[00:00] <Daviey> adam_g: it does,             i['dnsName'] = i['publicDnsName'] or i['privateDnsName']
[00:00] <Daviey> and public is null, therefore using private.. so it's an issue higher in the stack
[00:01] <adam_g> Daviey: in terms of that bug:  http://paste.ubuntu.com/845172/
[00:02] <adam_g> Daviey: in api/ec2/ec2utils.py   avoiding that first attempt calls to the network driver for the actual info, and returns the correct floating_ip
[00:02] <Daviey> adam_g: that function is failing?
[00:02] <adam_g> $ echo $(curl -s http://169.254.169.254/2008-02-01/meta-data/public-ipv4)
[00:02] <adam_g> 192.168.20.225
[00:02] <adam_g> its not failing
[00:03] <Daviey> adam_g: are you working on a fix for that?
[00:03] <adam_g> according to the cache, the floating ip is not associated with the instance, its associated some time after the cache is populated
[00:03] <adam_g> it gets an outdated cache hit
[00:03] <Daviey> adam_g: i wonder if, my case, it's because i keep restartng the api service
[00:04] <adam_g> Daviey: just trying to triage a bit
[00:05] <adam_g> Daviey: what do you mean? whats restarting the api server do?
[00:06] <Daviey> adam_g: where is the cache stored, ad flushed?
[00:06] <Daviey> and*
[00:08] <adam_g> Daviey: its stored in the database, see InstanceInfoCache(BASE, NovaBase) in db/sqlalchemy/models.py
[00:09] <adam_g> Daviey: AFAICS, it only stores network info. im not sure if or when the proper place to update it is
[00:15] <Daviey> adam_g: ah
[00:19] <Daviey> adam_g: confirmed, that if i force a fallback - it returns the correct value
[00:19] <Daviey> adam_g: therefore, i assume it used to faill, but now returns Null.
[00:20] <Daviey> I'm guessing after the first attempt, it's cached.
[00:22] <Daviey> adam_g: http://pb.daviey.com/bJ2Y/
[00:24] <adam_g> Daviey: the ip info returned form those get_* functions is more than just the IP you want, it contains both fixed and floating. so an empty floating_ip list just means there are no associated floating_ips, which there aren't when the instance is created / cache initially populated
[00:29] <Daviey> ah
[00:33] <Daviey> adam_g: http://pb.daviey.com/qcQx/
[00:34] <Daviey> adam_g: so the odd thing is, it IS cached after a successful run
[00:35] <adam_g> Daviey: weird. a successful run, meaning, a fallback to the network_api?
[00:36] <Daviey> adam_g: that is what i think
[00:36] <Daviey> adam_g: new instance, without a floating_ip and the original one with http://pb.daviey.com/zO5j/
[00:39] <Daviey> adam_g: so, run-instance produced a FAIL (expected, as it doesn't have an address at reserve time)
[00:39] <adam_g> Daviey: ah, hmm. ill tryo with a fresh instance
[00:46] <Daviey> adam_g: want my log from a fresh instance?
[00:46] <adam_g> Daviey: nah, i think we're seein the same thing now
[00:46] <Daviey> The first one shows a FAIL, then SUCCESS
[00:46] <Daviey> http://pb.daviey.com/WfEA/
[00:48] <uvirtbot`> New bug: #933818 in nova "public-ipv4 metadata url not working" [Undecided,Confirmed] https://launchpad.net/bugs/933818
[00:48] <adam_g> Daviey: right, the bug is that it will never fail and refresh its cache, just because its not up to date with floating ip info
[00:49] <Daviey> right
[00:49] <adam_g> Daviey: so, with your version, what happens if you deallocate the address and allocate a different ip ? :)
[00:55] <Daviey> adam_g: trying
[00:55] <Daviey> sorry, went afk
[00:56] <adam_g> Daviey: i think the whole meta-data handler needs a giant refactoring, it would solve this and (all of the) other problems
[00:57] <roaksoax> Daviey: still around?
[00:59] <Daviey> adam_g: http://pb.daviey.com/Tb4B/ .. see the logs inline
[00:59] <Daviey> roaksoax: yah
[01:00] <Daviey> adam_g: i'm not convinced it is floating_ip related
[01:00] <Daviey> i'd expect to see a FAIL there...
[01:02] <Daviey> adam_g: although... floating_ip is constantly null
[01:03] <Daviey> adam_g: ah, interesting... the cache is present, but not 'correct'
[01:04] <adam_g> Daviey: right, its not updated with floating_ips
[01:05] <zul> adam_g: i just pused a new python-keystoneclient to the ubuntu-server-dev
[01:06] <roaksoax> Daviey: pm
[01:07] <zul> adam_g: argparse has been removed from the egg as well
[01:07] <adam_g> zul: nice, thanks. hopefully ill finally be able ot try that soon
[01:07] <zul> adam_g: ill upload it to the archive tomorrow
[01:07] <adam_g> zul: just the client, right?
[01:07] <zul> right
[01:07] <adam_g> q!
[01:07] <adam_g> q
[01:26] <uvirtbot`> New bug: #927803 in postfix (main) "Fails to launch recovery menu when (at least) Postfix installed" [High,Triaged] https://launchpad.net/bugs/927803
[01:30] <adam_g> zul: what PPA are the new keystone packages going to?
[01:37] <zul> adam_g: it will be going in https://launchpad.net/~openstack-ubuntu-testing/+archive/openstack-ksl-testing starting tomorrow
[01:38] <adam_g> zul: cool
[01:38] <adam_g> zul: what time you doing the opentack snapshot?
[01:38] <zul> adam_g: tomorrow morning
[01:39] <adam_g> zul: https://review.openstack.org/#change,4149 <- i screwed pep8 with one freakin space on this, but hopefully it can get merged before
[01:39] <zul> okies
[03:07] <BuenGenio> is the PHP 5.3.10 update coming to 11.04/10 any time soon?
[04:30] <twb> ARGH.  Grub you're causing problems *AGAIN*
[04:31] <twb> I have this as a BCP: http://paste.debian.net/156590/
[04:31] <twb> But guess what, it looks like triggers (including the kernel post-hook to rewrite grub.cfg) happen *AFTER* the dpkg post-invoke
[06:05] <tewea> i have an external hard disck but i coudn't write to it  just iam using FAT32 fomat please help me?
[06:11] <twb> Scratch that, it looks like apt.conf isn't reread between aptitude sub-runs.  Mea culpa.
[06:18] <tewea> i have an external hard disck but i coudn't write to it  just iam using FAT32 fomat please help me?
[06:26] <BuenGenio> is the PHP 5.3.10 update coming to 11.04/10 any time soon?
[06:29] <twb> SpamapS might know, if he's awake.
[06:33] <journeeman> Hi. I am running 11.10 amd64-server. I had installed xubuntu on the machine so, xserver got installed automatically. I need xserver to run as a daemon after startup so I can forward graphics over ssh but, I also need to have the default runlevel as Multiuser text, i.e. `3' in the old unix way.
[06:33] <journeeman> How do I do this? Aprreciate any help :)
[06:36] <twb> journeeman: we discourage use of GUIs on servers
[06:36] <twb> runlevels 2 through 5 are identical on Ubuntu systems; "3 is text; 5 is GUI" is a RHism
[06:37] <twb> To disable GUI on startup, find /etc/init/gdm.conf and comment out the start-on line, where "gdm.conf" may be xdm.conf or kdm.conf or similar.
[06:37] <twb> ssh -X requires only the xauth program, not all of X.  xauth is in its own package as at 11.10.
[06:39] <twb> If this is a fresh install, recommend you reinstall from scratch, and only install xauth and the specific app you want (e.g. firefox), but *not* the xubuntu-desktop or xorg or gdm packages.
[06:40] <journeeman> twb: Okay. I uninstalled xubuntu later. Didn't uninstall lightdm or other X packages. This is not a fresh install. Can't afford to re-install. That's definitely the last option.
[06:42] <twb> if your server is just hosting the X apps, which X confusing calls the "client side", you only need the X app, xauth and openssh-server.  You *don't* need the X "server" side, i.e. lightdm and xserver-xorg
[06:42] <journeeman> Okay, cool.
[06:44] <journeeman> Okay, I commented out the `start on' line in lightdm.conf. So that should not start GUI on startup, right? Will it start X though? How do I stop xserver from starting?
[06:45] <koolhead11> hi all
[06:46] <journeeman> twb: Thank you very much for the help :)
[06:50] <koolhead11> SpamapS: do i have to add a diff repo to get the latest 5.3.10 pkg?
[06:57] <koolhead11> got it, apt-pinning :)
[06:59] <twb> koolhead11: the x server doesn't 'start' at all
[06:59] <twb> Sorry, bad completion
[06:59] <koolhead11> twb: np :)
[06:59] <twb> journeeman: lightdm starts an X server, but for ssh -X the X server is running on the place you're sshing *from*
[07:00] <journeeman> oh okay..
[07:09] <journeeman> twb: Okay, got it. Removed xserver-xorg-* and allied packages as well as lightdm. Everything works great. Thanks. :)
[07:36] <uvirtbot`> New bug: #934011 in postfix (main) "package postfix 2.7.0-1ubuntu0.2 failed to install/upgrade: subprocess installed post-installation script returned error exit status 75" [Undecided,New] https://launchpad.net/bugs/934011
[09:06] <jamespage> morning all
[09:49] <jamespage> Daviey: around?
[09:58] <jamespage> rbasak, hows openmpi looking?
[09:59] <rbasak> good!
[09:59] <rbasak> I needed one fix, a DSO-related thing
[09:59] <rbasak> Apart from that the only issue was build-deps and ordering as I haven't quite got a streamlined transition testing system
[10:00] <rbasak> Also I have news of a test suite, though I haven' tlooked at that yet
[10:05] <jamespage> rbasak, so do you think we will propose a full transition? or just upload a 1.5 versioned package to universe?
[10:07] <rbasak> jamespage: assuming that a full rebuild test succeeds, and I can operate the MPI test suite successfully, I see no reason why we shouldn't do a full transition - apart from the timing, but I don't think anyone would have done any more testing had we uploaded a transition earlier this cycle anyway.
[10:08] <Fudge> hiq uick question, is the console available from live cd ubuntu-server
[10:08] <rbasak> jamespage: but it may not be absolutely necessary - it just seems cleaner to me. What do you think?
[10:08] <jamespage> rbasak, I agreed that its cleaner
[10:08] <jamespage> we would probably push for the transition early next release anyway.
[10:09] <rbasak> My doubts were about testing, and now that I understand that there is an MPI test suite, it doesn't seem to be so much of an issue.
[10:09] <jamespage> rbasak, there is a transition tracker we can get this sort of stuff setup in which might help a bit
[10:09] <rbasak> Oh, really? That would be helpful
[10:10] <jamespage> rbasak, http://people.canonical.com/~ubuntu-archive/transitions/
[10:10] <koolhead11> https://launchpad.net/bugs/934064
[10:10] <uvirtbot`> Launchpad bug 934064 in horizon "Installing openstack-dashboard on Precise removes Keystone package" [Undecided,New]
[10:10] <koolhead11> is it?
[10:11] <rbasak> Just having a complete list of rdepends would be good. I'd like a recursive option to reverse-depends. I was going to write that today :)
[10:12]  * koolhead11 is trying it out
[10:12] <uksysadmin> koolhead11: I've just raised that bug
[10:13] <koolhead11> uksysadmin: am testing it. i doubt it could be the case
[10:13] <koolhead11> still
[10:15] <koolhead11> horizon uses keystone :)
[10:17] <koolhead11> uksysadmin: what is the pkgname 4 dashboard openstack-dashboard or openstack-horizon
[10:27] <jamespage> rbasak, http://paste.ubuntu.com/845579/
[10:27] <jamespage> I grabbed the codebase and tried it for openmpi
[10:28] <jamespage> that at least gives us the transition order
[10:29] <rbasak> jamespage: awesome, thanks! I found xdeb which can provide a build sequence, but that doesn't help with parallelism without using xdeb itself which doesn't use sbuild.
[10:29] <jamespage> rbasak, once you are happy that we can transition and we have a FFE exception I suggest we ask to get this setup in the transition tracker
[10:30] <rbasak> jamespage: I have a patch for blacs-mpi btw, that causes its rdepends to FTBFS
[10:31] <rbasak> jamespage: I'd like to do a full rebuild test which should work better this time with the dependency information. I'll set it up today with parallel builds to speed it up.
[10:31] <rbasak> jamespage: and then I'll get started on the test suite.
[10:31] <jamespage> rbasak, are you installing the resulting artifacts?
[10:32] <rbasak> jamespage: no, although I noticed that sbuild has a piuparts thing that can do that. Do you want me to enable it?
[10:32] <rbasak> jamespage: I take it you're asking me to check if the generated debs are installable?
[10:32] <jamespage> rbasak: that might be helpful with the FFE
[10:32] <rbasak> OK
[10:33] <jamespage> rbasak, no I'm saying that for later tiers of dependencies you will need to rebuild against the packages you just built
[10:33] <jamespage> make sense?
[10:33] <rbasak> Ah yes - I have been doing that
[10:33] <rbasak> I modified my sbuild to use add a local repo, and have that pinned at higher priority
[10:33] <rbasak> That's how I discovered the blacs-mpi problem.
[10:34] <jamespage> right - I see
[10:34] <jamespage> FYI the level 0 and 1 dependencies can be folder
[10:34] <jamespage> mpi-defaults does not need to be rebuilt
[10:34] <jamespage> *folded
[10:37] <uksysadmin> koolhead11: openstack-dashboard
[10:37] <uksysadmin> couldn't find one referencing horizon... lemme check again.
[10:38] <uksysadmin> nope - openstack-dashboard it is, references a python-django-horizon
[10:38] <Stef2> when i start ubu server some errors appear but they dissapear fast. how can i see all the text that appear on screen util ubu login?
[10:39] <Stef2> *screen until
[10:43] <Daviey> jamespage: always
[10:43] <jamespage> Daviey, good morning!
[10:43] <jamespage> late one last night?
[10:44] <Daviey> jamespage: sadly, all too often
[10:50] <koolhead11> uksysadmin: i can see keystone still there
[10:50] <koolhead11> i dont think its a bug
[10:50] <koolhead11> :)
[10:51] <koolhead11> Daviey: am bit famous today!! :P
[10:54] <jamespage> Daviey: I wanted your opinion on something in a new package I've reviewing
[10:55] <jamespage> bug 930422
[10:55] <uvirtbot`> Launchpad bug 930422 in ubuntu "[needs-packaging] rds Resara Server" [Wishlist,New] https://launchpad.net/bugs/930422
[10:55] <jamespage> all-in-all its looking OK; however I'm uncomfortable about the embedded source of qtrpc2 in this package
[10:56] <jamespage> Daviey, wanted to double check I'm not being irrational!
[10:57] <Daviey> jamespage: you seem to be being quite thorough :)
[10:58] <jamespage> Daviey: I learn't from the master of thorough!
[10:58] <uksysadmin> gets removed on my setup :(
[10:58] <Daviey> jamespage: Hah.. right, so - if it were me, i think i'd be tempted to, at this stage, just make the changes myself and upload
[10:59] <jamespage> Daviey: with a FFE exception of course :-)
[11:00] <Daviey> jamespage: it seems they went to the effort.
[11:00] <Daviey> https://launchpad.net/~resaraserver/+archive/resaraserver/+build/2046346
[11:01] <jamespage> Daviey: it looks so
[11:01] <jamespage> hence my slight discomfort with it being embedded
[11:02] <Daviey> jamespage: i'd get clarification on that, but the rest of the things i'd just change muyself.
[11:03] <jamespage> Daviey: OK - I'll catch brendan this afternoon (he's east coast)
[11:04] <jamespage> Daviey: can i get an FFE for this on #ubuntu-release or do I need to justify in the bug and sub ubuntu-release?
[11:05] <Daviey> jamespage: Add ratonal on the bug, subscribe ~ubuntu-release.. and give me a ping. :)
[11:05] <koolhead11> uksysadmin: gosh you are right :)
[11:05] <jamespage> Daviey: OK - I'll get the rationale for the embedded qtrpc2 library sorted out and do so later today
[11:05] <Daviey> jamespage: add [FFe] to the title
[11:06] <Daviey> jamespage: sorry, rational for a FFe
[11:06] <jamespage> Daviey: yeah - got that!
[11:06] <koolhead11> uksysadmin: i installed keystone after installing dashboard and its telling me it will remove whole horizon. gosh
[11:06] <Daviey> jamespage: groovy!
[11:07] <uksysadmin> koolhead11: yup. bug :p
[11:07] <koolhead11> uksysadmin: http://paste.ubuntu.com/845620/
[11:08] <uksysadmin> confirmed then.
[11:08] <koolhead11> yeah am adding my pastebin output there too in comment
[11:08] <uksysadmin> great
[11:10] <koolhead11> uksysadmin: am sure it would be fixed and in the pinned repo.
[11:11] <uksysadmin> pinned repo?
[11:12] <koolhead11> uksysadmin: https://help.ubuntu.com/community/PinningHowto
[11:12] <uksysadmin> cheers
[11:13] <uksysadmin> not sure I have a version I know works to pin though - does this suggest I install keystone, pin that and it won't get removed?
[11:13] <uksysadmin> sounds like a fudge when it comes to dependency checking
[11:14] <koolhead11> uksysadmin: i think this will keep you updated with the daily developments i suppose
[11:15] <uksysadmin> true - will see how it goes.  I appreciate that essex on precise is a moving target.
[11:16] <uksysadmin> keystone is getting a hammering in this release - dialogs boxes popping up asking user input, to getting removed by unrelated packages! ;-)
[11:16] <koolhead11> essex/precise/precious == win!! :)
[11:16] <uksysadmin> I for one is looking forward to the Randy Robin release in 2013 ;-)
[11:19] <Daviey> rvba: Hey, do you have a moment?
[11:22] <rvba> Daviey: sre, what's up?
[11:22] <rvba> sure*
[11:23] <Daviey> rvba: So, i wanted to run a patchset against you for django.
[11:23] <rvba> Daviey: all right…
[11:24] <Daviey> bug 905837 , has patches attached.. but differ from upstream.
[11:24] <uvirtbot`> Launchpad bug 905837 in python-django "python-django 1.3-2ubuntu1.1 is incompatible with Psycopg2 2.4.2-1" [Undecided,Confirmed] https://launchpad.net/bugs/905837
[11:25] <Daviey> rvba: specifically, from django.db.backends.postgresql.introspection import DatabaseIntrospection as PostgresDatabaseIntrospection , isn't removed.
[11:26] <uvirtbot`> New bug: #934118 in nova (main) "Floating ips not available to instances after upgrade" [Undecided,New] https://launchpad.net/bugs/934118
[11:26] <rvba> Daviey: I have some reading to do now :)
[11:27] <Daviey> rvba: sorry!
[11:27] <rvba> Daviey: That's ok ;).  I'll look into the patches.
[11:28] <Daviey> rvba: thanks
[11:28] <rvba> Daviey: I mean you just gave me some reading to do ;)
[11:28] <rvba> Daviey: so you want me to make sure the patch is fine is that it?
[11:29] <rvba> Daviey: The one proposed on the ticket.  Because the fact that it differs from the upstream patch worries you a bit.… right?
[11:29] <rvba> Proposed on the lp bug that is…
[11:30] <rvba> Daviey: just making sure because I'm kinda new to packaging (to say the least).
[12:05] <Daviey> rvba: right!
[12:05] <rbasak> jamespage: could you please review/sponsor bug 934138?
[12:05] <uvirtbot`> Launchpad bug 934138 in blacs-mpi "libblacs-openmpi.so.1.1 does not declare its dependencies" [Undecided,New] https://launchpad.net/bugs/934138
[12:05] <rvba> Daviey: cool, I'm on it…
[12:05] <Daviey> rvba: It's not really packaging problem, just that the proposed diff's are different to trunk
[12:06] <rbasak> jamespage: (this is affecting rebuild testing but the bug is relevant to the current archive version too)
[12:06] <Daviey> rvba: thanks!
[12:06] <rvba> Daviey: right.
[12:09] <jamespage> rbasak, why did you need to switch gcc -> ld?
[12:10] <rbasak> jamespage: it did confuse me. gcc in that form isn't creating the desired result. To simplify things, I used ld, and it worked. I'm still baffled as to why it doesn't work using gcc, but all that line does is link object files, so ld would be more suitable anyway.
[12:11] <jamespage> rbasak, ack
[12:11]  * jamespage has discovered if my laptop fan runs harder it stops making a grinding noise!
[12:12] <rbasak> jamespage: apt-get install cpuburn? :-)
[12:12] <jamespage> rbasak, its still under warranty - raised a support call last night
[12:13] <rbasak> I had an RMA for my laptop but failed to get round to backing it up, switching to my spare and sending it off in the 7-day window they gave me :-(
[12:13] <Daviey> jamespage: 'grinding'.. crikey
[12:13] <jamespage> rbasak, next business day onsite warranty is worth it...
[12:14] <rbasak> hmm
[12:16] <rvba> Daviey: the second patch on lp is the same as the fix 16520 on django.  The first patch is the minimal effort that needed to done for the second patch to apply successfully (it simply copies db/backends/postgresql/creation.py to db/backend/postgresql_psycopg2).  It's a very short "summary" of what happened between 1.3 and trunk.
[12:16] <rvba> In django upstream, between 1.3 and trunk, the whole module db/backends/postgresql has been removed.
[12:16] <rvba>  "from django.db.backends.postgresql.introspection import DatabaseIntrospection as PostgresDatabaseIntrospection" is still there because all this renaming/deleting has not happened in 1.3 and the first patch only took care of the file creation.py where the meat of the real fix (the second patch on lp or changeset 16250 in django) happens.
[12:17] <rvba> Daviey: I haven't run the tests or anything yet but the fix seems very sane to me.
[12:17] <rvba> Daviey: do you want me to run the tests & all or have you done that already?
[12:18] <Daviey> rvba: ahh, so if i fold both patches into one.. you think it makes sense?
[12:18] <Daviey> rvba: no, i'll run the tests.. don't worry about that
[12:18] <Daviey> it was more the logic that didn't make sense to me :)
[12:19] <rvba> Daviey: absolutely, like I said, the first is a summary of what happened to the Django code base between 1.3 and trunk (the minimal amount for the second patch to be applied).
[12:19] <rvba> the first patch*
[12:22] <Daviey> rvba: thanks :)
[12:22] <jamespage> rbasak, you can still use gcc - but the -l's need to be moved to after the .o files in the command
[12:23] <rvba> Daviey: Glad to help.
[12:23] <jamespage> rbasak, http://paste.ubuntu.com/845710/
[12:23] <rbasak> jamespage: aha, thanks! I thought it was an ordering issue from memory but tried it the wrong way round :)
[12:23] <jamespage> OK if I make that minor change and upload?
[12:24] <jamespage> I hit a similar issue with some hadoop stuff a few weeks ago
[12:24] <rbasak> Sure, thanks! Sorry about forgetting update-maintainer too.
[12:24] <jamespage> rbasak, np - sponsor-patch did that for me :-)
[12:24] <rbasak> btw, are you sure that this will work? I haven't tested it now :)
[12:26] <rbasak> Just testing it now.
[12:28] <jamespage> rbasak, I just tested it and uploaded
[12:28] <rbasak> jamespage: thanks. Just this second my test succeeded.
[12:28] <jamespage> rbasak, http://paste.ubuntu.com/845713/
[12:29] <jamespage> rbasak, please can you make sure that gets submitted back to debian
[12:29] <jamespage> I would suspect they have the same issue
[12:29] <rbasak> Interesting - I get something slightly different.
[12:30] <rbasak> http://pastebin.com/vPvEc0BU
[12:30] <rbasak> I wonder what libopen-pal is? Anyway I don't suppose it matters.
[12:30] <rbasak> I'll file a Debian bug now.
[12:30] <jamespage> rbasak, you know submittodebian?
[12:31] <rbasak> I'd heard of it, but never used it. I shall use it now, thanks :)
[12:32] <jamespage> saves alot of time!
[12:34] <rbasak> jamespage: hmm. It's not going to work from the mostly-unconfigured with no functioning mail cloud instance I have the sources on! :-/
[12:42] <jamespage> no
[12:45] <Daviey> rbasak: no functionig mail ?
[12:46] <rbasak> Daviey: not unless you appreciate emails from root@localhost and your spam filter allows them through, or I configure stuff correctly. The cloud I'm using has outbound SMTP blocked at the moment, too.
[12:49] <Daviey> rbasak: configuring outbound isn't  challenge, and the key is that port 25 is blocked, not SMTP :)
[12:50] <rbasak> Daviey: I don't want to put my outbound submission credentials on a cloud instance either!
[12:50] <rbasak> (nor have I got round to automating it in some way)
[12:51] <Daviey> rbasak: ah, i created a special user just for it.
[12:56] <rbasak> Daviey: I don't have that luxury :)
[13:52] <zul> morning
[13:53] <jamespage> hey zul
[13:58] <zul> hey jamespage
[14:18] <smoser> rbasak, http://paste.ubuntu.com/845863/
[14:18] <smoser> that works now with latest cloud image
[14:18] <smoser> basically a much saner way than OVF manipulation to seed the image.
[14:20] <rbasak> smoser: that's brilliant - thanks!
[14:26] <smb> smoser, Awesome, this answers my question that I was thinking to ask about a minimal set of data for meta and user data. :)
[14:31] <smb> smoser, Maybe one additional question about that: can the password that way set to be empty (just having nothing beyond the password:)?
[14:31] <smoser> smb, are you asking if you can set an empty password ?
[14:32] <smb> smoser, Yes, that was my intention
[14:32] <smoser> it doesn't look like it.
[14:32] <smoser> cloud-init just uses chpasswd
[14:32] <smoser> and i can't get chpasswd to take an empty passwd.
[14:32] <smoser> smb, are you wanting ssh login ?
[14:33] <smoser> that can be done easily enough by inserting keys (which you can via user-data)
[14:33] <smb> Ah ok. Well just trying to get an end result similar to ec2. Not really sure the pw there is empty actually or whether sudoers just allows anything without a pw
[14:34] <smoser> smb, ah. yeah.
[14:34] <smoser> sudoers for the ubuntu user does not require a password
[14:34] <smoser> so what you want is to do something like this:
[14:36] <smb> Up to now I did some sort of uncloudify and used "passwd -d ubuntu" which is wrong in several ways I guess. :)
[14:37] <smoser> smb, http://paste.ubuntu.com/845895/
[14:37] <smoser> use that to write your user-data and your ssh-keys will get imported.
[14:37] <smoser> or use ssh-import-id
[14:37] <smoser> (which can be used from user-data too)
[14:38] <smoser> #cloud-config
[14:38] <smoser> ssh_import_id: stefan-bader-canonical
[14:39] <smb> I see. Ok. I'll giive those a try
[14:39] <smb> thanks
[14:50] <jamespage> hallyn, around? seeing and odd issue with lxc with duplicated mac addresses
[14:56] <jamespage> hmm - maybe utlemming then?
[15:01] <MTecknology> I'm trying to make a preseed file but getting stuck on the partitioning piece. I think I have everything else figured out. Any chance I could convince someone to take a peak at what I have? http://dpaste.com/704419/
[15:01] <caribou> smb apw, I have a kernel question for you
[15:01] <hallyn> stgraber: workign o nthe lxc server guide - i'm going to work on getting all the other templates to work too, as part of that.
[15:01] <caribou> off the top of your head, would you remember any regression b/w Oneiric & Natty wrt the megasas module ?
[15:02] <stgraber> hallyn: good luck with fedora ;)
[15:02] <apw> caribou, not i
[15:02] <smb> not from the top of my head
[15:03] <caribou> The Oneiric installer is no longer to see the raid drives but it works well on Natty
[15:03] <hallyn> stgraber: ya
[15:03] <smb> caribou, Hm, I only remember some not yet supported cases for Lucid...
[15:05] <caribou> apw: smb here is the contents of the syslog with DEBCONF_DEBUG=5 on both installs : http://paste.ubuntu.com/845931/
[15:05] <caribou> it's using the same preseed file on both btw
[15:05] <smb> caribou, Hm... wasn't that maybe the driver that renamed devices... or was that another thing
[15:06] <caribou> I'll try the same with the precise daily build next week to compare
[15:07] <smb> caribou, One of your final lines in the oneiric case seems to have a scsi4 found and bound...
[15:08] <caribou> smb: yes, but after that there is nothing more and d-i goes ahead and ask which module should be loaded
[15:09] <caribou> smb: wait, no, the lines were much further & diluted in the log, my fault
[15:09] <smb> caribou, Right, so somehow the drive is not found, or none to say
[15:10] <caribou> smb: here is the rest of the log : http://paste.ubuntu.com/845939/
[15:11] <smb> caribou, Sorry got me a bit confused: is that oneiric or the natty one?
[15:12] <caribou> smb: sorry, this is the follow-up of the Oneiric syslog
[15:12] <smb> ah ok
[15:12] <smb> hm. there is some disk info...
[15:12] <caribou> smb: at the end it goes to SUBST disk-detect/module_select to ask for which module needs to be loaded
[15:13] <caribou> smb: with Natty, it continues without asking anything
[15:14] <smb> caribou, Yes. looks a bit weird. I see the info from the host scan, but not any sd devices being created
[15:15] <uvirtbot`> New bug: #934256 in lxc (universe) "lxc-clone should replace/change hwaddr when cloning containers" [Undecided,New] https://launchpad.net/bugs/934256
[15:16] <caribou> smb: could this be caused by a race condition because of the fact that the host scan happens much later in the Oneiric log than on Natty where it appears right after the module is loaded ?
[15:19] <smb> caribou, Not sure. Actually it does not really seem to be "much" later. Just about one second after the messages about the adapter init
[15:19] <smb> and that could be just becuase we were at the end of that second
[15:29] <hallyn> jinkeys man pages need updating too
[15:31] <michael_tn> good day all, is this the correct channel for orchestra/juju discussions?
[15:32] <kirkland> michael_tn: sure
[15:32] <michael_tn> :-)
[15:32] <michael_tn> great,
[15:32] <kirkland> michael_tn: this channel is a bit better for orchestra;  #ubuntu-juju is a bit better for juju
[15:32] <michael_tn> we're experimenting with the build on some cluster hardware
[15:33] <michael_tn> and i anticipate a bit of handholding may be required
[15:33] <michael_tn> we have a pretty nice cluster we're repurposing to openstack
[15:35] <michael_tn> we have orchestra up and are about ready to move on to juju, but were a bit unsure whether to go ahead and provision a group of nodes in orchestra before bootstrapping juju or to do it after
[15:41] <jamespage> michael_tn, do you want to manage all of the nodes you have through juju?
[15:43] <michael_tn> hmmm ...
[15:43] <michael_tn> not sure, not considered that really
[15:44] <jamespage> michael_tn, OK; so for the node that you want juju to manage they have to be in a certain class in cobbler (part of orchestra)
[15:44] <michael_tn> the entire cluster is commited to openstack (two head nodes and 64 cluster nodes)
[15:45] <SpamapS> michael_tn: you simply need to identify them in orchestra, and put them in the 'available-mgmt-class' that you defined in your juju environments.yaml...
[15:45] <michael_tn> ok
[15:45] <SpamapS> michael_tn: juju will do the provisioning
[15:45] <michael_tn> juju will select those nodes at random?
[15:46] <jamespage> michael_tn, at the moment yes
[15:47] <michael_tn> ok
[15:48] <michael_tn> should we instantiate a single node to bootstrap juju at the start or just let bootstrap pick one out of our group?
[16:03] <jamespage> michael_tn, juju will pick one
[16:16] <Daviey> SpamapS: would you be able to look at bug 934013 ?
[16:16] <uvirtbot`> Launchpad bug 934013 in mysql-5.5 "Mysql fails to start after upgrade to precise" [Undecided,New] https://launchpad.net/bugs/934013
[16:18] <michael_tn> ok, thanks, i'll spin up a few nodes and see if i can get juju bootstraped
[16:19] <michael_tn> our WOL is not working so i have to go boot things by had :-\
[16:25] <kyconquers> what does this error mean? http://pastebin.com/ANuDMQRR I have narrowed it down to one line, I am clicking on an input field.
[16:29] <brendan0powers> jamespage: Hi, thanks for taking another look
[16:31] <SpamapS> Daviey: I think I fixed that yesterday...
[16:31] <SpamapS> Daviey: oh, no.. hrm.. thats the new version
[16:32] <SpamapS> Looks like the apparmor profile is still borked
[16:32] <Daviey> :(
[16:36] <SpamapS> Daviey: I think its actually caused by incomplete /run migrations... but we should still be able to run things. :p
[16:36] <jamespage> hey brendan0powers: thanks for doing those updates
[16:37] <jamespage> brendan0powers, so - qtrpc2 - why do you need it within the rds packaging?
[16:37] <jamespage> it could be packaged separately?
[16:37] <Daviey> SpamapS: Make sure you thank lennart.
[16:37] <SpamapS> Daviey: whoa wait.. this is weird
[16:38] <SpamapS> Daviey: I think it may be a confflie transfer of ownership issue
[16:38] <brendan0powers> jamespage: It could be, but rds is the only software that uses it(that I'm aware of), and since it lives in the rds VCS, it's released at the same time Resara Server is
[16:39] <Daviey> SpamapS: This adds to my confidence of upgrade testing :)
[16:40] <robbiew> heh
[16:46] <jamespage> brendan0powers, so its part of rds? I think I'm misunderstanding how it relates to rds
[16:48] <brendan0powers> jamespage: QtRPC was a library Resara developed for our last product, which was closed source. At one point, we posted a version of it on google code with an open source license
[16:48] <jamespage> brendan0powers, ah - I see - so it versions and releases inline with rds
[16:48] <brendan0powers> jamespage: But now, it lives in the Resara server repository, with an open source license
[16:48] <brendan0powers> yes
[16:49] <jamespage> brendan0powers, OK - look I'll do the updates to the package as I documented in the bug report
[16:49] <brendan0powers> Oh, I'm already part way through that
[16:49] <jamespage> brendan0powers, excellent!
[16:50] <jamespage> brendan0powers, so the normal place to document things folk should understand about stuff like this is in debian/README.source
[16:50] <jamespage> could you add a few notes while you are doing the updates?
[16:50] <brendan0powers> Notes about what?
[16:50] <brendan0powers> QtRpc?
[16:50] <jamespage> brendan0powers, yep
[16:50] <brendan0powers> Ok, I will do that
[16:50] <jamespage> +1
[16:55] <michael_tn> off to data center, bbiaw with more questions i imagine :-)
[16:55] <michael_tn> thanks for the help!
[17:05] <MTecknology> preseed files seem to be nice and well documented until it comes to partitioning.... :(
[17:06] <uvirtbot`> New bug: #760443 in bacula (main) "Bacula installation fails at dbconfig-common stage" [Medium,Confirmed] https://launchpad.net/bugs/760443
[17:09] <brendan0powers> jamespage: I'm not sure I understand the copyright stuff
[17:09] <jamespage> brendan0powers, leave that with me then - I'll update
[17:09] <brendan0powers> jamespage: I modified the 2 LGPL-2 licenses to be LGPL-2+
[17:10] <brendan0powers> jamespage: ok, then all the issues should be fixed
[17:10] <brendan0powers> I need to run i through pbuilder to make sure
[17:12] <jamespage> brendan0powers, OK - let me know when you are happy
[17:12] <jamespage> I'll re-clone and update as required.
[17:31] <brendan0powers> jamespage: Ok, everything is pushed to the repo
[17:31] <jamespage> brendan0powers, ack
[17:31] <brendan0powers> jamespage: you can re-clone and take a look
[17:34] <hallyn> jamespage: good catch on lxc-clone and hwaddrs :)
[17:34] <hallyn> i'll add that to the HUGE SLEW of changes i'm doing today
[17:34] <jamespage> hallyn, w00t!
[17:34] <jamespage> took me a while to figure out what was going on
[17:53] <hallyn> stgraber: fedora template pretty much working.  though repos aren't quite working after start...  more work to do
[17:56] <Plizzo> I have an installation of Ubuntu Server 11.10 x64 which is freezing completely at random occasions. I've tried to modify the grub parameters, change kernel, running memtests, tailing logs, but nothing helps. Does anyone here know how to solve this?
[17:58] <brendan0powers> jamespage: It seems there is an issue with the packages in the repo
[17:59] <brendan0powers> jamespage: librdsserver1 is depending on librdslap, and not librdsldap1
[17:59] <brendan0powers> It's not listed in the control file, so debhelper must be generating the dependency
[18:00] <jamespage> brendan0powers, that maybe an shlibs problem
[18:01] <jamespage> brendan0powers, shlibs.local
[18:01] <jamespage> needs a tweak
[18:04] <brendan0powers> forgot about that file
[18:06] <jamespage> it kinda works i.e. it will install because librdsldap1 replaces librdsldap
[18:07] <jamespage> brendan0powers, anyway I have to finish reviewing this on Monday  - thanks for all of the updates you have done to  the package
[18:07] <brendan0powers> Ok
[18:07] <jamespage> once I've confirmed it builds and installs OK I'll get the FFe and get it uploaded.
[18:07] <brendan0powers> Awsome
[18:07] <brendan0powers> I'll continue poling around a little more and see if I find any more issues
[18:08] <brendan0powers> I'll update the bug report if I find anything
[18:18] <adam_g> zul: thoughts on reenabling the libvirt console patch on CI?
[18:19] <zul> adam_g: im trying it out locally still
[18:21] <zul> adam_g: if im happy with it then ill re-enable it
[18:22] <adam_g> hm ok
[18:24] <Potatoe> I am sending sip packets to my server. On my server I am running ngrep on the interface where the packets are coming from. Some packets are not arriving in ngrep even though I see them on tcpdump leaving the remote site. What's odd is it's always the same packets that are being dropped. is there any filter that might drop packets before ngrep ?
[18:34] <Potatoe> Can packets by dropped by the kernel before hitting ngrep or tcpdump ?
[18:35] <rbasak> Potatoe: as far as I'm aware, no, provided that you're capturing on a real interface (not a bridge or something).
[18:35] <rbasak> Potatoe: I'd use tcpdump on both ends with identical filters to verify.
[18:35] <Potatoe> eth0 which is a nic on the moboard
[18:47] <Potatoe> it says '0 packets dropped by the kernel;
[18:53] <brendan0powers> jamespage: I have updated the repo with the shlibs.local fix, rebuilt the packages, and verified they install
[18:56] <JanC> Potatoe: what type of traffic isn't seen by ngrep?
[19:09] <Potatoe> SIP Packets
[19:09] <Potatoe> but only some are disappearing, so it's not a nat issue
[20:01] <RoyK> Potatoe: anything in iptables_
[20:01] <Potatoe> No
[20:01] <Potatoe> I reset everything iptables-save and see nothing
[20:01] <Potatoe> accept by default
[20:01] <RoyK> then kernel really shouldn't drop anything
[20:20] <Potatoe> That's what I thought
[20:20] <Potatoe> It's my understanding that ngrep/tcpdump shows data before it hits iptables
[20:23] <RoyK> Potatoe: iirc before the filter table, yes, but not sure about the rest
[20:23] <RoyK> Potatoe: anyway - what sort of software is this?
[20:25] <Potatoe> Ubuntu 10.04 LTS with Kamailio 3
[20:27] <Potatoe> I think it might be related to the ethernet bonding rules which I know little about
[20:38] <patdk-wk> heh? your using bonding?
[20:38] <patdk-wk> you better be very careful what mode you use, and make sure your nic's and switch supports it
[20:39] <patdk-wk> plus whatever you talking to supports that mode also (for a few modes)
[20:41] <Potatoe> how do I get the network card back into 'master'
[20:41] <stgraber> hallyn: uploading udev now
[20:42] <patdk-wk> unplug the other network cards?
[20:42] <Potatoe> I tried to activate the bonding mode
[20:42] <Potatoe> it didn't work
[20:42] <Potatoe> now the network card won't work in non bonded mode
[20:42] <Potatoe> it says 'RUNNING SLAVE'
[20:42] <patdk-wk> heh?
[20:42]  * patdk-wk might recommend learning how bonding works, before you go and use it
[20:42] <Potatoe> I tried to activate bonding
[20:42] <Potatoe> now nothing works
[20:43] <Potatoe> I reset the interfaces script to what I had before
[20:43] <Potatoe> and ran /etc/init.d/networking restart
[20:43] <patdk-wk> that won't unbond
[20:43] <patdk-wk> in fact, that is a horrible idea to do, ever
[20:43] <Potatoe> how do I go back
[20:43] <patdk-wk> this is what ifup/ifdown is for
[20:43] <patdk-wk> at this point? reboot
[20:49] <Potatoe> that fixed the original config
[20:55] <patdk-wk> what bonding mode did you attempt?
[21:13] <joren> Hey, I've got a serve rwith a raid setup that I just replaced a disk in, and it's taking for ever to come up now (been waiting 5 minutes so far.) I don't see any disk activity, and the last output was /dev/md1 clean
[21:25] <joren> is there any way to get a more verbose output of what it's doing?
[21:26] <jacobw> what model of raid controller is it?
[21:30] <joren> software raid
[21:30] <joren> maybe the array is just being slow, but it really doesn't seem to be hitting the disks at all
[21:31] <joren> I'll give it a bit longer and then just remove the array from fstab I guess
[21:36] <uvirtbot`> New bug: #934515 in python-novaclient (main) "--meta key=value cannot accept '=' in value" [Undecided,New] https://launchpad.net/bugs/934515
[22:11] <uvirtbot`> New bug: #934534 in nova (main) "metadata ami-launch-index coming back off-by-one" [Undecided,New] https://launchpad.net/bugs/934534
[22:43] <uvirtbot`> New bug: #934540 in bacula (main) "package bacula-common-pgsql 5.2.3-0ubuntu2 failed to install/upgrade: trying to overwrite '/usr/lib/bacula/libbaccats.la', which is also in package bacula-common 5.2.3-0ubuntu2" [Undecided,New] https://launchpad.net/bugs/934540
[22:47] <vindav> Hello, I am a Ubuntu 11.10 newbie gaining experience and have a new project in which I want to back up my local drive to my local network access storage drive (LaCie d2).  Can anyone provide a solution or a step by step process of how this can be achieved please ?
[23:09] <vindav> I we'll if it can help the majority, 'Grsync' is the resident program that is used to accomplish the task I am looking to do to do local drive to a network access storage device.  I tested this out and found that I can do so far a single folder transfer with contents.  Now, if I can do multiple folders in a transfer with automatic timely transfers, I think we'll have something happening.  Any comment or advise ?
[23:21] <Plizzo> I have a server running Ubuntu Server 11.10 x64 and that is randomly freezing, any help?
[23:23] <Spanky> Plizzo -- Sounds like hardware crap to me...
[23:23] <Spanky> Memory?
[23:23] <Plizzo> Spanky: I did 23 passes of memtest, no errors
[23:23] <Spanky> Try running off a LiveCD and do a few things...
[23:24] <Spanky> Random freezing sounds like hardware....  These things can be a bitch to troubleshoot
[23:25] <Plizzo> I know, I've been trying to fix it for weeks.. :(
[23:25] <Spanky> See if the box runs 10.04LTS or run another server distro or WindowsServer or something and see what happens...
[23:25] <Spanky> You got any spare boxes in the closet....
[23:25] <Spanky> ?
[23:27] <Spanky> You could try a forked off server like Zentyal or even try Debian Server or what not...
[23:27] <qman___> if it freezes up without a kernel panic that's almost guaranteed hardware
[23:28] <Spanky> Just doesn't sound like a Server OS issue when you have random freezes...
[23:28] <qman___> even most freezes with kernel panics are hardware, but almost all freezes that are software will have a kernel panic
[23:28] <Spanky> Yep...
[23:28] <Plizzo> qman__: The logs are empty
[23:29] <Spanky> This is a new install right?
[23:29] <Plizzo> I've tried tailing all logs, but there are no irregulaatories
[23:29] <Plizzo> Spanky, not entirely new
[23:29] <Plizzo> I've had it since around new years
[23:29] <Spanky> Ewww...
[23:29] <Plizzo> And the problem seems to have started after I ran sudo update-grub
[23:29] <Plizzo> the first time
[23:29] <Plizzo> But at that time I had made no changes to the config file..
[23:30] <Spanky> Can you try a new install on a separate drive and just check the hardware...?
[23:30] <qman___> grub has no effect on the server while it's running
[23:30] <qman___> only when it's booting
[23:30] <Plizzo> Could the graphics drivers be causing this?
[23:30] <Spanky> Grub 2 has been pretty solid for me on the desktop.  I don't multi-boot my server hardware though so I don't get into the wooly stuff there...
[23:31] <qman___> without X, there aren't really any graphics drivers per se
[23:31] <qman___> if you're running KMS or a framebuffer you can try booting with it disabled
[23:31] <Plizzo> I run XBMC
[23:31] <Plizzo> XBMC-Live
[23:31] <Plizzo> But even without that and Xorg running, it still freezes
[23:32] <qman___> recent versions uses KMS by default, try booting with the option 'nomodeset'
[23:32] <Spanky> Yeah, I run Zentyal Server that uses LXDE and it's super low graphics overhead....  So even with X, not a lot of issues I've seen...
[23:32] <Plizzo> qman__ I have nomodeset already
[23:32] <Plizzo> I have an Nvidia 520 GPU in the server for 1080p decoding
[23:32] <Plizzo> And vdpau drivers installed
[23:33] <qman___> well, it could be a problem
[23:33] <qman___> but like I said
[23:34] <qman___> freezing up, if it's a software problem, almost always gives a kernel panic
[23:34] <qman___> like, five nines
[23:34] <Plizzo> Should I boot from a Live-CD usb and try having it on to see if it freezes?
[23:34] <Spanky> I'm out.... I don't put big graphics boards in my server boxes.
[23:34] <qman___> worth a shot
[23:35] <Plizzo> Alright, thanks
[23:35] <Spanky> As I suggested...  You have to nail down the problem... What server application you using that requires a NV 520?
[23:35] <qman___> while it's not really wrong in any way, media centers are a bit outside the normal scope of this channel
[23:36] <qman___> I'd bet money it's a hardware problem though
[23:36] <Spanky> Yeah, most of us shiners are just running a file, print server and maybe some FTP or WWW....
[23:36] <qman___> could try some sort of graphics card testing utility
[23:37] <Spanky> Can you boot Ubuntu desktop on that puppy?
[23:37] <Spanky> Live CD?
[23:42] <hallyn> so to be clear, we're not in ff, but right now we're not in any sort of upload freeze (for bugfixes) right?
[23:48] <hallyn> hm, lxc-start-ephemeral may need the macaddr fix too as lxc-clone did