[03:25] <test54321> has anyone tried to setup maas & juju with bonded nics?
[03:26] <sarnold> test54321: you may wish to try asking in #maas -- I think I heard something recently about dual-nic setups on maas being out-of-the-usual, but I don't think there was any discussion of bonded...
[03:27] <test54321> ok thanks!
[06:27] <jonascj> any libvirt users? What are the newest version of libvirt you've got on precise (12.04)?
[06:38] <crass> if there documentation on how to use fstab.d?
[08:08] <jibel> rbasak, Hey, I'm reviewing the lxc driver for autopkgtest but cannot seem to make it work from the source tree. How doyou invoke it usually?
[08:25] <jibel> rbasak, hm, it's because I used an ubuntu template for the container not a cloud template
[08:25] <jibel> rbasak, could we gracefully fail in that case ?
[08:55] <rbasak> jibel: that sounds reasonable. I could dpkg-query for cloud-init at the start I guess.
[09:33] <bariho> is there a way to prevent the display from switching off after some time when using the virtual consoles?
[09:41] <xnox> rbasak: jamespage: mysql--5.5 adt tests were failing before the borked 	5.5.32-0ubuntu2, and still fail after the revert in 	5.5.32-0ubuntu3. https://jenkins.qa.ubuntu.com/view/Saucy/view/AutoPkgTest/job/saucy-adt-mysql-5.5/
[09:42] <xnox> Completed: Failed 3/1713 tests, 99.82% were successful.
[09:42] <xnox> Failing test(s): main.symlink sys_vars.secure_file_priv_basic main.mysqltest
[09:42] <xnox> is anyone looking into fixing mysql adt test?
[09:47] <jibel> rbasak, can you remind me what was the requirement to use a cloud template instead of a debootrap based container?
[09:58] <jamespage> xnox: I think rbasak is
[09:59] <xnox> ok.
[10:31] <rbasak> xnox: do we know if the adt test has ever passed?
[10:32] <xnox> rbasak: history only goes back to build nr 39, all of which are red. Not sure if there are further logs. jibel do we have more adt history beyond what's visible in jenkins?
[10:32] <rbasak> Yeah that's what I was looking at.
[10:32] <rbasak> There's no "last successful build" noted in the status though.
[10:35] <rbasak> Looks like jenkins isn't blocking migration though.
[10:36] <rbasak> And that we need a no change rebuild for pinba-engine-mysql
[10:36] <jibel> xnox, no that's all we keep. I don't remember these tests ever passed
[10:42] <rbasak> jibel: thanks for the review!
[10:59] <jibel> xnox, I confirm that mysql tests never passed, neither on raring nor saucy
[11:02] <rbasak> Thanks
[11:02] <rbasak> We're running the same test suite in the build itself. So I'm not too worried about it.
[11:03] <rbasak> It should be fixed though of course
[11:18] <oz0ne> Hello, I wanted to add a notice under common error and fixes for others that might have been faighting the same issue as me with Postfix\opendkim\amavis. https://help.ubuntu.com/community/Postfix/DKIM (the issue is with enabling dkim signing like that guide if you allready have amavis doing spam checks is that the dkim will sign the header two times. one before amavis and then again one
[11:18] <oz0ne> more time after amavis) i wanted to put the note and how to correct the behaivor(eg. with adding -o no_milters when adding it back in postfix queue) under common errors and fixes. Reson for putting it under this article, cause many newcommers belives it is the issue of postfix\opendkim settings and not the way you feed messages back to queue from amavis.
[11:22] <rbasak> oz0ne: thanks for your input. The server guide is a volunteer effort and we appreciate all the help that we can get. Could you please file a bug against the serverguide project? Use this link: https://code.launchpad.net/~ubuntu-core-doc/ubuntu-docs/help.ubuntu.com
[11:23] <rbasak> If you are able, then a bzr merge proposal against the serverguide branch would be appreciated as well, those this is a bit more involved (editing the docbook xml source)
[11:23] <rbasak> Sorry, wrong link.
[11:24] <rbasak> THe bug reporting link is https://bugs.launchpad.net/serverguide/+filebug
[11:28] <oz0ne> i will file a bug and then i will check if i am capeable of doing a merge proposal. =)
[11:33] <rbasak> Thank you!
[11:52] <MACscr> how can i rsync files from one box to the other box, while preserving all ownership/perms, etc? Im trying to move LXC containers to a new ubuntu box
[11:55] <andol> MACscr: rsync --archive covers most of it. Depending on the situation you might also want --hard-links --xattrs and/or --acls
[11:55] <andol> MACscr: Depending on the situation, you may or may not want --numeric-ids
[11:55] <MACscr> andol: yeah, but since root is needed to create files other than itself and some owned by root, etc
[11:56] <MACscr> i think i might have to adjust sudoers or something
[11:58] <andol> MACscr: Or you could just allow yourself to use the root account for stuff like this? Otherwise you should be able to use the --rsync-path= option to include a sudo on the recieving end.
[12:37] <hallyn_> is there a kernel pkg that has a vmlinux.S?
[12:37] <hallyn_> gah
[12:37] <hallyn_> i'm apparently not awake yet.  sorry.
[12:41] <jamespage> rbasak, did you figure out the mysql test failure? I had a poke as well
[12:50] <jamespage> rbasak, --vardir=$WORKDIR/var/ -> --vardir=$WORKDIR/var (debian/tests/upstream)
[12:51] <jamespage> want me to upload?
[13:05] <rbasak> jamespage: wow. That's all? Please do.
[13:05] <rbasak> jamespage: I never looked. I was focused on getting it through proposed migration, and it wasn't blocked by that.
[13:06]  * rbasak wonders if anyone tests their uploads nowadays
[13:08] <jamespage> rbasak, OK _ I'll upload the fix
[13:09] <rbasak> Thanks!
[13:10] <rbasak> BTW, I uploaded a no change rebuild for pinba-engine-mysql which was blocking migration, and then it went through. So hopefully this upload will go straight through,.
[13:31] <jodh> rbasak/jamespage: hi guys - it appears mysql's failing ADT tests are blocking the latest upstart upload. Are you working on this failure?
[13:31] <jamespage> jodh, just uploaded a fix ^^
[13:31] <jodh> jamespage: any idea on turnaround?
[13:32] <jamespage> jodh, it takes about ~3 hours to build normally
[13:32] <jamespage> jodh, if its urgent please ask the release team to override - they can
[13:43] <phunyguy> quick question... I have some LVM volumes that were temporarily in the system, but have since been removed without restart (USB), but ubuntu keeps thinking it is still there... aside from a reboot, how can I force-remove it?
[13:55] <RoyK> philipballew: were the filesystem(s) mounted when you removed them?
[13:56] <RoyK> phunyguy: that was for you...
[13:56] <RoyK> phunyguy: echo "- - -" > /sys/class/scsi_host/host0/scan
[13:56] <RoyK> just choose the right host
[13:56] <phunyguy> RoyK: they were unmounted
[13:56] <phunyguy> define "host" in this context?
[13:57] <phunyguy> :P
[13:57] <RoyK> host0 is the first
[13:57] <phunyguy> right, but there are 8... how do I know which one is correct?
[13:58] <RoyK> phunyguy: perhaps in /dev/disk/by-id ?
[13:58]  * RoyK a bit unsure
[13:59] <RoyK> anyway - it won't hurt to run a rescan on another scsi host
[13:59] <phunyguy> so just hit em all
[13:59] <phunyguy> roger.
[14:00] <phunyguy> that worked!
[14:00] <phunyguy> GENIUS!
[14:00] <phunyguy> owe you beer.
[14:04] <RoyK> :)
[14:12] <diegonat> hi guys, i need a tip! Im going to develop a piece of software with a couple of friends so i was wondering what kind of resource i would need to have to keep in touch. Like a forum? or something else?
[14:13] <jkitchen> diegonat: with each other? github and an irc channel?
[14:13] <jkitchen> trello board?
[14:13] <Katafalkas> trello
[14:13] <diegonat> trello seems cool
[14:13] <jkitchen> trello is awesome
[14:13] <Katafalkas> its more then cool
[14:16] <diegonat> can u chat on trello
[14:16] <diegonat> ?
[14:17] <Katafalkas> sort of
[14:17] <Katafalkas> its real time
[14:17] <Katafalkas> so yeah
[14:50] <adam_g> jamespage, any idea what might have changed in the virt stack that would cause nova-compute to no longer pull in qemu-kvm?
[14:51] <jamespage> adam_g, on saucy?
[14:51] <adam_g> jamespage, about to test on saucy, but using havana-staging atm
[14:52] <jamespage> adam_g, qemu-kvm is a transitional package for saucy
[14:52] <jamespage> qemu-system-x86 is the one for saucy/havana-staging
[14:55] <adam_g> jamespage, which is responsible for actually getting /dev/kvm setup now?
[14:55] <jamespage> erm
[14:55] <adam_g> at the end of the day 'apt-get install nova-compute-kvm' has no kvm module loaded
[14:56] <jamespage> hallyn_, ^^
[14:56] <smb> Its automatically created when the kernel modules load
[14:56] <smb> and those load based on cpu type
[14:56] <adam_g> smb, right, but what package is modprobe'ing ?
[14:56] <smb> The kernel does
 at the end of the day 'apt-get install nova-compute-kvm' has no kvm module loaded
[14:56] <smb> well udev
[14:56] <jamespage> where are you doing that?
[14:57] <adam_g> jamespage, the jenkins lab. waiting for a qemu backport build
[14:57] <jamespage> adam_g, unsure then
[14:57] <jamespage> I'd defer to hallyn on that
[14:57] <smb> adam_g, check modinfo kvm-intel for example. That has an alias section in S
[14:57] <hallyn_> adam_g: yeah, qemu's postinst no longer modprobes kvm
[14:57] <adam_g> jamespage, but ive tried elsewhere from the H ppa
[14:58] <hallyn_> udev does it for us without qemu being installed
[14:58] <jamespage> adam_g, actually that might be it
[14:58] <hallyn_> adam_g: are you running linux-image-generic?
[14:58] <jamespage> hallyn_, is that udev in saucy doing that? does precise run the same way?
[14:58] <adam_g> hallyn_, yea.. the kvm_intel.ko is available
[14:58] <hallyn_> jamespage: i think it does, but i don't recall for sure when that was introduced.  smb?
[14:59] <hallyn_> adam_g: available, but not modprobed.  what does kvm-ok say?
[14:59] <smb> hallyn_, I would need to check git but in Saucy at least
[14:59] <smb> Maybe before
[14:59] <hallyn_> adam_g: removing the 'modprboe kvm_intel' from qemu.postinst was deliberate...
[14:59] <adam_g> hallyn_, it tells me that something is wrong. cpu supports it, but module is not loaded
[14:59] <hallyn_> adam_g: ok but loading the module by hand works ok?
[14:59] <adam_g> hallyn_, yup
[15:00] <hallyn_> ok.  then i guess we have a packaging issue.
[15:00] <adam_g> hallyn_, AFAICS everything is where it needs except the the module being loaded
[15:00] <hallyn_> adam_g: do you have
[15:00] <hallyn_> /lib/udev/rules.d/40-qemu-system-common.rules:KERNEL=="kvm", GROUP="kvm", MODE="0660"
[15:00] <adam_g> hallyn_, yeah
[15:01] <hallyn_> huh.  then why didn't it run?
[15:01] <hallyn_> oh i'm not thining right
[15:01] <adam_g> hallyn_, is that responsible for loading the module itself, or just setting the permission on /dev/kvm when it shows up?
[15:02] <hallyn_> adam_g: yeah the latter
[15:03] <smb> hallyn_, So that was introduced with the 3.5 kernel
[15:03] <smb> Which means Q, R and S
[15:04] <adam_g> ok, so the issue should be limited to running this on precise?  booting a saucy machine now
[15:04] <hallyn_> adam_g: but shoudldn't your cloud-wahtever ppa be depending on a newer kernel than 3.2?
[15:05] <adam_g> hallyn_, no, we haven't done any kernel backports
[15:05] <hallyn_> bleh
[15:06] <hallyn_> adam_g: how big a deal is it to add a delta to the backported qemu pkg?
[15:06] <adam_g> hallyn_, its not that we've had to do before, but i always assumed we would need to eventually
[15:07] <adam_g> *not something
[15:07] <hallyn_> adam_g: ok, well we *could* put the modprobing logic back int at saucy as well.  the problem was that in situations like containers the modprobe could fail, and it all just got too ugly to bother since we expect it to be loaded already
[15:08] <adam_g> what specifically is introduced in 3.5 that all of this depends on?
[15:09] <hallyn_> smb: ^ how is the kvm autoloading done?
[15:10] <smb> hallyn_, The modalias that triggers on cpu type and features
[15:10] <smb> alias:          x86cpu:vendor:*:family:*:model:*:feature:*0085*
[15:10] <smb> the amd module has a similar one for svx
[15:13] <smb> hallyn_, When the cpu is registered it shows up on cpu bus and emits a device add uevent. from there normal magic matches the alias to a module to probe
[15:14] <adam_g> smb, thanks
[15:15] <hallyn_> adam_g: what's your plan stan?
[15:15] <adam_g> hallyn_, trying to capture the issue in a bug
[15:15] <hallyn_> ok
[15:16] <adam_g> hallyn_, if its between backporting the kernel, or a small qemu delta, i vote for the latter
[15:16] <adam_g> i assume it would be a small delta?
[15:21] <hallyn_> adam_g: yup
[15:21] <hallyn_> to qemu-system-common.qemu-kvm.upstart or somesuch
[15:21] <adam_g> hallyn_, also, when checking saucy i noticed permissions: crw------- 1 root root 10, 232 Aug 29 15:09 /dev/kvm
[15:22] <hallyn_> adam_g: that's without qemu-system-x86 installed?
[15:23] <hallyn_> kvm group doesn't exist at that point, so that would be correct
[15:23] <adam_g> hallyn_, oh duh. ya
[15:23] <adam_g> without that missing udev rule
[15:33] <prime> I have a storage array I am trying to install ubuntu 12.04.3 precise pangolin LTS. I've used the Universal-USB-Installer which seems slightly outdated. The boot menu had a hard time finding the installation files. Any tips?
[15:35] <prime> I'm using a usb thumb drive. This time I'll format it prior to copying the slightly newer 12.04.03 precise pangolin lts as Unlisted rather than the known 12.04.02 server.
[16:57] <oz0ne> rbasak: i registred the bug, but now i do see I can actually do the change directly on the wiki page. It just took some time after registring as a part of the team until i could actually edit the page. This is not directly related to the server-guide but rater the wiki page. Tho if you do follow the instructions, install amavis opendkim and postfix. and you decide to enable signing you
[16:57] <oz0ne> will most likly get this double sign behavior as this line will make the milts run two times. http://bazaar.launchpad.net/~ubuntu-core-doc/serverguide/raring/view/head:/serverguide/C/mail.xml#L1616
[17:04] <rbasak> oz0ne: which wiki page?
[17:04] <oz0ne> https://help.ubuntu.com/community/Postfix/DKIM
[17:04] <rbasak> Ah, right.
[17:05] <rbasak> oz0ne: sorry, I was mistaken. You're right - you can just edit that. Please do!
[17:06] <oz0ne> rbasak: https://bugs.launchpad.net/serverguide/+bug/1218335 read my correction: it kinda also apply to the server guide, if you stilla gree will i make a patch\merge request.
[17:07] <rbasak> oz0ne: I see. Yes, that makes sense. Please go ahead with the merge request.
[17:07] <rbasak> (I'm not sure about the issue itself - just that you're filing the bug, editing the wiki, submitting the merge proposal in the right place, etc)
[17:07] <koolhead17> adam_g: ping
[17:20] <adam_g> koolhead17, hey
[17:20] <koolhead17> adam_g: will you be in town on 6th?
[17:20] <adam_g> koolhead17, should be
[17:21] <koolhead17> cool. lets catch up for coffee/breakfast then
[17:21] <koolhead17> DM me your number
[17:47] <adam_g> jamespage,  thoughts on carrying a delta re:  https://bugs.launchpad.net/bugs/1218459 ?
[17:48] <jamespage> adam_g, +1
[17:49] <adam_g> jamespage, where should track it? should we keep stacked branches in lp:~ubuntu-cloud-archive repositories for these types of packages?
[17:58] <arrrghhh> hello all.  I have mdadm running a raid1 array on two 1tb hdd's.  I kind of want more speed, as I have my source for Android on this array... I don't so much care if I lose the data.  Can I convert it to raid0 "in place" or do I have to migrate the data to another disk/array, and format/rebuild the array new as raid0?
[18:00] <rbasak> I'm not aware of any way to convert md in place, but perhaps I'm missing latest developments there.
[18:00] <arrrghhh> yea, based on my knowledge/what I've read I have to rebuild the array
[18:00] <sarnold> throwing it on an ssd would probably give you better performance improvement for less effort.
[18:01] <arrrghhh> I'm wondering if it'll help my build times that much even... hum.  Just annoying.  Rig used to build in about 20 minutes, and now it takes nearly 40
[18:01] <arrrghhh> ssd is there, it is used for ccache and out/
[18:01] <arrrghhh> I guess I could get another and put the source on it
[18:03] <RoyK> arrrghhh: ccache?
[18:04] <arrrghhh> RoyK, my rig builds Android, among other things.
[18:04] <RoyK> arrrghhh: iirc you can convert it to raid0, but I haven't tried it
[18:05] <arrrghhh> based on my readings, I will have to backup the data if I want to preserve it... which I'd rather not redownload the source if I can help it.
[18:07] <RoyK> I have a vm with 13.04 (or was that 13.10)
[18:07] <RoyK> my raidtest vm
[18:07] <arrrghhh> nice
[18:10] <RoyK> arrrghhh: ok, so two drives in raid1 and you ant to split and merge to raid0?
[18:11] <arrrghhh> yes, basically.  space is not a concern, I just want moar speed.
[18:11] <arrrghhh> which another ssd would be better... but I am cheap :)
[18:14] <RoyK> arrrghhh: hrmf - seems I bricked that vm somehow
[18:17] <RoyK> arrrghhh: give me 10 minutes
[18:21] <RoyK> arrrghhh: shouldn't take too long to setup a new vm :P
[18:22] <RoyK> arrrghhh: there's something about zfs on a few drives and an ssd for caching ;)
[18:34] <RoyK> arrrghhh: took a bit more time to reinstall that vm - 20 mins
[18:39] <RoyK> arrrghhh: still waiting for resync on that mirror
[18:42] <RoyK> arrrghhh: still here?
[18:55] <cyphermox> utlemming: ping?
[18:56] <utlemming> cybermox: pong
[18:57] <utlemming> cyphermox: pong
[18:57] <cyphermox> utlemming: hehe. this is about https://code.launchpad.net/~med/ubuntu-seeds/sosreport/+merge/179485
[18:58] <utlemming> cyphermox: yeah, I have spoken to med about that
[18:59] <utlemming> med: ^
[18:59] <utlemming> cyphermox: I had suggested to med that it being in the ubuntu-server seed might be more appopriate, given that it is useful on servers, not just cloud instances
[18:59] <cyphermox> ok to add to the cloud-image seed? I think that's all that's left... though I'd be more conforable for one of you guys to handle it rather than I change the seed :)
[19:00] <utlemming> smoser: ping
[19:00] <cyphermox> sure
[19:00] <cyphermox> server is a good idea as well
[19:00] <utlemming> cyphermox: I have no objections, I just want to check with smoser, et al of the server team first
[19:00] <cyphermox> np
[19:01] <utlemming> smoser: sosreport to ubuntu-server or cloud-image seed?
[19:01] <smoser> yeh
[19:01] <smoser> what is the added delta ?
[19:01] <smoser> i have only objected in the past based on python2.
[19:04] <utlemming> smoser: sosreport does not include any new dependency in the images
[19:05] <utlemming> smoser: in checking on a fresh saucy, its just an extra package
[19:05] <smoser> is there an mp ?
[19:05] <smoser> i'l just do it now if there is.
[19:06] <utlemming> smoser: for cloud-images, https://code.launchpad.net/~med/ubuntu-seeds/sosreport/+merge/179485
[19:06] <utlemming> smoser: I think there is one for ubuntu-server, but I need to check
[19:06] <utlemming> smoser: I would rather see it in server, though
[19:07] <utlemming> smoser: there isn't one for server
[19:08] <clusty> hey
[19:09] <smoser> there is a bug on this.
[19:09] <smoser> https://bugs.launchpad.net/ubuntu/+source/sosreport/+bug/1206106
[19:10] <clusty> last upgrade hung so i had to close the console. now every time i run "sudo dpkg --configure -a" as suggested it still hangs at "Setting up apache2-mpm-worker (2.2.22-6ubuntu5.1)"
[19:10] <clusty> clues how I could flush all apaches so i can restart ?
[19:12] <smoser> utlemming, i just pushed cloud-utils -> cloud-guest-utils in images
[19:12] <smoser> in cloud seed
[19:12] <utlemming> smoser: ack
[19:14] <clusty> anybody ?
[19:20] <arrrghhh> RoyK, sorry I was at work.. then lunch.
[19:20] <arrrghhh> now back to work, but I'll be semi-here
[19:20] <RoyK> arrrghhh: I quit testing - don't know
[19:20] <arrrghhh> hah np
[19:42] <med_> smoser, utlemming cyphermox : catching up on sosreport...
[19:42] <med_> it's already in Main per that MIR smoser found
[19:42] <cyphermox> yes
[19:42] <med_> and it is still python 2.7 based though is scheduled to be reworked ahead of 14.04
[19:42] <cyphermox> med_: I'm going through the sponsoring overview, the seed change was up there
[19:42] <med_> nod.
[19:42] <cyphermox> I just wanted to get an ack from people before changing seeds I don't normally touch
[19:43] <med_> nod.
[19:43] <smoser> cyphermox, you can have ack from me.
[19:43] <smoser> i was about to do it.
[19:43] <smoser> you can do i.
[19:43] <smoser> it.
[19:43] <cyphermox> can you?
[19:43] <cyphermox> I'm beyond busy with Touch stuff :)
[20:22] <LargePrime> I cant find a google answer to this.
[20:22] <LargePrime> What is the server load number mean?
[20:48] <genii> LargePrime: The wikipedia page covers it fairly well. http://en.wikipedia.org/wiki/Load_%28computing%29
[20:48] <LargePrime> genii: Thanks so much
[21:08] <LargePrime> i have a problem  did an apt-get upgrade and i am getting bitched at cause  /boot/initrd.img-3.2.0-40-generic updated?
[21:12] <LargePrime> Configuration file `/etc/issue'
[21:12] <LargePrime>  ==> Modified (by you or by a script) since installation.
[21:12] <LargePrime>  ==> Package distributor has shipped an updated version.
[21:12] <LargePrime> what should i choose?
[21:12] <LargePrime> sorry for the noob question
[21:19] <sarnold> LargePrime: probably doesn't matter -- it just shows a login banner to the login prompt on virtual consoles.
[21:19] <sarnold> LargePrime: if you accept the package version, you probably won't be bothered in the future. :)
[22:28] <mgw> so I have what may be an odd question. What would be the performance implications of an lvm over a loopback inside a qcow image?
[22:32] <sarnold> it probably won't be fast, but you knew that right? :)
[22:40] <MACscr> I want to run ubuntu server from a usb thumb drive, but more importantly, i want it to mostly be running in just a ram disk for speed and to minimize writes to the usb drive. I do though still want some persistency. Any suggestions on making that happen? Its going to be for a storage node.
[22:41] <mgw> sarnold: I assumed that, yes… here's what I'm trying to accomplish: I want to run lxc containers inside a host which is a kvm guest.
[22:46] <sarnold> mgw: should be alright, people use nested kvm for testing / development work all the time. I don't know if I'd want to rely on that for deployment, but lxc-in-kvm ought to be less painful than that :)
[22:46] <sarnold> MACscr: look into "laptop mode"
[22:49] <MACscr> sarnold: i think thats power power related and since this is a server, its going to be running 24/7
[22:50] <sarnold> MACscr: limiting the number of disk writes feels like a useful thing to aim for however