablablala/bla/preseed/
[21:22] <mtaylor> I thought it did the preseed as a bit of the dhcp netboot interaction
[21:22] <RoAkSoAx> or something like that
[21:22] <segv> cobbler is awesome
[21:22] <segv> :)
[21:22] <mtaylor> yes - no, you're right
[21:22] <kirkland> mtaylor: i'm confirming with RoAkSoAx, but it looks like it's using the url= param on the kernel line
[21:23] <mtaylor> but this is the funny part - because of the netboot part - the network is actually WORKING just fine
[21:23] <RoAkSoAx> kirkland: yes it does
[21:23] <mtaylor> which is what makes the netcfg/choose_interface=default thing not working all the more funny
[21:23] <kirkland> mtaylor: okay, i'm disecting this ....
[21:23] <segv> you know you can just grab the pxe grub init etc. and use grub to netboot yeah?
[21:23] <segv> I do it all the time :)
[21:23] <kirkland> mtaylor: basically, i think the assumption should be in debian-installer
[21:24] <kirkland> mtaylor: that if you netbooted using some interface, then that's your damn default!
[21:24] <RoAkSoAx> kirkland: http://paste.ubuntu.com/694771/
[21:24] <mtaylor> kirkland: YES
[21:26] <kirkland> mtaylor: hmm, any idea, top your head, how to determine that inside of the OS?
[21:27] <mtaylor> uhm
[21:28] <mtaylor> netstat ?
[21:28] <mtaylor> kirkland: ip route | grep default | awk '{print $5}'
[21:29] <kirkland> mtaylor: heh, we have to do this *before* ip has a route :-)
[21:29] <kirkland> mtaylor: we're trying to determine which interface to use, right?
[21:30] <kirkland> mtaylor: fwiw, i've reproduced this in a VM with multiple nics
[21:31] <kirkland> RoAkSoAx: so...here's what i'm thinking
[21:31] <RoAkSoAx> kirkland: shoot
[21:31] <kirkland> RoAkSoAx: i _think_ we should perhaps set that netcfg/choose_interface=eth0 in the kernel opts, by default
[21:31] <mtaylor> kirkland: YAY
[21:31] <kirkland> RoAkSoAx: and document that that needs to be changed for machines where that's a bad assumption
[21:31] <kirkland> mtaylor: ^
[21:31] <kirkland> mtaylor: what do you think?
[21:32] <mtaylor> kirkland: I think that's an excellent default for orchestra to do
[21:32] <mtaylor> kirkland: it catches the 90% case and is inline with your assumtions already
[21:32] <kirkland> mtaylor: i thought you might like that :-)
[21:32] <mtaylor> kirkland: you've figured me out
[21:32] <kirkland> mtaylor: yeah, i think so
[21:32] <mtaylor> sensible defaults ftw
[21:35] <RoAkSoAx> kirkland: we can do that
[21:39] <RoAkSoAx> kirkland: I'll look into doing it on orchestra or simply just changing a template whithin cobbler
[21:42] <segv> hmmm
[21:43] <segv> What's a stock mongrel-cluster enabled site config look like, just a default mongrel cluster file?
[21:57] <jdstrand> Daviey: I'm looking at the iscsitarget mir (bug #843808)
[21:58] <jdstrand> Daviey: why did tgt get demoted and we are asking for a mir of iscsitarget?
[22:01] <jdstrand> zul: ^
[22:03] <jdstrand> well, I guess because nova doesn't have tgt support
[22:04] <jdstrand> zul: I found this actually: https://bugs.launchpad.net/nova/+bug/819997
[22:07] <Daviey> jdstrand: tgt was considered to be less reliable than iscsitarget
[22:07] <Daviey> IIRC
[22:07] <Daviey> ah no
[22:07] <kirkland> jamespage: don't you have an enjumble charmula for jenkins?
[22:08] <mtaylor> jamespage: hey
[22:08] <Daviey> jdstrand: i might be talking tosh
[22:08] <jdstrand> Daviey: I'm trying to find the origins of tgt, but I thought it was a rewrite and supported by RHEL
[22:08] <jdstrand> well, I know it is supported by RHEL
[22:08] <Daviey> jdstrand: https://bugs.launchpad.net/nova/+bug/819997
[22:08] <jdstrand> I don't know its origins
[22:08] <jdstrand> Daviey: yes, I just pasted that bug :)
[22:08] <mtaylor> jamespage: I was just thinking that, in addition to me ever finishing the jclouds plugin - we should write a jenkins plugin that just makes juju calls
[22:09] <Daviey> https://code.launchpad.net/~markmc/nova/iscsi-tgtadm-choice/+merge/75906
[22:09] <Daviey> good description
[22:10] <kirkland> mtaylor: https://code.launchpad.net/~charmers/charm/oneiric/jenkins/trunk
[22:11] <jdstrand> Daviey: ok, I updated the bug asking for server team imput regarding tgt
[22:11] <jdstrand> input
[22:14] <mtaylor> kirkland: so, uh, can't use orchestra to deploy anything less than oeneric, apparently, because it wants to install ubuntu-orchestra-client on the machine. ?
[22:15] <kirkland> mtaylor: good point -- RoAkSoAx we need to fix that
[22:15] <kirkland> mtaylor: you'll need to drop that out of the preseed file
[22:15] <mtaylor> kirkland: ok. it'll still work and stuff?
[22:15] <kirkland> mtaylor: we'll ge that fixed
[22:15] <kirkland> mtaylor: yeah
[22:15] <mtaylor> cool.
[22:15] <RoAkSoAx> kirkland: uhmmm i can do something else to work around that problem
[22:16] <kirkland> RoAkSoAx: cool -- what?
[22:16] <mtaylor> kirkland: need a bug filed?
[22:16] <kirkland> mtaylor: please
[22:16] <kirkland> mtaylor: assign it to RoAkSoAx (andreserl)
[22:16] <RoAkSoAx> kirkland: put that in a preseed, and do and if/else to see the version of the OS
[22:16] <RoAkSoAx> s/preseed/snippet
[22:16] <RoAkSoAx> so if it is natty or lower, not install the client
[22:17] <kirkland> mtaylor: we definitely want orchestra to be able to install lucid - oneiric
[22:17] <kirkland> RoAkSoAx: rock!
[22:18] <mtaylor> https://bugs.launchpad.net/orchestra/+bug/855956
[22:20] <RoAkSoAx> mtaylor: thanks
[22:21] <kirkland> mtaylor: cool
[22:21] <kirkland> mtaylor: you coming to Orlando?
[22:21] <mtaylor> kirkland: yup.
[22:21] <kirkland> mtaylor: cool!  I'll make RoAkSoAx buy you a beer :-D
[22:21] <kirkland> LoL
[22:21] <mtaylor> kirkland: whole openstack community team is coming, actually - me, and jeblair  and jaypipes and ttx and reed
[22:21] <kirkland> neat
[22:22] <kirkland> I'll be at ODS too
[22:22] <mtaylor> yup. it's gonna be baller
[22:22] <mtaylor> sweet
[22:23] <RoAkSoAx> coolio
[22:23] <RoAkSoAx> cool*
[22:23] <RoAkSoAx> lol
[22:23] <Daviey> thanks jdstrand
[22:29] <Daviey> jdstrand: Do you have any findings from glance?
[22:34] <jdstrand> Daviey: haven't looked yet. rest assured, I will get there. other than 3 phone calls today, all I have been doing is server team MIRs (and that doesn't count all the hours on monday and tuesday)
[22:35] <jdstrand> (please bear with me, code audits, even at this shallow depth take a while)
[22:40] <Daviey> jdstrand: Yes, sorry - i didn't mean to seem to be keeping on, I was really just seeing if there is anything we can do based on initial findings.
[22:41] <Daviey> We do appreciate your time :)
[22:42] <jdstrand> no worries
[22:49] <zul> jdstrand: yeah thats a rewrite of what i did (the nova branch)
[23:03] <jeblair> so another workaround to bug 855956 is to install oneiric... it prompted with an option to configure iscsi volumes, which isn't desirable in an unattended orchestra installation...
[23:09] <jeblair> kirkland: does that sound like an installer bug, or something we should be setting in orchestra?
[23:41] <bfreis> Is there a Natty cloud image very very small, smaller than those available at http://cloud-images.ubuntu.com/server/natty/current/ ? I'd like it to have only cloud init, and a very small set of packages
[23:42] <bfreis> Or how could I build such an image?