[14:23] <Daviey> jamespage: who do you need at Ceph review?
[14:24] <jamespage> Daviey, hmm - don't have any upstream attendance
[14:24] <Daviey> jamespage: currently you, and I
[14:24] <jamespage> sounds good - if anyone else wants to jump in then +1
[14:33] <med_> I was planning to watch.
[14:33] <med_> maybe ivoks ? or dosaboy?
[14:34] <med_> I think dosaboy wants an invite to HA later.
[14:44] <med_> Join ~ubuntu-etherpad in your launchpad id to have access to etherpad
[14:56] <smoser_> 'oomselfmg' was one of the words in my captcha
[14:59] <med_> still pending Daviey jamespage
[14:59] <jamespage> o/
[14:59] <med_> "live event will begin in a few moments"
[14:59] <med_> make the on air live at some point
[14:59] <marrusl-uds> "the conference will begin when the leader arrives"
[15:00]  * nwl waves
[15:00] <Daviey> jamespage: ?
[15:01] <Daviey> any other fishbowl people needing to be in te inner circle
[15:01] <med_> Daviey, still not getting a stream
[15:01] <med_> on the ceph page
[15:01] <med_> jamespage, ^
[15:02] <Daviey> 2 moments
[15:02] <nwl> rturk: morning
[15:02] <med_> put SpamapS in the fishbowl
[15:02] <rturk> nwl: howdy
[15:02] <med_> are there any inktankers here?
[15:02] <SpamapS> I don't have much to say :)
[15:02] <med_> heh.
[15:02] <SpamapS> and I don't kow where the video stream is
[15:02] <nwl> med_: me and rturk
[15:03] <med_> Daviey, you can start the livestream even if the meeting isn't quite ready to go.
[15:03] <med_> unless there are tech issues of course....
[15:03] <SpamapS> inktankers are likely still rolling out of bed this early
[15:03] <SpamapS> (might have made more sense to put this one later in the day)
[15:03]  * rturk- yawn\s
[15:03] <Daviey> SpamapS: possibly :(
[15:03] <rturk-> :)
[15:04] <nwl> we're made of hardy stuff at inktank
[15:04] <med_> jamespage, Daviey : dosaboy needs the hagnout
[15:04] <smoser> anyone here able to say "yes, stream is working" ?
[15:04] <Daviey> Helo?
[15:04] <dweaver> Nothing here
[15:04] <med_> nope
[15:05] <marrusl-uds> working here
[15:05] <med_> is now
[15:05] <med_> dosaboy would like the Hangout URL
[15:05] <med_> cna someone PM him
[15:05]  * nwl just stares at static
[15:05] <dweaver> refresh the page if it is not working for you
[15:05] <med_> yes dosaboy
[15:05] <med_> nwl
[15:05] <med_> rturk
[15:05] <dduffey_> yes, shift-reload did work for me
[15:05] <med_> dosaboy wants a link
[15:06] <rturk> I have video now
[15:06] <dosaboy> med_: got it, thanks
[15:06] <kentb> me too
[15:06] <med_> nice lounge jamespage
[15:07] <nealmcb> I just got the video up via a page reload - it was just saying that it would be starting soon, but no indication that I'd need to refresh it.
[15:09] <nwl> this is the first UDS session I have attended in my dressing gown. feels liberating.
[15:09] <med_> nwl, TMI. :^)
[15:12] <jamespage> https://blueprints.launchpad.net/ubuntu/+spec/servercloud-r-ceph
[15:14] <Daviey> nealmcb: feedback taken, i'll pass it on
[15:15] <nwl> jamespage: longer term i am hoping we move off fastcgi
[15:16] <nwl> i am the product manager at Inktank :-)
[15:16] <nwl> and ex-canonical-er
[15:16] <nwl> otherwise known as Neil Levine
[15:16] <med_> Neil Levine
[15:16] <rturk> :)
[15:16] <med_> :)
[15:16] <nwl> because i am sitting in  my pyjamas
[15:16] <med_> I tried to tell you guys earlier to add the inktakners
[15:16] <rturk> you don't want to look at me now
[15:17]  * med_ turns off his camera when he joins
[15:17] <nwl> trying that link
[15:17] <nwl> 2 secs
[15:17] <med_> yes sound good
[15:17] <med_> yep, blacked out
[15:17] <dweaver> lol
[15:18] <med_> automatically push CEPH into older versions of ubuntu PRECISE LTS
[15:18] <med_> MRE is different than SRU
[15:18] <med_> Minor Release Exception is like an Auto SRU.
[15:22] <nealmcb>  I'm lurking, trying to add a bit of info to help lurkers. Failing.... Can't even find a web link to put in the public pad to define what Ceph Bobtail is.... Seems to be the second release of Ceph, but I don't see where they ever say so....
[15:23] <Daviey> nealmcb: footer of http://summit.ubuntu.com/uds-1303/meeting/21652/ceph-review/
[15:23] <Daviey> http://pad.ubuntu.com/uds-1303-ceph-review
[15:24] <dosaboy> Ceph Bobtail Info - http://ceph.com/releases/v0-56-released/
[15:24] <rturk> or: http://ceph.com/docs/master/release-notes/#v0-56-1-bobtail
[15:25] <rturk> btw, SSO on the pad doesn't work for me :-/
[15:25] <nwl> me neither - get into an endless openid loop
[15:25] <nealmcb> thanks folks.  Google would figure it out also if someone put that on the ceph home page or somewhere close that that.  Or noted it in wikipedia.  Cheers!
[15:25] <nealmcb> Daviey: thanks - perhaps
[15:26] <nealmcb> ... it would help if the video was ready to go at the start of each session, but I don't know
[15:26] <rturk> http://ceph.com/dev-notes/deploying-ceph-with-juju/
[15:26] <rturk> Patrick McGarry -> scuttlemonkey on irc
[15:27] <med_> tx rturk
[15:30] <nwl> stream just crashed on me
[15:31] <med_> nwl, can you point to ... what was that you wanted Daviey
[15:31] <med_> actually, whole stream just froze for me too
[15:31] <dosaboy> hangout died?
[15:31] <Daviey> med_: the supported timefrae for each release
[15:31] <yolanda_> for me too
[15:31]  * dosaboy rejoins
[15:32] <med_> It's Dead Jim.
[15:32]  * rturk refreshes the stream too
[15:32] <Daviey> can anyone see us?
[15:32] <med_> nyet, youtube stream froze
[15:32] <dosaboy> "The hangout ended because of a server error"
[15:32] <nwl> can't get back into the stream
[15:32] <med_> blackened
[15:32] <nwl> "hangout ended because of a server error"
[15:32] <nwl> and justice for all
[15:32] <med_> you may want to close that stream and either call it or just start another one
[15:32] <jamespage> yeah - going for that now
[15:32] <med_> put a link in the ether pad
[15:33] <jamespage> please hold the line callers
[15:33] <med_> for the new youtube stream
[15:33] <rturk> pls also paste the new link here, the pad doesn't work for me and nwl
[15:33] <med_> nod.
[15:33] <med_> rturk, will do
[15:33] <jamespage> pad: http://pad.ubuntu.com/uds-1303-ceph-review
[15:34] <med_> jamespage, I suspect they need to be added to the launchpad group
[15:34] <Daviey> apparently i have 3 viewers
[15:34] <jamespage> ah - yes
[15:34] <med_> Join ~ubuntu-etherpad in your launchpad id to have access to etherpad
[15:34] <med_> nwl, rturk ^
[15:34] <nwl> med_: i can't access the etherpad. get into an openid loop
[15:34] <med_> nwl, no worries
[15:35] <med_> so the original youtube stream is in post processing
[15:35] <med_> You will be forced to create a new hangout
[15:35] <med_> I assume you're on that Daviey
[15:35]  * med_ goes for more coffee
[15:35] <rturk> joined ~ubuntu-etherpad, no luck yet
[15:36] <rturk> ah! there we go.  got the pad now
[15:36] <nwl> rturk: how did you do that?
[15:36] <Daviey> http://youtu.be/jGxNSNLIL-4
[15:37] <nwl> i get "Either you have not been granted access to this resource or your entitlement has timed out. Please try again."
[15:37] <dweaver> I'm back on now
[15:38] <dweaver> Yes I can
[15:38] <marrusl-uds> yes.  works for me.
[15:39] <med_> what page do you want us to refresh?
[15:39] <udsbotu> uds-servercloud-1: 5 minutes left in this session!
[15:40] <med_> http://www.youtube.com/watch?v=jGxNSNLIL-4&feature=player_embedded
[15:40] <adam_g> question for when things settle: if we're backporting ceph to the cloud archive, which cloud archive pocket do new backports go? a requirement that X ceph release is shipped /w Y openstack release?
[15:40] <udsbotu> uds-servercloud-1: 4 minutes left in this session!
[15:41] <adam_g> so an upgrade between openstack release will also require a ceph upgrade?
[15:41] <udsbotu> uds-servercloud-1: 3 minutes left in this session!
[15:42] <Daviey> adam_g: yes
[15:42] <Daviey> adam_g: it's the tested stack
[15:43] <med_> Where is CEPH required in an OpenStack cloud archive (LTS) based deployment?
[15:44] <med_> jamespage, ^ Daviey ^
[15:44] <med_> good.
[15:45] <med_> "incredibly useful" is what I want.
[15:45] <udsbotu> uds-servercloud-1: This session has ended.
[15:45] <med_> yep, you go withy custom
[15:46] <med_> lots of cinder plugins
[15:46] <med_> thanks guys
[15:46] <rturk> thanks, all :)
[15:46] <dweaver> Thanks all, where's the catering!!
[15:46] <Daviey> thanks!
[15:46] <med_> thanks for managing through diversity Daviey
[15:47] <med_> dweaver, snacks in the Blue Ale room
[15:47] <Daviey> nealmcb: Thanks for your help
[15:47] <med_> Daviey, smoking out behind the barn.
[15:47]  * Daviey is a non-smoker :)
[15:47] <med_> j/k
[15:48] <med_> many thanks to inktank/CEPH upstream
[15:55] <med_> Daviey, arosoales, utlemming :i'm looking for hangout invite for this one.
[15:56] <Daviey> med_: sure
[15:57] <Daviey> Anyone else want to be in the fish bowel for this?
[15:57] <Daviey> Cloud archive review
[15:57] <jamespage> Daviey, yes
[15:58] <Daviey> jamespage: you should already have the URL :)
[15:58] <jamespage> med: "Managing through Diversity" sounds like a book title - when are you going to write it
[16:00] <Daviey> Starting shortly
[16:00] <med_> jamespage, it's a pretty standard MBA term
[16:00] <Daviey> Anyone else want to attend?
[16:01] <Daviey> yolanda_: ?
[16:01] <Daviey> smoser: ^?
[16:01] <Daviey> roaksoax: ^?
[16:01] <smoser> Daviey, i'm in #ubuntu-uds-community-1 (rolling release)
[16:02] <Daviey> ok, cool
[16:02] <yolanda_> Daviey, is there any hangout link? i'm playing the video
[16:02] <Daviey> starting
[16:03] <Daviey> please refresh those attending remotely
[16:08] <caribou_> jamespage: So the standard SRU process does apply to u-c-a ?
[16:09] <caribou_> jamespage: thanks
[16:12] <dosaboy> jamespage: is this the correct place for Openstack SRU - https://wiki.ubuntu.com/ServerTeam/OpenStack
[16:12] <jamespage> http://reqorts.qa.ubuntu.com/reports/ubuntu-server/cloud-archive/folsom_versions.html
[16:12] <Daviey> dosaboy: can you expand your question ?
[16:13] <dosaboy> Daviey: just curious as to how the openstack-ubuntu-testing project fits in with the SRU process
[16:15] <melmoth> yep
[16:20] <melmoth> hmm
[16:20] <dosaboy> Daviey: still live for me
[16:20] <melmoth> does the video works ?
[16:20] <jamespage> guess you guys lost video?
[16:21] <jamespage> staring a new hangout
[16:21] <melmoth> yep
[16:21] <melmoth> the live recordig blablabla... is being processed
[16:21] <med_> shortly melmoth
[16:21] <med_> just a sec
[16:21] <melmoth> no problemo :)
[16:21] <dedalus> no longer live, showing recorded session now
[16:22] <Daviey> reload
[16:22] <dedalus> now back to "live event will begin in a few moments"
[16:23] <dedalus> ok, back
[16:24] <med_> http://www.youtube.com/watch?v=f3QosKQLuTU&feature=player_embedded
[16:24] <med_> melmoth, reload or http://www.youtube.com/watch?v=f3QosKQLuTU&feature=player_embedded
[16:26] <mdeslaur> Daviey: yo
[16:27] <Daviey> hey
[16:27] <mdeslaur> \o
[16:27] <Daviey> mdeslaur: you have a pm
[16:30] <jamespage> mdeslaur, http://reqorts.qa.ubuntu.com/reports/ubuntu-server/cloud-archive/folsom_versions.html
[16:30] <jamespage> for example
[16:37] <Daviey> BD - Build depend
[16:41] <Daviey> I wonder if we should talk about.. armhf :)
[16:46] <Daviey> zul: since you joined, you brought noise with you
[16:46] <med_> vacuum cleaner noise
[16:46] <zul> im muted
[16:47] <med_> that fixed it.
[16:47] <Daviey> zul: try again
[16:47] <zul> hold on
[16:47] <Daviey> perfect
[16:47] <Daviey> :/
[16:48] <med_> g+ mute should work (vs h/w muting the wrong h/w)
[16:49] <zul> lets try one more time
[16:49] <med_> better
[16:49] <med_> no noise
[16:50] <zul> figured it out
[16:50] <zul> i had the radio on
[16:52] <bkerensa> ciao
[16:54] <med_> Daviey, jamespage : wrap up.
[18:10] <adam_g> is there a G+ URL ?
[18:11] <roadmr> adam_g: see here: http://summit.ubuntu.com/uds-1303/meeting/21653/openstack-infrastructure-ha-review/
[18:11] <roadmr> adam_g: the youtube thingy on the left should mutate once the session is live, you can click on the title to access the hangout
[18:12] <adam_g> roadmr, i meant a URL to actually join the hangout. :)
[18:12] <adam_g> jamespage, ^
[18:12] <jamespage> adam_g, Daviey dishes those out
[18:13] <roadmr> oh fall asleep sounds good
[18:14] <danjared> roadmr: I thought this talk is going to be about high availability, not knocking out :)
[18:15] <roadmr> danjared: heh :) so how do you h/a when a node falls asleep?
[18:15] <jamespage> o/
[18:15] <jamespage> should start in a few seconds
[18:16] <danjared> roadmr: more coffee :)
[18:16] <jamespage> refresh now for video
[18:16] <roaksoax> https://blueprints.launchpad.net/ubuntu/+spec/servercloud-r-openstack-ha
[18:17] <zul> im around if you guys need me
[18:18] <Daviey> zul: jump in
[18:23] <Daviey> jamespage: STONITH should also help avoid split brain?
[18:23] <Daviey> Do we carry any risk of split brain?
[18:24] <danwest> well the risk of split brain is more via a third or odd numbered vote
[18:25] <danwest> else you can get each node trying to  STONITH each other
[18:25] <danwest> hence, how do we handle the 3rd voting member?
[18:26] <SpamapS> danwest: first stonith wins :)
[18:26] <jamespage> danwest - have your question
[18:29] <SpamapS> I believe there are some preliminary patches to make quantum-dhcp active/active btw
[18:29] <Daviey> interesting
[18:29] <SpamapS> Probably won't land till early H
[18:33] <danwest> how do we expose HA/cluster admin events to openstack sysadmins?
[18:33] <danwest> ie. failure event
[18:33] <SpamapS> lol.. sounds a lot like nova to me. :)
[18:35] <fghaas> there is a blueprint for this for havana
[18:35] <fghaas> (about a week or so old, from john griffith)
[18:35] <Daviey> fghaas: link?
[18:35] <fghaas> https://blueprints.launchpad.net/cinder/+spec/ha-cinder-services
[18:36] <Daviey> thanks!
[18:36] <fghaas> actually older than a week, as I've noticed :)
[18:38] <fghaas> crm_mon has a general event notification facility
[18:39] <danwest> part of the bigger picture
[18:39] <danwest> well for us it may better live in landscape
[18:40] <fghaas> yes it is
[18:40] <dosaboy> can't directly tie haproxy with crm though afaik
[18:40] <dosaboy> crm produces xml
[18:40] <fghaas> html is one option
[18:40]  * dosaboy may be wrong
[18:40] <fghaas> but there is also crm_mon --daemon and writing an event handler for it
[18:41] <Daviey> fghaas: want to join the innser circler?
[18:41] <Daviey> ^ or anyone else?
[18:41] <Daviey> inner*
[18:42] <fghaas> for now I'd be happy if I could at least view the pad :)
[18:43] <jamespage> fghaas, join https://launchpad.net/~ubuntu-etherpad
[18:44] <popey> approved
[18:44] <popey> fghaas: reload
[18:44] <Daviey> ta popey
[18:45] <fghaas> thanks :)
[18:49] <fghaas> jftr, "multiple" rings == 2 rings (corosync does not support more than 2)
[18:49] <fghaas> :)
[18:50] <fghaas> silly question, ceph integration does mean rbd, correct? (not cephfs)
[18:50] <jamespage> fghaas: yes
[18:50] <fghaas> +1
[18:50] <fghaas> yes :)
[18:51] <fghaas> everyone aware of the pacemaker integration for rbd? (just so no-one reinvents the wheel)
[18:52] <jamespage> fghaas, ceph-resource-agents package right?
[18:52] <fghaas> jamespage: correct
[18:52] <fghaas> it's upstream
[18:53] <fghaas> sure, let me know and I'll send them a pull req, or have a go at it yourself, either way is fine
[18:53] <fghaas> note of course that rbd version 2 is not yet supported for mapping from the kernel
[18:53] <fghaas> it's librbd/qemu/kvm only
[18:53] <Daviey> fghaas: are you sure you won't join us on the hangout?
[18:53] <Daviey> Might make more sense..
[18:53] <fghaas> sec
[18:54] <fghaas> Daviey: prolly 2l8 :)
[18:55] <Daviey> if you just abbreviated too-late to 2l8, i retract the invite. :)
[18:58] <fghaas> Daviey: was just trying to save time on typing :)
[19:00] <fghaas> for those of you coming to Portland, there will be a bunch of HA related material in the talks there. probably worthwhile to attend.
[19:03] <Daviey> Anyone else want to be in the fish bowl?
[19:03] <zul> why do i feel like i been stuck in the same room at uds all day?
[19:03] <Daviey> starting shortly
[19:03] <zul> Daviey:  ill hop ib
[19:14] <smoser> stgraber, could you post a link to the per-architecture apt format? or info on that.
[19:14] <smoser> (sources.list)
[19:14] <rbasak> smoser: example: deb [arch=armhf] http://ports.ubuntu.com/ubuntu-ports quantal main universe
[19:15] <rbasak> If that's what you mean. I'm in a different session.
[19:15] <smoser> ah
[19:15] <smoser> yeah, i think lso
[19:21] <jdstrand> stgraber: rebooting is ok. hard shutdown loses it
[19:23] <Daviey> hallyn: is TPM interesting here?
[19:23] <Daviey> stgraber:
[19:28] <jdstrand> hallyn: that's correct
[19:34] <jdstrand> probably more than 5 minutes
[19:34] <jdstrand> but I'm not overly concerned about the concept of an setuid binary. would this actual be chmod 750 with root:lxc or something similar?
[19:36] <jdstrand> hallyn: I thought there was an idea of pushing the lxc tools to libvirt
[19:36] <jdstrand> hallyn: ah, there you are saying it :)
[19:36] <stgraber> jdstrand: we'll probably store quotas under /etc/lxc/ (number of veth per user, that kind of thing), so I'm not sure the :lxc bit will be needed but it certainly wouldn't hurt
[19:37] <jdstrand> stgraber: well, I was more just thinking of a way to limit access to the setuid binaries to a group, but that is an implementation detail
[19:38] <gavinguo> Could I ask if there is plan about ARM server, especially for ARM virtualization extension?
[19:38] <Daviey> jdstrand: do you want to join the inner circle?
[19:38] <Daviey> gavinguo: related to LXC or Qemu?
[19:38] <jdstrand> Daviey: maybe? I think my questions were answered. If I have significantly more to add, I'll hop on
[19:38] <Daviey> jdstrand: ok, cool
[19:39] <gavinguo> LXC
[19:39] <zul> stgraber:  where is the url for the liblxc library?
[19:39] <Daviey> hmm, i didn't know virt extensions was suitable for lxc?
[19:40] <gavinguo> I saw there is maintained by Columbia university.
[19:40] <Daviey> gavinguo: for LXC, or Qemu?
[19:42] <gavinguo> qemu arm
[19:43] <gavinguo> http://systems.cs.columbia.edu/projects/kvm-arm/
[19:45] <rbasak> I'm not in this session. But tomorrow's Linaro Connect session on KVM might be relevant: http://lca-13.zerista.com/event/member/72523
[19:45] <jdstrand> gavinguo: LXC are qemu are different. LXC uses Linux namespaces and other things and qemu does virtualization. you don't need virt extensions for LXC
[19:46] <Daviey> rbasak: thanks
[19:46] <jdstrand> LXC is a lightweight technology (compared to virtualization) that may be used when you don't have virt extensions, if it fits your particular use case
[19:46] <jdstrand> gavinguo: ^
[19:47] <rbasak> Also http://lca-13.zerista.com/event/member/72369
[19:47] <stgraber> zul: liblxc0 is the binary package, lxc-dev contains the header. We don't have html doc generated for the API (sadly) but you can figure most of it looking at the lxccontainer.h header and how it's used in the python3/lua binding
[19:47] <zul> stgraber:  cool thanks
[19:48] <Daviey> thanks
[19:48] <jdstrand> thanks! :)
[19:49] <gavinguo> jdstrand, so if I want to run the LXC on Arndale platoform which support A15. If there is anything I should do ?
[19:50] <hallyn> gavinguo: so long as you have CONFIG_*_NS and CONFIG*CGROUP* enabled int he kernel, lxc will just work
[19:50] <gavinguo> Because I saw the qemu are modified heavily in the Linaro git repo which supported by Christoffer dall's team.
[19:51] <gavinguo> hallyn, ok, but that does not support ARM virtualization extension, right?
[19:52] <hallyn> correct.  there's no need for virtualization extensions with xlc
[19:53] <gavinguo> hallyn, that's my qestion about how LXC support ARM server in the future? If there is any progress with ARM virutalization extension?
[19:54] <gavinguo> hallyn, I think that ARM virtualization extension will be a trend in the future.
[19:54] <hallyn> gavinguo: for the last time, lxc is based on resource namespaces and cgroups.  there is no emulation, and therefore no need or use for virtualization extensions.
[19:54] <hallyn> lxc will never use kvm virt exntensions
[19:54] <hallyn> qemu will, and when that becomes available, we should pretty quicklyi cherrypick the patches from qemu-linaro tree
[19:56] <gavinguo> hallyn, the code architecture is different between qemu and lxc?
[19:56] <hallyn> they do completely different things, yes
[19:57] <hallyn> kvm extensions are so that a cpu emulator can avoid emulating some instructions, and have the host run those instructiosn natively
[19:57] <hallyn> lxc is not an emulator, never emulates any instructions, always runs natively
[19:57] <hallyn> well, heh, when you use qemu-user-static to back a container, then yes, but all that is then in qemu :)
[19:57] <hallyn> so summary is: we will support it in qemu as soon as we can :)
[19:57] <gavinguo> hallyn, ok, understand.
[19:58] <hallyn> cool - ttyl
[19:58] <gavinguo> hallyn, that's what I'm expected.
[19:58] <gavinguo> hallyn, looking forward to the support for ARM VE. :-)