[00:01] <hallyn> Daviey: ikonia: my last round of kvm perf runs didn't do lvm.
[00:02] <hallyn> Daviey: ikonia: but my first (much less useful) set of runs implied lvm to be not really faster than raw file ( http://s3hh.wordpress.com/2012/01/07/some-first-performance-tests-of-various-kvm-backing-stores/ )
[00:02] <hallyn> Daviey: ikonia: can't wait to see ikonia's results - thanks!
[00:03] <hallyn> stgraber: ok, i think i'm going to push my lxc.  i'm using /usr/share/lxc/hooks/.  shoult if you really feel example-hooks is clearer
[00:04] <stgraber> hallyn: I'm fine with lxc/hooks/ we'll probably have to make sure to document it properly though (server guide I guess)
[00:12] <hallyn> stgraber: d'oh, yeah, the 12.10 server guide updates should probably be in work items
[00:13] <stgraber> hallyn: indeed
[00:14] <stgraber> hallyn: btw, lxc was moved to -updates today, I'll prepare a branch for the next one, see if we already have enough for another SRU
[00:14] <hallyn> stgraber: i'm going to hope not :)
[00:15]  * hallyn would like things to calm down a bit in precise-updates
[00:15] <hallyn> stgraber: also, given how long the lxc api work item is, it really should be broken up.  do you have a concrete plan in your midn of remaining steps?
[00:16] <stgraber> hallyn: looking at the current changelog, I have 4 SRU-able changes already, but fairly small, so might while a little while to see if we get more
[00:17] <stgraber> hallyn: might upload one at the end of this week or early next week then
[00:17] <hallyn> ok
[00:18] <stgraber> (changes I'm looking at are the fstype=fuse.* apparmor change, fixing lxc-list/upstart-job to work with both the Ubuntu and Deian way of doing auto-start, depend on adduser (for postinst script) and fix a dh_apparmor call)
[00:19] <stgraber> oh, and fix dpkg to use --add-architecture on >= quantal
[00:19] <stgraber> that last one is issuing a warning on every single dpkg call, so I guess it's quit important to SRU :)
[00:21] <hallyn> wait.  i thought ubuntu and debian autostart were now the same?
[04:46] <tarvid> how do I determine which device is connected to an interface?
[04:46] <tarvid> ntop does not like eth1, I have an onboard nic and a usb network interface
[04:50] <jmarsden> tarvid: /sbin/ifconfig -a     # should help?
[04:54] <tarvid> that shows eth1 but when I configure ntop with eth0 eth1 it fails saying invalid interface
[04:55] <jmarsden> Maybe ntop needs the device names /dev/eth0 and /dev/eth1 ?
[04:57] <tarvid> jmarsden: thanks turns out to have been a comma separated list
[04:58] <jmarsden> tarvid: You're welcome -- at least I put you on the right track :)
[04:58] <tarvid> yes you did
[04:59] <tarvid> want to solve #2? my logs are empty after logrotate but service restart rsyslogd starts them
[05:05] <jmarsden> tarvid: Sounds like a logrotate configuration issue of some sort, but I'm not really sure what to suggest you look for.
[05:06] <tarvid> I have seen it before, perhaps a couple of years ago
[05:06] <tarvid> may be a sequence issue
[05:08] <tarvid> now ntop says I must set -M (merging?)
[05:08] <jmarsden> You could try mv'ing everything in /etc/logrotate.d/ somewhere else, then run logrotate and see if the issue still occurs... then you can move the individual config files back in until the issue happens, and so (hopefully) figure out which of them is causing the problem.
[05:09] <tarvid> a manual service rsyslogd restart always works
[05:13] <matt_keys> Having some issues with the kernel talking to my sata controller. See : http://pastebin.com/vNWS99Vt . Anybody have any suggestions on what to try?
[05:21] <tarvid> have you had a peek at gsmartctl?
[05:25] <tarvid> jmarsden: had to patch /etc/init.d/ntop to include -M option to get eth1 stats
[05:27] <jmarsden> tarvid: Seems odd, but OK... it's been a while since I used ntop, and I think when I did it was on machines with only eth0
[05:27] <tarvid> something about netflow plugin
[05:28] <tarvid> I've got a convoluted fiber and wireless network behind eth1 that I would like to know more about
[05:30] <tarvid> eventually identify clients by mac address of CPE which might be two or three hops inside
[05:31] <tarvid> they typically have one or more computers connected to a wireless router connected to a radio or a fiber termination which I know nothing about
[05:31] <tarvid> the radio mac is static (until I change them out) everything else could change
[05:32] <tarvid> jmarsden: matt_keys question prompts one more - is gsmartctl the best disk subsystem diagnostic tool?
[05:35] <jmarsden> tarvid: This is #ubuntu-server, so I use smartctl rather than the GUI interface :)  But yes I think for 'normal' disk drives at least, smartctl from the smartmontools package works fine.
[05:37] <tarvid> good point, I use the cli version on the headless servers, I questioned a few kernel errors until I read the SMART error reports, didn't want to believe both hard drives in RAID1 were bad
[06:24] <matt_keys> tarvid : i'll try out smartctl, bios is set to ehci on these disks and it's got two controllers; a marvell and intel. when disks are on marvell they don't show up at all, had to move them all to intel
[06:59] <tarvid> Good luck matt_keys and thanks jmarsden. A good night for me. Wishing you all well.
[07:16] <soren> Daviey, rbasak: uvirtbot announces bugs as new the first time it hears about them. Under normal circumstances this is when the bug is first reported. At times, it can be delayed if a bug is wrongfully reported against a package uvirtbot doesn't care about and then gets reassigned.
[07:18] <soren> Daviey, rbasak: Other times still, it could be a bit of a backlog from the days when the uvirtbot server was off-line for long enough that lists.ubuntu.com stopped sending it e-mails on the ubuntu-server-bugs list. In those cases, the initial bug report never reached uvirtbot and so now it announces them when it gets  anew e-mail about it, i.e. when someone changes its status, comments on it, etc.
[07:19] <soren> Daviey, rbasak: Uh, yeah, the core issue is that it's all e-mail driven. It's subscribed to packages' bugs, gets e-mail and polls the mailbox over imap. Cool, huh?
[07:20] <soren> One of these days, I'll get off my lazy behind and get it to talk to the LP API. Not today.
[07:26] <koolhead11> hi all
[07:46] <Daviey> soren: slacker :)
[07:52] <rbasak> soren: what kind of link does the server have?
[07:57] <koolhead11> Daviey: hello sir
[08:07] <Daviey> koolhead11: hey!
[08:07] <Daviey> rbasak: probably ethernet :)
[08:07] <koolhead11> Daviey: what about my early morning wish, i pinged you!! :)
[08:10]  * Daviey look back
[08:11] <Daviey> koolhead11: < koolhead11> Daviey: hello sir .. i said hello back!
[08:12]  * koolhead11 needs a mentor
[08:12] <Daviey> koolhead11: roaksoax and ivoks mostly drove clustering, but have since been working on other things.
[08:13] <Daviey> i think roaksoax is probably still the closest to the issue, but i don't believe it's getting as much attention as it was.
[08:13] <koolhead11> ooh okey. :(
[08:14] <Daviey> smb: i386 kernel seems to be rocking!
[08:15] <smb> Daviey, Glad to hear. :) Though I was somewhat sure it would at least perform better than the previous one as I had it up on an instance. ;)
[08:16] <Daviey> smb: heh
[08:16] <koolhead11> Daviey: for the Folsom i was thinking to integrate a chapter on HA on the beginners guide. I saw the bluprint so i felt if some magic can happen :D
[08:16] <Daviey> koolhead11: that sounds like a great idea.
[08:17] <Daviey> koolhead11: roaksoax will be around a little later, he is undoubtedly the best person to speak to.
[08:17] <koolhead11> Daviey: awesome will ping him and plan ahead
[08:17] <Daviey> super
[08:18] <koolhead11> we need quantim/cinder and HA additnitonal component if i can make it
[08:18] <koolhead11> *quantum
[08:30] <lynxman> morning o/
[08:38] <koolhead11> hi lynxman
[08:41] <lynxman> koolhead11: ello o/
[09:35]  * koolhead11 is having bad day with java
[09:35] <koolhead11> die oracle
[09:35] <koolhead11> phewwwwwww
[09:36] <glance> sounds more like a regular day with java =)
[09:38] <matti> Haha
[09:43] <eagles0513875_> !raid
[09:45] <eagles0513875_> !openstack
[09:46] <koolhead11> eagles0513875_: we need to add more brain to the stupid bot
[09:46] <eagles0513875_> koolhead11: ya im noticing
[09:46] <koolhead11> glance: matti :(
[09:46] <eagles0513875_> has anyone in here used ubuntu-vm-builder package?
[09:46] <eagles0513875_> and or openstack?
[09:47] <koolhead11> eagles0513875_: what about it
[09:47] <eagles0513875_> does the ubuntu-vm-builder need a DE to use it?
[09:47] <koolhead11> eagles0513875_: http://docs.openstack.org/essex/openstack-compute/starter/ this might help
[09:47] <eagles0513875_> koolhead11: thing is its already in the repos of 12.04
[09:48] <koolhead11> eagles0513875_: no idea Ubuntu Devs would be right folks
[09:48] <ikonia> it's a gui based system
[09:48] <eagles0513875_> ok
[09:48] <ikonia> but there is a command line interaction
[09:49] <eagles0513875_> i think i might given openstack a shot
[09:49] <eagles0513875_> just not sure out of the decent list of packages what needs to be installed
[09:49] <ikonia> you install the package and it will pull in the dependencies for you
[09:50] <eagles0513875_> i did an apt-cache search open | grep stack and it pulled a fair number of packages all for openstack
[09:51] <ikonia> ok...?
[09:51] <eagles0513875_> except mfor two packages
[09:52] <ikonia> have you read any of the documentation on this before rushing in ?
[09:54] <eagles0513875_> i have the documentation up and reading through it but it seems like there is no mention of quantum at least not what I have read so far or see in the contents of the documentation
[09:54] <ikonia> eagles0513875_: what are you actually trying to install ?
[09:54] <eagles0513875_> openstack
[09:54] <ikonia> so why are you referencing quantum ?
[09:54] <eagles0513875_> from 12.04 repositories
[09:54] <eagles0513875_> its in the repositories
[09:55] <eagles0513875_> ikonia: do apt-cache search Openstack and youll see quantum listed
[09:55] <eagles0513875_> oops
[09:55] <ikonia> eagles0513875_: right ? but why are you looking at that ?
[09:55] <eagles0513875_> i stand corrected
[09:55] <eagles0513875_> my first search was an incomplete one
[10:01] <jamespage> Daviey: I have a work item to send a call for testing for A2 images - are we in a good position todo that now?
[10:09] <Daviey> jamespage: Today we are..
[10:10] <Daviey> jamespage: I'm sending a general one out soonly.
[10:10] <Daviey> but if you want to specifically rally server, that is awesome.
[10:10] <jamespage> Daviey, ack - will do
[10:10]  * jamespage goes to steal smosers copy from last time
[10:11] <Daviey> jamespage: smoser tried to rally for A1.. SO if you do the same for A2, that is great... Please share with balloons (Community/QA) what you have done.
[10:11] <Daviey> Thanks.
[10:11] <jamespage> Daviey, will do
[10:22] <jamespage> Daviey, done - cc'ed balloons
[11:19] <koopa58> Hi all, does anyone know how I can find out more on a 404 entry in logwatch httpd section? I can't find the log file it was taken from
[11:40] <koopa58> is it me, or is it quiet in here :)
[11:50] <IdleOne> koopa58: it is always you :)
[11:51] <koopa58> haha :P
[11:51] <IdleOne> is quiet though. someone will be along sooner or later
[11:52] <koopa58> okay, I'll wait a while :)
[12:21] <smoser> yippee! green: https://jenkins.qa.ubuntu.com/view/ec2%20AMI%20Testing/view/Overview/job/quantal-server-ec2-daily/
[12:21] <smoser> thank you smb ^
[12:22] <smb> smoser, Welcome. :) We all like green. D
[12:22] <smoser> jamespage, can you go ahead and start a quantal-server-ec2 run of 20120627 ?
[12:23] <smoser> or, if you'd rather, you can teach me if that is possible
[12:29] <jamespage> smoser, sure
[12:36] <Daviey> jamespage / smoser: Are all ec2 builds not auto tested now?
[12:36] <jamespage> Daviey, they all get a smoke test
[12:36] <jamespage> but not the full service
[12:36] <Daviey> ahh
[12:36] <Daviey> jamespage: Give me a full service please!
[12:36] <jamespage> Daviey, would do if I could get to jenkins...
[12:45] <zul> Daviey: hehe...you said full service
[13:01] <koopa58> Hi all, is it possible to get more info from logwatch? I need IP and user agent from the httpd section (404 errors and such). Currently it only shows the request-uri
[13:12] <RoyK> koopa58: not sure, but logwatch is written in perl, so it shouldn't be very hard to add that functionality if wanted :)
[13:22] <SpamapS> koopa58: you could go look through the logs directly too :)
[13:24] <eagles0513875_> hey guys my raid array is degraded how do I go about fixing the situation? all i did during reinstall was use the same raid setup as previous install
[13:25] <eagles0513875_> hey SpamapS  :D
[13:26] <RoyK> eagles0513875_: pastebin content of /proc/mdstat, please
[13:26] <eagles0513875_> atm rebooted it and now its hanging for some odd reason
[13:26] <eagles0513875_> will try a regular kernel instead of the xen kernel
[13:27] <RoyK> !kvm
[13:27] <RoyK> ;)
[13:27] <eagles0513875_> whats so great about kvm
[13:27] <eagles0513875_> actually it came up
[13:27] <RoyK> kvm works well, but xen may be faster in some circumstances
[13:28] <eagles0513875_> i have had great experiences with xen havent worked much with kvm
[13:28] <eagles0513875_> kvm is kernel based virtualization right
[13:28] <eagles0513875_> RoyK: what would you like to know from cat /proc/mdstat
[13:28] <eagles0513875_> wait a min lol
[13:29] <RoyK> but I like kvm because of its simplicity - it doesn't annex the entire computer like xen does ;)
[13:30] <eagles0513875_> RoyK: http://paste.ubuntu.com/1062537/
[13:30] <eagles0513875_> forgot i had ssh on the server which is sitting right next to me lol
[13:30] <RoyK> so I guess sdb is dead?
[13:30] <RoyK> does it show up under /proc/partitions?
[13:31] <eagles0513875_> RoyK: on this install they are sda
[13:31] <eagles0513875_> actually sdb does show up ya
[13:31] <RoyK> yeah, but that shows a mirror with one side gone
[13:31] <RoyK> or two mirrors
[13:31] <eagles0513875_> let me paste ya proc partitions
[13:32] <eagles0513875_> http://paste.ubuntu.com/1062540/ RoyK
[13:33] <RoyK> eagles0513875_: was the missing side of those mirrors on sdb?
[13:33] <RoyK> if so, something is badly messed up there, since sdb doesn't list any partitions
[13:33] <eagles0513875_> RoyK: nothign was missing prior to my reinstall
[13:34] <eagles0513875_> when reinstalling it asked me if i wanted to use the same raided partition layout on the installer and i told it to do so
[13:34] <RoyK> well, there are no partitions on sdb
[13:34] <eagles0513875_> can delete the raid setup and recreate it from scratch?
[13:34] <RoyK> no need
[13:34] <eagles0513875_> how can this be fixed?
[13:35] <RoyK> you can create partitions on sdb of equal size of those on sda and mdadm --add the partitions to those degraded mirrors
[13:35] <eagles0513875_> do you have a good document I could follow?
[13:36] <RoyK> eagles0513875_: just use fdisk to create those partitions
[13:36] <eagles0513875_> ok
[13:37] <RoyK> first look at sda to find the right size, then run it with sdb to create those
[13:37] <patdk-wk> if your really lazy, dd the first sector :)
[13:37] <RoyK> hehe
[13:37] <eagles0513875_> lol
[13:37] <patdk-wk> after your all done, make sure you redo grub install
[13:38] <RoyK> bbl
[13:38] <eagles0513875_> ok
[13:38] <eagles0513875_> think ill stick with fdisk
[14:04] <Jeeves_> http://www.change.org/petitions/canonical-ltd-make-all-ubuntu-related-services-reachable-over-ipv6
[14:27] <koopa58> @SpamapS Sorry was away for a while.. I have tried, but I can't find the entries that logwatch reports :)
[14:41] <koopa58> it's not in the site's error.log :S
[14:55] <RoyK> eagles0513875_: did it sort out?
[15:05] <stgraber> smw: hey there
[15:05] <stgraber> oops, wrong sm* sorry
[15:05] <stgraber> smb: hey there :)
[15:06] <smb> stgraber, Hi
[15:06] <stgraber> smb: I subscribed you to bug 930962 as it mentioned xen/virtio
[15:06] <smb> stgraber, Yep, noticed that. I think I have been there before but that feels long time ago
[15:07] <smb> stgraber, I might get back after todays attempted iso testing
[15:07] <stgraber> smb: sure. The bug has been around for a few years, so we can wait a bit longer ;)
[15:08] <smb> stgraber, :-P Yeah, can only get more-important... ;)
[15:33] <smoser> Daviey, jamespage utlemming. i will update the iso tracker results for the 20120627 run of cloud-images.
[15:33] <smoser> i have to go to lunch now, but will do so when i get back.
[15:33] <jamespage> smoser, marvellous!
[15:33] <smoser> had some issues with jibel's script immediately, though
[15:33] <smoser> https://code.launchpad.net/~jibel/+junk/qatracker
[15:34] <jibel> smoser, there was a change in the API, I just pushed rev5 that shoudl fix it
[15:34] <smoser> jibel, after fixing one global use of 'INSTANCE' (http://paste.ubuntu.com/1062710/)
[15:35] <jibel> smoser, fixed too
[15:35] <smoser> jibel, ok. http://paste.ubuntu.com/1062711/
[15:35] <smoser> htats what i get
[15:35] <smoser> if that isn't fixed, then i'll figure it out or bother you.
[15:35] <smoser> thanks
[15:37] <smoser> jibel, ok. so i still get the same error with revno 5
[15:37] <smoser> possibly bad usage
[15:37] <smoser> but i'll look later. have to run.
[15:39] <jibel> smoser, ./tracker_update_result -ad -u USERNAME -p APIKEY "Quantal Alpha 2" "Ubuntu Server EC2 EBS (Europe) amd64" "EC2 User Data" ami-a55f5bd1 Passed
[15:39] <jibel> don't forget the double-quotes
[15:40] <hallyn> stgraber: bug 1003656 - interesting, bad parsing of interfaces depending on indentation
[15:43] <stgraber> hallyn: I'll have a look later today. I'm going through all the network stack packages this week, bridge-utils is on the list for this afternoon :)
[15:45] <hallyn> stgraber: so then bug 231060 (sru of libvirt's /ec/dnsmasq.d file) is of interest to you this week?  :)
[15:45] <hallyn> the last comment in there makes me wonder if i s hould just pull that bit up to get the rest of the changes sru'd for now
[15:46] <stgraber> hallyn: dnsmasq isn't in that list (it's a universe package) :)
[15:46] <hallyn> but libvirt is
[15:47] <hallyn> all right, i had to try :)
[15:47] <stgraber> hallyn: hmm, did you see my discussion with RAOF wrt that change?
[15:48] <stgraber> hallyn: because IIRC we then agreed that the change was OK for SRU
[15:49] <hallyn> stgraber: ok, pinging him, thx
[15:50] <hallyn> otherwise long as it's fixed in q i'm fine dropping that bit from sru, i guess
[15:50] <stgraber> hallyn: http://paste.ubuntu.com/1062735/
[15:52] <hallyn> stgraber: thanks
[15:53] <stgraber> hallyn: we have the exact same fix in lxc and nobody complained. I can agree with the policy problem of removing the file on remove (instead of purge) but that's required to ensure dnsmasq's behaviour gets back to normal on removal.
[15:54] <stgraber> hallyn: but I'd certainly be fine adding a big warning in that file though (but if we do that, we should also include that warning in the next lxc SRU for consistency)
[15:56] <hallyn> stgraber: hm
[15:56] <hallyn> stgraber: add a warning saying "if you edit this, changes will be lost when libvirt is removed" ?
[15:57] <hallyn> stgraber: but if that file remains on remove, and is only removed on purge, nothing breaks, right?
[15:58] <hallyn> just if some joker then creates his own virbr0, it will continue to not be served by dnsmasq
[15:58] <stgraber> hallyn: nothing breaks but dnsmasq remains in bind-interfaces mode
[15:58] <hallyn> ah
[16:01] <hallyn> stgraber: all right, so, warning in dnsmasq.d/libvirt file saying "changes made to this file will be lost when libvirt package is removed" ?
[16:02] <hallyn> or ignore and resubmit as was?
[16:03] <stgraber> hallyn: I think the comment is a good idea in all cases
[16:05] <hallyn> ok thx
[17:20] <zul> Daviey: ping
[17:21] <souliaq> can I give "restart" permission for "non-sudoer" user?
[17:25] <greppy> souliaq: you could give it to just that one user using sudo.  it's in the man page.
[17:34] <zul> adam_g: ping
[17:35] <adam_g> zul: pong
[17:36] <zul> adam_g: im just upload a branch to openstack-ubuntu-testing for cinderclient should i call it master?
[17:39] <adam_g> zul: oh. hmm. im not sure how to name the 'master' branch. how about just openstack-ubuntu-testing/python-cinderclient/folsom, which will then merge into .../{precise, quantal}-folsom-proposed
[17:40] <zul> adam_g: sure
[17:50] <Daviey> zul: hey
[17:51] <zul> Daviey: so openstack clients version numbers are changes so im proposing we have something like 2012.2~<upstream version>
[17:56] <Daviey> oh ffs
[17:56] <Daviey> zul: not seen the mail yet, what is it changing to?
[17:56] <zul> Daviey: depends on the client
[17:56] <zul> i think
[17:56] <Daviey> uh
[17:57] <Daviey> zul: where did you hear this?
[17:57] <zul> adam_g
[17:57] <Daviey> adam_g: can you expand?
[17:58] <zul> Daviey: https://github.com/openstack/python-glanceclient/commit/03efd1689616ada606cb4cd7c0b51d2e1935ecfa
[17:58] <adam_g> Daviey: only that client (glanceclient, novaclient) builds started failing last night because there was a versioning change upstream. theres been a discussion on the list about this, but i didnt know anything was decided
[17:59] <zul> Daviey/adam_g: novaclient is 2.6.1.4 :)
[18:00] <adam_g> yea..
[18:01] <Daviey> adam_g: I haven't followed the discussion, but did it seem reasonable to move away from the current convention ?
[18:01] <Daviey> ttx: Are you tracking this?
[18:07] <Daviey> adam_g: what was the Subject?
[18:07] <Daviey> I haven't read about this.
[18:07] <adam_g> Daviey: https://lists.launchpad.net/openstack/msg13307.html
[18:07] <adam_g> Daviey: it seems reasonable given the client tools are (in theory) independent of the rest of the release.
[18:08] <Daviey> adam_g: well, the same could be said for mysql.. but i don't believe they ship independent versions
[18:08] <adam_g> unless im overlooking something, fooclient_2012.2-$version seems like a decent way of keeping the clients pegged to a release, at least on our end. that is,
[18:09] <adam_g> we'll at least have some sense of what our targetted (supported?) client version is for a given release
[18:09] <Daviey> well, we are now introducing arbitrary version numbering, if they insist on this seemingly silly change.
[18:10] <Daviey> Ie, we will be doing what they should be doing :)
[18:10] <adam_g> Daviey: arbitrary in what sense?
[18:11] <adam_g> i dont have a strong opinion either way, im just trying to avoid what i assume to be a huge headache (superceding a python-glanceclient_2012.2 version with 0.1.0.4)
[18:12] <Daviey> adam_g: I agree with you, but my limited knowledge on the reasoning for this change suggests that *clients might want to move to an independent release schedule ?
[18:18] <adam_g> Daviey: yeah, i suppose. AFAICS, for released versions of Ubuntu it doesn't really make a difference. you get the client version that made it into the release, presumably tested well against the rest of the openstack release we're shipping.
[18:18] <Daviey> It very much seems that i am in the markmc camp
[18:27] <kyle__> I'm trying to install and run ubuntu-orchestra on a server that through various dist-upgrades now thinks it's a desktop (ltsp is what caused it I think).
[18:28] <kyle__> Anyway, how does one go about starting orchestra?  I have a web server running, but don't see any orchestra component..
[18:36] <kyle__> window 2
[18:36]  * RoyK hands kyle__ a slash
[18:42] <kyle__> Thanks
[19:56] <kyle__> Anyone here familiar with ubuntu orchestra?  What did you have to install/do to get it working?  It appears the 'just apt-get install orchestra -y' that the tutorials are saying doesn't install everything in 12.04
[19:58] <Daviey> kyle__: Orchestra is pretty much deprecated in 12.04
[19:58]  * Daviey makes a note to remove the packaes from 13.10
[19:58] <Daviey> err, 12.10
[19:58] <kyle__> Daviey: Oooh.  Sooo, what's the official ubuntu mechenism for deployment en-masse then?
[20:00] <Daviey> kyle__: i see what you did there... MAAS it is called
[20:01] <Daviey> https://wiki.ubuntu.com/ServerTeam/MAAS
[20:01] <kyle__> OK, I thought that was still in development or beta or somefin.
[20:05] <kyle__> Humm
[20:05] <kyle__> Know if MAAS plays well with being on a server that thinks it's a desktop?
[20:07] <jpds> kyle__: It's all 1s and 0s at the end of the day.
[20:08]  * kyle__ nods
[20:10] <kyle__> jpds: Some packages are more fraigle and finicky than others.  Especially when they're special purpose, new, or not installed often.
[20:11] <Daviey> kyle__: The only issue i have come across is tftpd not working without a restart on Desktop
[20:11] <kyle__> OK.
[20:11] <kyle__> Easy enough.  Basically I'm going to set it up to install a batch of hadoop nodes, see if I like it.  And if I do, build it on a proper system for the larger cluster.
[20:11] <kyle__> This is really really temporary.
[20:12] <Daviey> kyle__: How many nodes?
[20:15] <kyle__> Daviey: Only 20.
[20:15] <kyle__> I figure it will give me a good flavor for if I want to use it for deployment.  Of course I was origonally thinking that of orchestra...
[20:24] <Daviey> kyle__: well, using MAAS and juju to deploy hadoop would work really well
[20:24] <Daviey> it's designed specifically for this use case :)
[20:28] <kyle__> :) sweet.
[20:36]  * kyle__ is restorting to apt-get purge.... this system is screwed up something fierce
[20:42] <alket> Hi , can I install ubuntu desktop but change it to server by removing GUI ?
[20:45] <kyle__> alket: I don't really thing you can in any reasonable way, but possible, probably... It would be easier to install minimal server and add a desktop to it than the other way around.
[20:55] <Daviey> kyle__: if the system is in that bad a state, i'd be tempted to re-install.
[20:56] <kyle__> Daviey: I am, but it's currently my DHCP server... may still be worth it.
[20:56] <kyle__> Especially since it appears MAAS wants to run it's own dhcp to be happy.
[20:57] <Daviey> kyle__: no, running it's own dhcp is optional.
[20:58] <kyle__> Daviey: Yea, but there are no example configs for that, nor do they explain really what the settings should be.
[20:58] <kyle__> At least not in the pages I found.
[21:00] <Daviey> kyle__: documentation is being improved, but really, if you already have dhcp/dns on a serpate machine.. and each node has a static hostname already set.. then maas will just slot in
[21:00] <Daviey> kyle__: note, the on-stage demo's have all been based on a SOHO router.
[21:00] <kyle__> Besides, all the hosts I'm running now have mac based reservations, I can use the pool addresses for MAAS, comment out the reservation in isc-dhcp, and then all should be happy.
[21:08] <kyle__> and hit my nodes with wol packets.....
[21:16] <kyle__> Humm.  How long after netbooting would you expect it to take for a node to be added?
[21:20] <kyle__> oooh.  Nevermind I see what's taking so long: it's installing ubuntu on each node before registering it.
[21:24] <Daviey> kyle__: no, it's abusing the installer to do initial enlistment
[21:25] <Daviey> doesn't touch the disk
[21:25] <Daviey> when i say abuse, that is a little extreme.. it's using the installer environment as an execution enviroment
[21:25] <kyle__> Humm.
[21:26] <kyle__> Daviey:  At what stage does maas let you configiure kickstart or other options for specifying what you want on a node?
[21:31] <Daviey> kyle__: use juju..
[21:32] <gelachs> hello dears, I have a problem installing ubuntu server on HP Prolian DL360p G8, the installed doesn't detect the disk drive and it asks for driver in a list. any help?
[21:33] <Daviey> kyle__: accept the node in the maas ui, it will then commission, briefly.. the grab the api key from the maas ui, and configure juju.. then juju bootstrap, then juju deploy hadoop
[21:33] <Daviey> kyle__: brief description there
[21:34] <kyle__> Hum.  OK.
[21:35] <Daviey> juju bootstrap ; juju deploy hadoop hadoop-master ; juju deploy -n17 hadoop hadoop-slavecluster ; juju add-relation hadoop-master:namenode hadoop-slavecluster:datanode ; juju add-relation hadoop-master:jobtracker hadoop-slavecluster:tasktracker ; you now have an 18 node hadoop cluster
[21:35] <Daviey> kyle__:
[21:36] <kyle__> Neat.
[21:36] <kyle__> Does MAAS allow for any specific configurations, or does it only do one stock base OS config on all systems?
[21:37] <Daviey> kyle__: Generally, it does stock.. it can be customised slightly, but not recommended
[21:37] <Daviey> kyle__: the idea is to get the os on there as quickly as possible, and customise on first boot
[21:37]  * kyle__ thinks his setup has some cruft, as at least one nodes was in a fully fledged install screen, wanting me to provide a username.
[21:37] <Daviey> kyle__: it makes use of cloud-init
[21:37] <kyle__> Ah.
[21:38] <Daviey> If you are being asked for a username, something has gone wrong
[21:38] <kyle__> Kindof what I thought.
[21:38] <Daviey> Anyway, i need to dash.. MAAS specific support can be had in #maas btw
[21:38] <Daviey> kyle__: hope it works out well.
[21:38] <kyle__> Awesome.  Thanks so much!
[21:48] <gelachs> please, do you see this message, I think I can't send to freenode. please if some one see this message tell me
[22:36] <daff> anyone know what the right way of configuring snmpd on ubuntu 12.04 is? the default configuration seems completely broken, making snmpd complain loudly about unknown OIDs. the symptoms are exactly the same as described in this 1.5 year old debian bug: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=605769
[23:12] <mdeslaur> daff: you need to install the snmp-mibs-downloader package, and then download the mibs, and then comment out the "#export MIBS=" line in /etc/default/snmpd
[23:13] <mdeslaur> daff: I think you have to comment out other mibs lines in the config files too
[23:13] <mdeslaur> daff: possibly other insanity too
[23:13] <daff> mdeslaur: thanks for the idea, but I already did that. still getting "net-snmp: 33 error(s) in config file(s)"
[23:13] <daff> insanity is what I am close to
[23:14] <mdeslaur> daff: wait one sec, let me try
[23:14] <daff> that would be great, thanks
[23:17] <kieppie1> hi everybody
[23:18] <kieppie1> I'm thinking of scratching my old VM host & looking towards the new MaaS infrastructure. Is anyone online able/willing to talk a bit of shop with me, please?
[23:20] <mdeslaur> daff: so, 1- apt-get install snmp-mibs-downloader, 2- comment out "export MIBS=" line in /etc/default/snmpd, 3- comment out "mibs :" in /etc/snmp/snmp.conf, 4- /etc/init.d/snmpd restart
[23:20] <patdk-lap> maas looks interesting
[23:21] <kieppie1> yea - it kinda encapsulates where I think I should be heading
[23:24] <kieppie1> I've built & run my old 10.04 LTS server with KVM & libvirt (mostly by hand), but after some reading (many threads leading me to the Canyon's Edge posts by kirkland ), I think I should be shifting direction - OpenStack on to op 12.04, KVM, libvirt with Juju, possibly Orgestra, Nagios, PXE, etc - all tighly integrated
[23:24] <patdk-lap> if I wasn't using esxi, I likely would go smartos
[23:27] <kieppie1> smartos looks interresting...
[23:27] <daff> mdeslaur: thanks, 3 was missing. but there are still hundres of lines worth of "/etc/snmp/snmpd.conf: line 101: Warning: Unknown token: disk." and "/var/lib/snmp/snmpd.conf: line 62: Warning: Unknown token: ifXTable."
[23:27] <daff> kieppie1: you could also look closer at puppet and foreman (theforeman.org)
[23:29] <kieppie1> yea - puppet forms a part of the new paradigm, as per kirkland's posts.
[23:31] <kieppie1> cthanks for the foreman lead - more to take into consideration
[23:33] <kieppie1> I think my setup may vary slightly though. I'll probably rely on my FreeNAS as the storage pool over iSCSI, and use my existing trusty pfSense as firewall. I'm not all that familiar with LDAP just yet, but I may need to dig into it in more detail to help manage resources
[23:38] <mdeslaur> daff: hrm, are you sure those aren't old lines? I just did the steps on a fresh precise VM, and I didn't get any errors
[23:41] <daff> mdeslaur: yeah, there must be something cached or old lying around
[23:51] <daff> mdeslaur: any ideas where I could look for further clues? I now have /etc/default/snmpd and /etc/snmp/*.conf identical to my home system, but I still get "net-snmp: 33 error(s) in config file(s)"
[23:52] <mdeslaur> daff: maybe one or more of the mibs files failed to download automatically when you installed snmp-mibs-downloader?
[23:52] <mdeslaur> daff: try "sudo download-mibs"
[23:53] <mdeslaur> daff: if that doesn,t solve it, I'm not quite sure what else you can try
[23:53] <daff> mdeslaur: did that repeatedly, the only thing that looks interesting is this:
[23:53] <daff> NOTE: PW-STD-MIB: ignored.
[23:53] <daff> gzip: /tmp/tmp.j0Vo8SsVkK/ianaiprouteprotocol-mib: unknown suffix -- ignored
[23:53] <daff> gzip: /tmp/tmp.j0Vo8SsVkK/ianamalloc-mib: unknown suffix -- ignored
[23:53] <daff> but I get that on all systems
[23:54] <mdeslaur> yeah, probably not relevant
[23:54] <mdeslaur> daff: sorry...I can't think of anything else to try for no
[23:54] <mdeslaur> now
[23:54] <daff> what I find strange is also the MIB search path. syslog says "MIB search path: /usr/share/snmp/mibs"
[23:55] <daff> but that directory doesn't exist, instead, shouldn't it be /usr/share/mibs?
[23:55] <mdeslaur> maybe compare those directories with your other system?
[23:56] <daff> and in addition to the errors also seen in the debian bug report, I get lots of these:  "Cannot find module (SNMPv2-SMI): At line 0 in (none)"
[23:56] <daff> can't make heads or tails of this
[23:56] <daff> the systems seem identical
[23:56] <daff> but obviously they aren't
[23:58] <Grenade> ubuntu 12.04 desktop with xen
[23:58] <Grenade> how to connect via ssh
[23:58] <Grenade> and configure xen
[23:59] <Grenade> anyone ?