Thursday, 2012-11-29

*** colinmcnamara has quit IRC00:00
*** colinmcnamara has joined #openstack-meeting00:01
*** dwcramer has joined #openstack-meeting00:02
*** jcooley is now known as jcooley_afk00:06
*** colinmcnamara1 has joined #openstack-meeting00:08
*** colinmcnamara has quit IRC00:08
*** KurtMartin has joined #openstack-meeting00:09
*** KurtMartin has quit IRC00:10
*** krtaylor has joined #openstack-meeting00:20
*** colinmcnamara has joined #openstack-meeting00:22
*** colinmcnamara1 has quit IRC00:24
*** colinmcnamara has quit IRC00:27
*** lloydde has quit IRC00:34
*** lloydde has joined #openstack-meeting00:34
*** colinmcnamara has joined #openstack-meeting00:36
*** lloydde has quit IRC00:39
*** hemna has joined #openstack-meeting00:41
*** colinmcnamara has quit IRC00:43
*** colinmcnamara has joined #openstack-meeting00:43
*** mikal has quit IRC00:44
*** KurtMartin has joined #openstack-meeting00:45
*** gyee has quit IRC00:46
*** salv-orlando_ has joined #openstack-meeting00:52
*** salv-orlando has quit IRC00:52
*** salv-orlando_ is now known as salv-orlando00:52
*** sarob has joined #openstack-meeting00:55
*** colinmcnamara has quit IRC01:00
*** sarob has quit IRC01:01
*** sarob has joined #openstack-meeting01:02
*** ryanpetr_ has joined #openstack-meeting01:03
*** lloydde has joined #openstack-meeting01:04
*** mnewby has quit IRC01:06
*** ryanpetrello has quit IRC01:06
*** vipul is now known as vipul|away01:07
*** roadTripper has quit IRC01:08
*** lloydde has quit IRC01:09
*** markwash has quit IRC01:09
*** lloydde has joined #openstack-meeting01:09
*** lloydde has quit IRC01:14
*** vishy is now known as vishy_zz01:18
*** ryanpetr_ has quit IRC01:20
*** ywu_ has joined #openstack-meeting01:21
*** dendrobates has joined #openstack-meeting01:23
*** spiffxp-away has joined #openstack-meeting01:26
*** vipul|away is now known as vipul01:32
*** jcooley_afk is now known as jcooley01:39
*** metral_ has joined #openstack-meeting01:42
*** metral has quit IRC01:47
*** metral_ is now known as metral01:47
*** sarob has quit IRC01:47
*** markmcclain has quit IRC01:48
*** metral_ has joined #openstack-meeting01:52
*** salv-orlando has quit IRC01:54
*** KeithT has quit IRC01:55
*** metral has quit IRC01:56
*** metral_ is now known as metral01:56
*** alexpilotti has quit IRC01:58
*** jcooley is now known as jcooley_afk02:04
*** jhenner has joined #openstack-meeting02:04
*** cp16net is now known as cp16net|away02:09
*** cp16net|away is now known as cp16net02:10
*** blamar has quit IRC02:18
*** jog0 has quit IRC02:26
*** milner has quit IRC02:29
*** dwcramer has quit IRC02:38
*** ywu_ has quit IRC02:44
*** ywu_ has joined #openstack-meeting02:44
*** jhenner has quit IRC02:44
*** ywu_ has quit IRC02:49
*** dwcramer has joined #openstack-meeting02:52
*** blamar has joined #openstack-meeting03:01
*** arata has joined #openstack-meeting03:02
*** ryanpetrello has joined #openstack-meeting03:15
*** colinmcnamara has joined #openstack-meeting03:25
*** blamar has quit IRC03:32
*** bitblt has joined #openstack-meeting03:51
*** jcooley_afk is now known as jcooley03:55
*** maurosr has quit IRC03:59
*** martine has joined #openstack-meeting04:01
*** danwent has quit IRC04:02
*** markmcclain has joined #openstack-meeting04:04
*** redthrux has joined #openstack-meeting04:06
*** colinmcnamara has quit IRC04:10
*** KurtMartin has quit IRC04:15
*** colinmcnamara has joined #openstack-meeting04:24
*** ryanpetrello has quit IRC04:30
*** redthrux_is_miss has joined #openstack-meeting04:37
*** redthrux_is_miss is now known as Guest3692904:37
*** Guest36929 has quit IRC04:40
*** zbitter has joined #openstack-meeting04:40
*** redthrux has quit IRC04:41
*** hemna_ has joined #openstack-meeting04:43
*** zaneb has quit IRC04:43
*** blamar has joined #openstack-meeting04:44
*** hemna has quit IRC04:47
*** stevebake has quit IRC04:49
*** redthrux has joined #openstack-meeting04:51
*** redthrux has quit IRC04:52
*** ryanpetrello has joined #openstack-meeting04:56
*** samkottler is now known as samkottler|out05:00
*** blamar has quit IRC05:06
*** stevebake has joined #openstack-meeting05:07
*** martine has quit IRC05:13
*** bitblt has quit IRC05:14
*** adjohn has quit IRC05:18
*** arata has quit IRC05:26
*** maoy has quit IRC05:35
*** lloydde has joined #openstack-meeting05:40
*** lloydde has quit IRC05:43
*** colinmcnamara has quit IRC05:45
*** colinmcnamara has joined #openstack-meeting05:50
*** jcooley is now known as jcooley_afk06:27
*** ryanpetrello has quit IRC06:32
*** markmcclain has quit IRC06:35
*** jcooley_afk is now known as jcooley06:36
*** colinmcnamara has quit IRC06:51
*** ociuhandu has quit IRC06:54
*** adjohn has joined #openstack-meeting07:18
*** dendrobates has quit IRC07:20
*** garyk has joined #openstack-meeting07:23
*** ociuhandu has joined #openstack-meeting07:31
*** mnewby has joined #openstack-meeting07:48
*** dendrobates has joined #openstack-meeting08:11
*** dendrobates has quit IRC08:11
*** dendrobates has joined #openstack-meeting08:11
*** ttrifonov_zZzz is now known as ttrifonov08:13
*** dendrobates has quit IRC08:16
*** rafaduran has joined #openstack-meeting08:42
*** rafaduran has left #openstack-meeting08:43
*** flaper87 has joined #openstack-meeting08:46
*** garyk has quit IRC08:59
*** jcooley is now known as jcooley_afk09:01
*** afazekas has joined #openstack-meeting09:01
*** garyk has joined #openstack-meeting09:01
*** afazekas has quit IRC09:06
*** salv-orlando has joined #openstack-meeting09:09
*** garyk has quit IRC09:12
*** derekh has joined #openstack-meeting09:16
*** mnewby has quit IRC09:19
*** afazekas has joined #openstack-meeting09:21
*** garyk has joined #openstack-meeting09:27
*** darraghb has joined #openstack-meeting09:30
*** adjohn has quit IRC09:32
*** garyk has quit IRC09:38
*** garyk has joined #openstack-meeting09:42
*** garyk has quit IRC09:48
*** garyk has joined #openstack-meeting09:49
*** almaisan-away is now known as al-maisan10:34
*** al-maisan is now known as almaisan-away10:37
*** dendrobates has joined #openstack-meeting10:49
*** dendrobates has quit IRC10:54
*** jhenner has joined #openstack-meeting11:09
*** maurosr has joined #openstack-meeting11:09
*** saschpe has joined #openstack-meeting11:13
*** flaper87 has quit IRC11:13
*** garyk has quit IRC11:15
*** garyk has joined #openstack-meeting11:30
*** maurosr has quit IRC11:37
*** maurosr has joined #openstack-meeting11:42
*** rkukura has quit IRC11:44
*** salv-orlando_ has joined #openstack-meeting11:46
*** salv-orlando has quit IRC11:46
*** salv-orlando_ is now known as salv-orlando11:46
*** salv-orlando_ has joined #openstack-meeting11:46
*** salv-orlando has quit IRC11:46
*** salv-orlando_ is now known as salv-orlando11:46
*** maurosr has quit IRC11:51
*** maurosr has joined #openstack-meeting11:52
*** alexpilotti has joined #openstack-meeting12:00
*** garyk has quit IRC12:14
*** jhenner has quit IRC12:18
*** milner has joined #openstack-meeting12:21
*** roadTripper has joined #openstack-meeting12:27
*** dhellmann-afk has quit IRC12:39
*** jrd has quit IRC12:52
*** martine has joined #openstack-meeting12:58
*** salv-orlando_ has joined #openstack-meeting13:02
*** salv-orlando has quit IRC13:02
*** salv-orlando_ is now known as salv-orlando13:02
*** markvoelker has joined #openstack-meeting13:04
*** blamar has joined #openstack-meeting13:11
*** blamar_ has joined #openstack-meeting13:12
*** krtaylor has quit IRC13:15
*** krtaylor has joined #openstack-meeting13:15
*** blamar has quit IRC13:15
*** blamar_ is now known as blamar13:15
*** samkottler|out is now known as samkottler13:20
*** maoy has joined #openstack-meeting13:22
*** sandywalsh has joined #openstack-meeting13:24
*** arbrandes has joined #openstack-meeting13:25
*** garyk has joined #openstack-meeting13:27
*** dolphm has joined #openstack-meeting13:28
*** martine has quit IRC13:29
*** arbrandes has quit IRC13:40
*** martine has joined #openstack-meeting13:41
*** jhenner has joined #openstack-meeting13:43
*** flaper87 has joined #openstack-meeting13:43
*** sandywalsh has quit IRC13:46
*** flaper87 has quit IRC13:46
*** maoy has quit IRC13:47
*** jrd has joined #openstack-meeting13:48
*** samkottler is now known as samkottler|bbl13:55
*** jhenner has quit IRC13:56
*** jhenner has joined #openstack-meeting13:58
*** jhenner has quit IRC14:04
*** jhenner has joined #openstack-meeting14:05
*** dprince has joined #openstack-meeting14:09
*** milner has quit IRC14:09
*** milner has joined #openstack-meeting14:11
*** primeministerp has joined #openstack-meeting14:18
*** dhellmann has joined #openstack-meeting14:27
*** jaypipes has joined #openstack-meeting14:38
*** annegentle_ has joined #openstack-meeting14:39
*** ayoung has joined #openstack-meeting14:40
*** jcru has joined #openstack-meeting14:43
*** yjiang51 has joined #openstack-meeting14:44
*** yjiang51 is now known as yjiang5_away14:45
*** markmcclain has joined #openstack-meeting14:45
*** markmcclain has quit IRC14:46
*** dhellmann has quit IRC14:47
*** rkukura has joined #openstack-meeting14:48
*** n0ano has joined #openstack-meeting14:49
*** mattray has joined #openstack-meeting14:49
*** mtreinish has joined #openstack-meeting14:51
*** jcooley_afk is now known as jcooley14:51
*** eglynn has joined #openstack-meeting14:53
*** mestery has quit IRC14:55
*** Pete_ has joined #openstack-meeting14:59
nijaba#startmeeting Ceilometer15:00
nijaba#meetingtopic Ceilometer15:00
nijaba#chair nijaba15:00
openstackMeeting started Thu Nov 29 15:00:02 2012 UTC.  The chair is nijaba. Information about MeetBot at
openstackUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.15:00
*** openstack changes topic to " (Meeting topic: Ceilometer)"15:00
openstackThe meeting name has been set to 'ceilometer'15:00
*** openstack changes topic to " (Meeting topic: Ceilometer)"15:00
openstackCurrent chairs: nijaba15:00
nijabaATTENTION: please keep discussion focussed on topic until we reach open discussion topic15:00
nijabaHello everyone! Show of hands, who is around for the ceilometer meeting?15:00
nijaba#topic actions from previous meeting15:01
nijaba#topic eglynn drive fleshing out of nova-virt API to completion15:01
*** openstack changes topic to "actions from previous meeting (Meeting topic: Ceilometer)"15:01
*** openstack changes topic to "eglynn drive fleshing out of nova-virt API to completion (Meeting topic: Ceilometer)"15:01
eglynnSo lots of circular discussion on the upstream ML15:01
eglynnand we discussed this further on #openstack-metering yesterday15:01
eglynnthe new preferred approach is to code directly to the underlying hypervisor layer15:01
eglynn(e.g. the libvirt API)15:01
eglynnso as to avoid the release mgmt / versioning complexities of nova releasing a separate library15:02
nijabahow does this differ from what we have today?15:02
eglynnwe currently code to the hypervisor driver in nova15:02
eglynn(not to the libvirt-python API)15:02
*** dhellmann has joined #openstack-meeting15:03
nijabaah, k15:03
eglynnthe latter is packaged as a separate library15:03
eglynn(independently versioned etc.)15:03
nijababut would we be able to retrieve openstack context that way (ie meta-data)?15:03
eglynnwe only use a tiny chunk of the nova hypervisor driver15:03
n0anodoes libvirt control Xen & qemu VMs in addition to KVM?15:03
eglynnn0ano: it can do15:03
*** davidf_76 has joined #openstack-meeting15:04
eglynnso the idea would be also take a similar approach with xenapi etc.15:04
n0anodoes that mean it doesn't right now?15:04
nijaban0ano: it's one of 2 possibilities for xen, and it is the case for lxc15:04
n0anoOK, but what about qemu?15:04
eglynnit can do, if configured to do so15:04
nijabaeglynn: not sure you got to answer my q:15:05
eglynnyeah qemu too15:05
nijababut would we be able to retrieve openstack context that way (ie meta-data)?15:05
eglynnnijaba: so that's one wrinkle15:05
nijabaa big one...15:05
eglynnthe libvirt domain representation doesn't give us the flavor or display name15:05
dhellmannthose are pretty important details for billing15:06
eglynn(both of which we need for metering, certainly the flavor for the instance.<type> meter is needed, display name could be dropped maybe...)15:06
nijabahmmm... important as in essential, yes15:06
eglynnso we'll still have to call out to nova-api to get this info15:06
eglynn(though we may not have to call out to it on every polling cycle)15:06
dhellmannotoh, if we query nova for a list of instances first we wouldn't have to worry about the formula for vm names, right?15:06
eglynni.e. only when we see a new instance UUID15:06
nijabaeglynn: ah, and then we could retrieve the info through standard api calls?15:07
dhellmannhow would we see that? through notifications?15:07
eglynnnijaba: standard API, as in nova public API?15:07
yjiang5_awaydhellmann: Through novaclient or notification, either way15:07
*** andrewbogott_afk is now known as andrewbogott15:07
nijabaeglynn: yep?15:07
eglynnnijaba: sure15:07
eglynnnijaba: (that's how we list the instances currently)15:08
dhellmanndoes the nova api allow queries for instances on a given host?15:08
eglynn(i.e. we no longer go direct to the nova DB as before)15:08
yjiang5_awaydhellmann: that's curent method, nova client to get all instance on given host15:08
dhellmannyjiang5_away: current where? in ceilometer?15:08
dhellmannI thought we were using an internal API to talk to nova now.15:09
eglynnyep it does15:09
yjiang5_awaydhellmann: yes, I remember we get instance for one host through novaclient?15:09
eglynndhellman not to grab the instance list, that goes thru the public nova API15:09
eglynndhellmann ^^^15:09
dhellmanneglynn: ah, I thought everything was internal15:09
dhellmannok, so we have a way to get the data without relying on internals, we just need to work out how often15:09
eglynnso it used to be grabbed via the nova DB, but that changed recently15:09
eglynndhellmann: so I'm thinking every time we see a new instance UUID15:10
dhellmannI suggest we start simple and continue to ask every time we start to poll, then optimize as a separate step after we have the libvirt wrapper built15:10
eglynnyep reasonable15:10
dhellmanneglynn: I don't know what you mean by "see a new instance UUID", though15:10
*** shardy has joined #openstack-meeting15:10
dhellmannwhat bit of code is seeing it?15:10
eglynndhellmann: when libvirt reports an instance UUID that wasn't included in the instances described by the *last* call to the public nova API15:11
dhellmannaha, so libvirt knows the uuid?15:11
nijabaright, we don't currently maintain any list, IIRC...15:11
dhellmannI thought it only had a "name" or something15:11
eglynndhellmann: yep15:11
eglynndhellmann: nova uuid, and a libvirt ID, and a formated instance-000000%x tyep name15:12
dhellmanneasy, then, we can just ask libvirt for instances and then ask nova for details as we discover new ones15:12
yjiang5_awayeglynn: we need check  nova code invoked by ceilometer currently, to see if any thing nova added after they got information from libvirt?15:12
n0anoso, if I understand this, we're now requiring libvirt for things to work, if you configure to use xenapi you won't get this info15:12
nijabaok, so who takes the action to transform the agent that way?15:12
eglynnn0ano the idea to do the same for xenapi also15:13
nijaban0ano: it's already like this anyway.  we currently only spport libvirt15:13
eglynnnow this may not be quite as neat with the xenapi version15:13
eglynn(where there's some RRD file parsing needed to get the hypervisor stats reliably)15:13
*** markmcclain has joined #openstack-meeting15:13
* eglynn needs to look into that in more detail15:13
eglynnyjiang5_away: I didn't understand the question15:13
n0anojust so long as we don't forget the hypervisor APIs15:14
n0anos/hypervisor/other hypervisor15:14
eglynnso I have a prototype of this working for libvirt, need to polish a bit more before proposing for review15:14
*** gpernot has joined #openstack-meeting15:14
eglynnn0ano: agree15:14
eglynnshould have a patch tmrw15:14
yjiang5_awayeglynn: currently we invoke nova code to fetch libvirt information,right? Did nova added anything before return to ceilometer? If yes, that added stuff should be handled by ceilometer then.15:14
nijabaeglynn: cool, thanks.  should you action yourself on this, of you may need more than a week?15:14
dhellmannn0ano: yep, we want to support the other hypervisors, eventually15:14
nijabahey jd__!15:15
eglynnyjiang5_away: so I've removed the direct nova usage, instead calling libvirt API directly, but based on the original nova code (the v. small subset we use...)15:15
dhellmanneglynn: is there a blueprint for this stuff? that may make more sense than an action item15:15
eglynnyjiang5_away: so I don't think I've missed any nova "secret sauce"15:16
nijabadhellmann: true15:16
eglynndhellmann: I'll file a BP15:16
yjiang5_awayeglynn: great.15:16
dhellmanneglynn: cool, thanks15:16
nijabashall we move on?15:16
dhellmannalso, eglynn, double extra plus thanks for taking the lead on this. I'm *so* tired of nova breaking us.15:16
eglynnyeah, I hear ya!15:16
nijaba#topic yjiang5 to start a thread on transformer15:17
*** openstack changes topic to "yjiang5 to start a thread on transformer (Meeting topic: Ceilometer)"15:17
* nijaba thanks eglynn warmly too15:17
yjiang5_awaynijaba: yes, send mail to ML for discussion15:17
nijabaany conclusion to report?15:17
jd__I think we now have a good overview of what we want15:18
yjiang5_awayno objection to the transfomer method, only concern is dhellmann think it's too flexible15:18
jd__yeah, how "too flexible" can be a concern, really, dhellmann ;)15:18
nijabadhellmann: really ? ;)15:18
dhellmannI'm still a little worried about the notion of chaining transformers, and the complexity of configuring that. But I'm content to wait for a patch.15:18
yjiang5_awayI plan to use this method to finish CW publisher and then sent out for discussion. At that time, we will have a solid base patch for future discussion.15:19
jd__works for me15:19
dhellmannthe more complicated it is, the more likely someone will have trouble setting it up15:19
eglynnyjiang5_away: cool15:19
dhellmannyjiang5_away: excellent15:19
yjiang5_awaydhellmann: yes, so I think mostly they should use default one.15:20
nijabayjiang5_away: action?15:20
dhellmannyjiang5_away: right. but then what's the point of building a configuration system? :-)15:20
yjiang5_awaydhellmann: some advanced user, or someone may want to add special metrics themselves through configuration15:20
jd__dhellmann: evolution! :)15:20
yjiang5_awaynijaba: yes, my action to send patches for transformer15:21
dhellmannyjiang5_away: OK. I just prefer to add features one at a time. Multiple publishers first, then deployer configuration for them.15:21
nijaba#action yjiang5_away to send patches for transformer15:21
eglynnso in terms of configuration, should be mostly pre-canned right?15:21
nijaba#topic eglynn to report on synaps' decision15:21
*** openstack changes topic to "eglynn to report on synaps' decision (Meeting topic: Ceilometer)"15:21
eglynnthe synaps guys have come to a conclusion15:22
eglynnand they want to come on board15:22
shardy+1 :)15:22
eglynnbut have limited bandwidth initially (other internal demands on their time)15:22
jd__great :)15:22
dhellmanndon't we all15:22
eglynnso I'm proposing to have the conversation now on the upstream ML15:22
nijabathat's fine.  Let's make sure we help them feel "at home"15:23
eglynn(re-)invite them on board and discuss the mechanics of bringing their code under the ceilo umbrella15:23
yjiang5_awaydhellmann: yes, I'm now working on multiple publisher15:23
eglynnso we'll have a good starting point I think for a standalone monitoring solution15:23
eglynnI have a lundry list of initial tasks that I'll translate to blueprints15:23
eglynn#action eglynn kick off Synaps discussion on upstream ML15:24
nijaba#topic nijaba to get rid of roadmap page content to point to lp15:24
nijabaThis was done, see:15:24
*** openstack changes topic to "nijaba to get rid of roadmap page content to point to lp (Meeting topic: Ceilometer)"15:24
nijaba#topic dhellmann test more lenient anyjson support for folsom and grizzly compatibility15:25
*** openstack changes topic to "dhellmann test more lenient anyjson support for folsom and grizzly compatibility (Meeting topic: Ceilometer)"15:25
nijabaI think this was done and merged15:25
dhellmannyes, that's in now15:25
nijabadhellmann: correct?15:25
dhellmannor at least up for review15:25
nijabaperfect, thanks!15:25
nijaba#topic dhellmann update user-api blueprint15:25
*** openstack changes topic to "dhellmann update user-api blueprint (Meeting topic: Ceilometer)"15:25
dhellmannjust checked and it has merged15:25
dhellmannI did update that blue print with a link to a wiki page where I described the work a bit15:26
*** samkottler|bbl is now known as samkottler15:27
nijabaany action, appart from coding, left?15:27
dhellmannno, I don't think so15:27
dhellmannI'm waiting to start coding until my WSME conversion is completed15:27
dhellmannbaby steps...15:27
nijaba#topic dhellmann update pecan port blueprint15:27
*** openstack changes topic to "dhellmann update pecan port blueprint (Meeting topic: Ceilometer)"15:27
dhellmannI've updated the blueprint with some details15:28
nijabadhellmann: you had a lot of action!15:28
dhellmannI have the port to pecan complete, but ran into an issue with WSME so haven't finished that part yet15:28
dhellmannI hope to have that done before I go on vacation next weekend15:28
dhellmannthen we can look at asalkeld's ideas for changes to the API and terminology (most of which I really like)15:29
nijabasounds good15:29
jd__"action week"15:29
nijaba#topic zykes to report status on Bufunfa / BillingStack port15:29
*** openstack changes topic to "zykes to report status on Bufunfa / BillingStack port (Meeting topic: Ceilometer)"15:29
*** mestery has joined #openstack-meeting15:29
* dhellmann is cramming work in before going out for 2 weeks15:29
eglynnzykes- there?15:30
* nijaba think dhellmann won't be able to do any work after having dinner with jd and him15:30
nijabadoes not look like he is around, let's move on15:30
* dhellmann wasn't planning on it15:30
nijaba#topic Bug squashing day15:30
nijabaIt was proposed to organize a bug squashing day, or bugfest...15:30
*** openstack changes topic to "Bug squashing day (Meeting topic: Ceilometer)"15:30
nijabaso, when should this be?15:31
eglynnwhat's our bug queue length like?15:31
nijabajust before g2 release?15:31
dhellmannjust before g2, during g3, both?15:31
eglynnit's good to have a pre-prepared list of low hanging fruit15:31
eglynn(for noobs who happen upon the bug squshing day ...)15:32
nijabaeglynn: 64 open bugs of all sorts15:32
jd__nijaba: :))15:32
eglynnso we should triage this queue and try to identify some nice self-contained fixes15:32
jd__a first one I think before g2 would be good idea15:32
eglynn(and resist the temptation to fix them ourselves!)15:33
eglynnjd__ agree15:33
nijabaeglynn: only 9 of them need triaging (new)15:33
nijaba30 are either new, confirmed or triaged, quite a few of them wishlist15:34
eglynnso at least two things we can acheive with such a bug day: 1. drive quality 2. find some new active contributors15:34
eglynnfor #2, wishlist low-priority bugs are fine15:35
dhellmann12 are effort-s, too15:35
nijabaok, so g2 is jan 10.  bug day on jan 7th (monday)?15:35
eglynnhmmm, that seems a bit tight15:35
eglynnmaybe late on the week before?15:36
dhellmann3rd or 4th?15:36
nijabaeglynn: I am afraid of hangover or vacations if we do that the week before15:36
eglynn(otherwise we may be getting jumpy about regressions)15:36
eglynnnijaba: a-ha, Ok, forgot about that ;)15:36
jd__oh I think 3rd or 4th should be almost hangover free15:37
jd__not sure about vacations though15:37
nijabaok, let's vote then ;)15:37
jd__anyway, both are fine for me15:37
eglynnso if we go with the 7th, some of the bug day fixes will prolly miss g215:37
eglynn(but maybe that's OK)15:37
eglynn(g3 will be along soon enough...)15:38
yjiang5_awayeglynn: agree15:38
nijaba#startvote when to do bugday? jan4, jan715:38
openstackBegin voting on: when to do bugday? Valid vote options are jan4, jan7.15:38
openstackVote using '#vote OPTION'. Only your last vote counts.15:38
nijaba#vote jan715:38
yjiang5_away#vote jan715:38
eglynn#vote jan415:38
n0ano#vote jan415:38
dhellmann#vote jan415:38
jd__#vote jan415:38
nijaba20 sec countdown15:38
openstackVoted on "when to do bugday?" Results are15:39
openstackjan4 (4): jd__, n0ano, dhellmann, eglynn15:39
openstackjan7 (2): yjiang5_away, nijaba15:39
*** ryanpetrello has joined #openstack-meeting15:39
jd__eat that January 7th!15:39
nijaba#agreed bugday on jan 4th15:39
* n0ano we have a mandate :-)15:39
nijabaok, I guess we'll define the date for g3 later?15:40
nijababased on our experience?15:40
eglynnyep makes sense15:40
dhellmannnijaba: +115:40
dhellmannmaybe not so close to the deadline next time :-)15:40
nijabalet's move on then15:40
nijaba#topic Review blueprints and progress15:40
*** openstack changes topic to "Review blueprints and progress (Meeting topic: Ceilometer)"15:40
nijabaso, I think we should define a close date for blueprint validation15:41
nijabag2 and g3 are coming up fast15:41
*** rohitk has joined #openstack-meeting15:41
nijabaso I think if we have not agreed on a bp fore dec 15th, we should probably can it....15:41
dhellmannthat seems reasonable15:41
dhellmannwhat's the process for "agreeing"?15:42
dhellmanndiscusson the list?15:42
jd__I guess15:42
eglynnthat's a deadline for g2 only? (or for grizzly in general?)15:42
nijabawe don't have a formal agreement process, but consensus on the features and how to achve it is what we have done bfore15:42
nijabaeglynn: in general15:42
dhellmannok, I just want to make sure I get the agreements I need15:43
nijaba#startvote agree on dec 15th for bluprint "freeze"? yes, no, abstain15:44
openstackBegin voting on: agree on dec 15th for bluprint "freeze"? Valid vote options are yes, no, abstain.15:44
openstackVote using '#vote OPTION'. Only your last vote counts.15:44
nijaba#vote yes15:44
dhellmann#vote yes15:44
eglynn#vote yes15:44
jd__#vote yes15:44
n0ano#vote yes15:44
yjiang5_away#vote yes15:44
openstackVoted on "agree on dec 15th for bluprint "freeze"?" Results are15:44
openstackyes (6): n0ano, jd__, nijaba, eglynn, yjiang5_away, dhellmann15:44
*** jcru is now known as jcru|away15:45
nijaba#agreed blueprint freeze on dec 15th15:45
nijaba#topic Discuss suggested re-narrowing of project scope for grizzly versus user-oriented monitoring15:45
*** openstack changes topic to "Discuss suggested re-narrowing of project scope for grizzly versus user-oriented monitoring (Meeting topic: Ceilometer)"15:45
eglynnso this topic was motivated by markmc's observations on the ML:
nijabaI think this was proposed by someone outside of the project15:45
nijabaI am currently not too afraid of this15:45
nijabawhat do you guys think?15:46
nijabaspecially now that we have a bp freeze15:46
*** alexpilotti has quit IRC15:46
dhellmannI think some of this was brought about by the "thrashing" of the discussion of how to handle monitoring15:46
eglynnwell Synaps will give us a (user-oriented) monitoring service, and multi-publish will give the ability to push metrics into that15:47
dhellmannyes, so now we have a more defined direction15:47
eglynnbut I agree that extending ourselves to also cover system oriented monitoring / instrumentation now would be a bridge too far15:47
*** dendrobates has joined #openstack-meeting15:47
*** dendrobates has joined #openstack-meeting15:47
eglynn(for grizzly anyway)15:47
nijabaeglynn: on this I tend to agree, but we don't even have the start of a bp for it15:48
eglynnby system oriented monitoring I mean cloud-operator-focussed15:48
dhellmannyes, I think we need to start differentiating between short term and long term goals for these new features15:48
nijabaso I was seeing it as dicussions for h15:48
n0anoI'm more interested in system monitoring but I'm more than willing to address that post-grizzly15:48
nijababut we should definitily continue the discussion15:48
nijabaok, cool!!!15:49
n0anodiscussion - yes, actively push code upstream - not just yet15:49
nijaba#topic multi-publisher blueprint15:49
*** openstack changes topic to "multi-publisher blueprint (Meeting topic: Ceilometer)"15:49
jd__until there's any code, that's just kind of useless discussion anyway :)15:49
yjiang5_awayI'm working on the patch now15:49
nijabaI think we are getting close to have an agreement on this on right?15:49
yjiang5_awaynijaba: yes15:49
jd__yeah, yjiang5_away is doing the grunt work :)15:49
nijabathanks yjiang5_away15:50
nijaba#topic multi-dimensions bp15:50
*** openstack changes topic to "multi-dimensions bp (Meeting topic: Ceilometer)"15:50
nijabaso this is a new one that jd and I came up with15:50
*** jcru|away is now known as jcru15:50
*** maoy has joined #openstack-meeting15:51
*** cp16net is now known as cp16net|away15:51
*** grapex1 has joined #openstack-meeting15:51
nijabaso the question left on this one is wether we need to express complex dimension with & nd | or not15:51
nijabaI think we could propose only & for a first version15:52
jd__I think too15:52
dhellmannnijaba: +115:52
nijabaand extend it later if needed15:52
jd__or can be achieved with multiple request easily for now15:52
dhellmannI don't care for the syntax described, but the intent is good15:52
nijaba#action nijaba to update the bp to specify complex request in a future version15:53
nijabadhellmann: feel free to prppose a syntax :)15:53
dhellmannnijaba: something more closely resembling a regular GET request would be easier to handle15:53
nijabadhellmann: that reminds me of what jd was suggesting15:54
dhellmannnijaba: right. let me work up an example using a syntax WSME supports :-)15:55
nijabaok, I'll wait for your example then. thanks15:55
dhellmannthe controller method would get an array of dimension objects as an argument if we do it right15:55
nijabathat was the idea15:55
*** cp16net|away is now known as cp16net15:55
dhellmann#action dhellmann to prepare example of dimension query syntax that will work with WSME15:56
nijaba#topic Discuss adopting asalkeld's client implementation officially15:56
*** openstack changes topic to "Discuss adopting asalkeld's client implementation officially (Meeting topic: Ceilometer)"15:56
nijabaI am actually all for this proposal15:57
nijabaanything blocking?15:57
jd__what's asalkeld version?15:57
eglynnso when zykes- asked about CLIs a few days ago, I was surprised there were so many15:57
* dhellmann looks for link15:57
jd__I though dhellmann had one under its foot15:57
*** hemna_ has quit IRC15:57
dhellmannjd__: ours is much more primitive15:57
dhellmannasalkeld used the "standard" client model15:58
dhellmannmuch more complete15:58
dhellmannwe really need a client lib15:58
jd__ok :)15:58
eglynnsounds like the one to row in behind15:58
nijabaanyone against?15:58
jd__didn't look at the code, but I'm probably not against anyway15:59
* nijaba needs to run to deliver a speech, sorry to be in a hurry15:59
dhellmannI'm obviously in favor15:59
jd__can we put that into openstack'core/incubated?15:59
dhellmannthat's the idea, to have another git repo setup for it with CI integration, etc.15:59
yjiang5_awayjd__: normally client is a seperated project?15:59
jd__yjiang5_away: yes15:59
nijaba#agreed adopt asalkeld's client officially15:59
jd__dhellmann: perfect, what do we have to do for that?15:59
dhellmannjd__: dunno16:00
eglynnyjiang5: seperate repo, but same overall project?16:00
dhellmanntalk to the infra team, I guess16:00
jd__#action move asalkeld's client into openstack incubation/core16:00
nijabado you guys mind to continue the discussion on our chan?  I need to run :/16:00
dhellmannthat needs an owner16:00
jd__#action jd__ move asalkeld's client into openstack incubation/core16:00
jd__nijaba: #chair me I'll finish16:00
dhellmannnijaba: I think we're done16:00
nijaba#chair jd__16:00
dhellmannor maybe not16:00
openstackCurrent chairs: jd__ nijaba16:00
jd__run nijaba! run!16:00
nijabajd__: thanks, did not know I could do that during the meeting16:00
* nijaba waves and run16:01
* eglynn wonders if we've any sharable presentation collateral?16:01
eglynn(giving a talk to an Irish openstack user group next week on ceilo/monitoring etc.)16:01
jd__#topic open discussion16:02
*** openstack changes topic to "open discussion (Meeting topic: Ceilometer)"16:02
yjiang5_away eglynn: I remmeber  jd__ sent some in IRC before16:02
jd__eglynn: I've the ones we used with nijaba a few weeks ago16:02
eglynnjd__, yjiang5: cool16:02
eglynnthanks, I may borrow liberally ;)16:03
yjiang5_awayjd__: can we put into ceilometer launchpad?16:03
dhellmannyjiang5_away: good idea16:04
yjiang5_awayjd__: thanks.16:04
dhellmannyjiang5_away: even better, let's add a link to our docs16:05
yjiang5_awaydhellmann: yep, and eglynn can add his new one after finish the talking :)16:05
dhellmannspeaking of docs, the jenkins job to build them is now working16:05
eglynnyjiang5: yep, will do!16:05
dhellmannI opened a bug to fix the styling16:06
dhellmanndoes someone have time to talk to the doc team about whether something will link to us?16:06
dhellmannI don't know what we have to do to get a link on the project list, for example16:07
jd__bribe someone I guess16:07
jd__somebody for some #action ?16:07
yjiang5_awaydhellmann: So basically contacct with doc team?16:08
dhellmannyjiang5_away: yes, we need someone to email them and ask what to do next16:08
yjiang5_awaydhellmann: I'm glad to help16:08
dhellmanngreat! want to take an #action item?16:09
yjiang5_awaydhellmann: yes, but I don't know how to #action yet :$16:09
dhellmannjust type it out :-)16:09
eglynnyjiang5_away: #action yjiang description16:09
yjiang5_away#action talk with doc team to link to ceilometer docs16:10
*** dolphm has quit IRC16:10
jd__you forgot your nickname after #action I think :)16:10
yjiang5_away#action yjiang5 talk with doc team to link to ceilometer docs16:10
jd__that's one seems right!16:10
jd__anything else for today?16:10
*** redthrux has joined #openstack-meeting16:11
jd__last call!16:11
eglynnand that's a wrap I think ...16:11
dhellmannI'm done16:11
*** openstack changes topic to "OpenStack meetings || Development in #openstack-dev || Help in #openstack"16:11
openstackMeeting ended Thu Nov 29 16:11:49 2012 UTC.  Information about MeetBot at . (v 0.1.4)16:11
openstackMinutes (text):
jd__thanks guys!16:11
eglynnyep thanks all, productive as usual ...16:12
dhellmannthanks, everyone16:12
*** jrodom has joined #openstack-meeting16:13
*** jrodom has quit IRC16:13
*** jrodom has joined #openstack-meeting16:13
*** yjiang5_away has quit IRC16:13
*** afazekas has quit IRC16:16
*** dolphm has joined #openstack-meeting16:17
*** alexpilotti has joined #openstack-meeting16:17
*** mnewby has joined #openstack-meeting16:18
*** ryanpetrello has quit IRC16:20
*** ryanpetrello has joined #openstack-meeting16:21
*** openstack_newbe2 has quit IRC16:22
*** cbaeseman has joined #openstack-meeting16:24
*** samkottler is now known as samkottler|out16:24
*** sandywalsh has joined #openstack-meeting16:25
*** grapex1 has quit IRC16:27
*** Pete_ has quit IRC16:32
*** Mandell has quit IRC16:38
*** torandu has joined #openstack-meeting16:40
*** chunwang has joined #openstack-meeting16:41
*** mestery has quit IRC16:43
*** ryanpetrello has quit IRC16:47
*** jrodom has quit IRC16:47
*** spiffxp-away has left #openstack-meeting16:51
*** ravkumar_hp has joined #openstack-meeting16:55
*** jcru has quit IRC16:56
*** jcooley is now known as jcooley_afk16:58
*** donaldngo has joined #openstack-meeting16:58
*** jcru has joined #openstack-meeting16:58
*** donaldngo has quit IRC17:00
*** donaldngo has joined #openstack-meeting17:00
*** markvan has joined #openstack-meeting17:00
sdaguedavidkranz: qa meeting time?17:01
davidkranzAny other QA'ers here?17:01
ravkumar_hpRavi here17:01
*** openstack_newbe1 has joined #openstack-meeting17:01
chunwangChun Wang (English name is April)  here.17:02
davidkranz#startmeeting qa17:02
openstackMeeting started Thu Nov 29 17:02:34 2012 UTC.  The chair is davidkranz. Information about MeetBot at
openstackUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.17:02
*** openstack changes topic to " (Meeting topic: qa)"17:02
openstackThe meeting name has been set to 'qa'17:02
donaldngoDonald here17:02
*** jrodom has joined #openstack-meeting17:03
sdague<- here17:03
*** dwalleck has joined #openstack-meeting17:03
davidkranzjaypipes, dwalleck: You guys here?17:03
*** TheEstablishment has joined #openstack-meeting17:03
dwalleckjust got here17:03
*** rax-sam has joined #openstack-meeting17:04
davidkranzLooks like the tempest swift tests will be part of the gate as soon as the devstack fix is approved.17:04
*** rax-Jose has joined #openstack-meeting17:04
*** danwent has joined #openstack-meeting17:04
ravkumar_hpdavidkranz: that would be grat17:05
sdaguedavidkranz: which devstack fix is that?17:05
sdagueI can take a look17:05
ravkumar_hpwe are working on some new test for Swift17:05
davidkranzsdague: Any progress on the "instances go to ERROR" issue?17:05
*** blogan-RAX has joined #openstack-meeting17:05
sdaguedavidkranz: mtreinish is looking into it today, the problem is it's hard to reproduce17:05
sdagueI expect our systems are too fast17:06
sdagueso he was going to try to run it in a kvm guest to slow it down17:06
sdaguewhich is the way ci is run anyway17:06
davidkranzsdague: Yeah, this is probably acting as a lame stress test in the ci setup.17:07
davidkranzsdague: I would really like to get some stress tests in the nightly build.17:07
sdagueok, devstack patch approved17:07
dwalleckI'm going to kick a merge prop in today that should add the reason for a server going into error status into the logging as well, should help wiht debugging17:07
*** jrodom has quit IRC17:07
davidkranzsdague: Thanks. We'll see if anything blows up :)17:07
*** longshot has joined #openstack-meeting17:07
sdaguedwalleck: great17:07
*** jrd has quit IRC17:07
sdaguedavidkranz: well that devstack patch was just adding another variable to config tempest, so I hope not :)17:08
*** TheEstablishment has quit IRC17:08
davidkranzsdague: Yeah. I just meant that those tests will now start running in all project gates.17:08
*** jrodom has joined #openstack-meeting17:08
sdaguedavidkranz: cyeoh on my team is starting to look at the blueprint for converting tempest to testr/testtools. So expect to start to see some activity on that one in the next week or so.17:09
*** openstack_newbe1 has quit IRC17:09
davidkranzchunwang put in a blueprint for customized-test-launcher script17:10
*** rkukura has quit IRC17:10
*** krtaylor has quit IRC17:10
*** dwcramer has quit IRC17:10
*** lifeless has quit IRC17:10
*** ndipanov has quit IRC17:10
*** cp16net has quit IRC17:10
*** radez_g0n3 has quit IRC17:10
*** radez_g0n3 has joined #openstack-meeting17:10
*** dwcramer has joined #openstack-meeting17:11
*** lifeless has joined #openstack-meeting17:11
*** rkukura has joined #openstack-meeting17:11
*** krtaylor has joined #openstack-meeting17:11
davidkranzI think that blueprint needs some more information about how it integrates with parallel execution and what the main purpose is.17:11
*** ndipanov has joined #openstack-meeting17:11
*** cp16net has joined #openstack-meeting17:11
chunwangyes, the blueprint is a work around and improvement based on the issues currently we found during tempest execution...17:11
dwalleckdo we know yet if a testtools solution is viable? I thought someone was going to do a prototype17:12
chunwangActually I think it's not parallel execution, but a batch run of the tempest cases, and with the customized test case list and environment clean module17:12
*** ociuhandu_ has joined #openstack-meeting17:12
sdaguedwalleck: that will be cyeoh17:13
sdaguegoing to assume that it's workable, and if not, he'll flag that as an issue :)17:13
ravkumar_hpsdague: cyeoh wil analyze testtools for parallel execution also . right?17:13
sdaguechunwang: url for your blueprint?17:14
*** samkottler|out is now known as samkottler17:15
dwalleckI'm just a bit wary of diving into a solution without a full plan of attack. I'm curious though if any of this will tie us to testtools. It'd be nice if in the stripping away of nose any standard unittest test runner just worked17:15
*** enikanorov has joined #openstack-meeting17:16
sdaguedwalleck: fair, though I think in reality you can't really know unless you try17:16
chunwangwhy the parallel exeution is so important? May I know how long will the whole tempest test cases take in your environment?17:17
sdagueif it's not any good we won't force it in just because it's on the list. :-) But this will at least make it possible to evaluate17:17
sdaguedavidkranz: you have the timings for the gate?17:18
dwallecksdague: True. I guess what I was thinking was to perhaps start with converting say one project's tests first before doing the whole lot17:18
*** ociuhandu has quit IRC17:18
*** ociuhandu_ is now known as ociuhandu17:18
sdaguedwalleck: from the experience I've seen in the folks working on converting nova here, you more or less have to just do it in one batch17:18
*** jcooley_afk is now known as jcooley17:18
dwalleckchunwang: Because in a Devstack environment the tests take about 45 minutes. In a full deployed environment, especially using Windows images, it might almost take a day17:19
davidkranzsdague: The last hourly run took just under 1000 seconds for the full part.17:19
rohitkchunwang: I think in the longer term parallelism would prove more beneficial17:19
*** mestery has joined #openstack-meeting17:19
ravkumar_hpdwalleck: yes. but convert one test before converting project's all tests17:19
*** rick_rax has joined #openstack-meeting17:19
sdagueanyway, cyeoh is in australia, so I'll proxy his updates here, because it's some ungoddly hour of the night17:19
dwalleckThe goal is to just to get more done quickly so folks are more inclined to run them17:19
davidkranzsdague: and 200s for the smoke tests17:20
dwalleckAll fair points. I'm just a cautious one :-)17:20
sdaguedwalleck: right, which is why we've got a good review process :-)17:20
*** colinmcnamara has joined #openstack-meeting17:20
dwalleckyup, good point17:20
chunwangSimlar with your our environemnt, it will take about 25 hours to finish the whole tests...17:21
*** shardy has left #openstack-meeting17:21
*** colinmcnamara1 has joined #openstack-meeting17:21
sdaguechunwang: 25 hours? what environment is that?17:22
dwalleckBut in parallel, I'm running my Tempest tests (with real Linux/Windows images) in < 45 min, so getting there is the goal17:22
chunwanga Essex version Openstack, with 16 cn, 1 cc...17:22
*** dwalleck has quit IRC17:22
*** dwalleck has joined #openstack-meeting17:23
*** dwalleck has quit IRC17:23
*** dwalleck has joined #openstack-meeting17:23
davidkranzdwalleck: When we get parallelism for devstack gate we will likely run into
uvirtbotLaunchpad bug 1016633 in nova "Bad performance problem with nova.virt.firewall" [Medium,Incomplete]17:24
chunwangbut it may not similar with normal environment, becasue some of the OS image will take about 20 min to boot up at 1st time boot17:24
*** colinmcnamara2 has joined #openstack-meeting17:24
*** colinmcnamara has quit IRC17:24
sdaguechunwang: gotcha17:24
davidkranzFiring up a bunch of instances on a single compute node is slow.17:24
davidkranzI guess we will deal with that when it happens.17:25
dwalleckHmm, interesting.17:25
sdaguefor the gate, were you going to also make it error if there were stack traces in the daemons?17:25
davidkranzsdague: I really want to do that but there are still errors in the nova logs as far as I know.17:26
sdaguedavidkranz: yes, there are.17:26
*** colinmcnamara1 has quit IRC17:26
sdagueit would be great to get each one of those distinct stack traces as a nova bug17:26
davidkranzsdague: As soon as they are clean I would go for it.17:26
sdagueseperately they are solvable, I did fix one a couple weeks ago, which was serious enough to be a folsom backport as well17:27
*** Gordonz has joined #openstack-meeting17:28
*** Gordonz has quit IRC17:28
*** dhellmann has quit IRC17:28
davidkranzsdague: Yeah. It would be best if the nova team as a whole made this higher priority.17:28
*** Gordonz has joined #openstack-meeting17:28
*** markmcclain has quit IRC17:29
sdaguedavidkranz: yeh sure, just saying that if they get broken up as discreet issues, I can probably get folks looking at them17:29
*** eglynn has quit IRC17:29
davidkranzsdague: I agree. I can do look at the latest logs and do that.17:30
*** colinmcnamara2 has quit IRC17:30
sdaguethat would be great17:30
*** primeministerp has quit IRC17:30
*** colinmcnamara has joined #openstack-meeting17:31
dwalleckI need to duck out folks. I have some updates, but I'll send those in an email this afternoon. Adios!17:31
*** dwalleck has quit IRC17:31
davidkranzI am going to take a look at the multi-node-testing blueprint.17:32
*** jcooley is now known as jcooley_afk17:32
davidkranzAny one know if anything is happening with fuzz testing?17:32
chunwangMay I know when will the quantum scripts be ready for use?17:32
rohitkI will be jumping onto some quantum tests next week too17:33
davidkranzchunwang: Not sure. I think Nachi Ueno is the lead ono that.17:33
*** colinmcnamara1 has joined #openstack-meeting17:34
*** milner has quit IRC17:34
davidkranzrohitk: Great. Make sure to touch base with Nachi to avoid duplication.17:34
rohitkdavidkranz: yup :), mnewby was working on a patch I believe17:35
davidkranzAny other topics for today?17:35
rax-Joseravkumar_hp:  What features are you targeting w/ your latest swift tests?17:35
*** colinmcnamara has quit IRC17:35
rax-JoseJust curious17:35
rohitkdavidkranz: I can't see the fuzz testing/randgen coming in anytime soon17:35
rohitkwhat's the strategy to accept more negative tests?17:36
ravkumar_hprax-Jose: tempurl +
rax-Josecoolbeans, thanks.17:36
sdaguedavidkranz: nothing more from me17:37
davidkranzrohitk: Is some one working on fuzz testing at all? There isn o assignee for the blueprint.17:37
rohitkdavidkranz: I don't think so, but the old style negative API validation tests were not being accepted as they slowed down the runtime17:38
*** gyee has joined #openstack-meeting17:38
*** colinmcnamara1 has quit IRC17:38
chunwangif there is any scripts need help, I will try to help on that...17:38
davidkranzrohitk: That is still a problem. It is also a problem that they take much longer to write than if using a negative test-generator.17:38
davidkranzrohitk: But there is the up-front, non-distributed cost of doing that.17:39
rohitkdavidkranz: Agreed, I don't think we should afford that17:39
davidkranzchunwang: Is proposing and creating a negative test runner something you could do?17:39
davidkranzchunwang: There are some tools out there that do this kind of thing.17:39
sdaguerohitk: I think the important thing to is that the negative tests do more than just test fail, they need to get fails the way they expect17:40
davidkranzsdague: A negative test runner would take care of that if you specify the space of parameters and the expected result.17:40
rohitksdague: Yes, more to do around bad input17:40
chunwangok, currently I didn't proposing more negative test myself, but if there is any existing requirement there, I will try to look into it...17:40
*** jhenner has quit IRC17:42
sdaguedavidkranz: oh, that's the last thing. The coverage reporting for tempest is actually coming along by mtreinish. He's got a nova extension in final stages of review that will let us get nova coverage from an external test runner17:42
davidkranzsdague: Great.17:42
sdagueso I'm hoping that's available in a couple of weeks as part of normal tempest runs17:42
rohitksdague: Sounds good!17:43
davidkranzLast call for new issues...17:43
sdagueit doesn't really add much to the test runs time wise, so it should be something we can enable in the default runs17:43
chunwangsdague: Good news, we also need this kind of data17:43
*** colinmcnamara has joined #openstack-meeting17:45
davidkranzOK, see you all next week.17:46
sdaguesee you then17:46
*** openstack changes topic to "OpenStack meetings || Development in #openstack-dev || Help in #openstack"17:47
chunwangsee you17:47
openstackMeeting ended Thu Nov 29 17:47:00 2012 UTC.  Information about MeetBot at . (v 0.1.4)17:47
openstackMinutes (text):
*** donaldngo has quit IRC17:47
*** rax-sam has quit IRC17:47
*** blogan-RAX has quit IRC17:47
*** longshot has quit IRC17:47
*** rax-Jose has left #openstack-meeting17:47
*** annegentle_ has quit IRC17:48
*** ravkumar_hp has quit IRC17:48
*** eglynn has joined #openstack-meeting17:49
*** colinmcnamara has quit IRC17:50
*** rohitk has quit IRC17:50
*** colinmcnamara has joined #openstack-meeting17:52
*** rick_rax has quit IRC17:53
*** jcru is now known as jcru|away17:56
*** chunwang has quit IRC17:56
*** davidf_76 has quit IRC17:57
*** andrewbogott is now known as andrewbogott_afk17:58
*** jcru|away is now known as jcru18:01
*** eglynn has quit IRC18:05
*** hemna has joined #openstack-meeting18:08
*** derekh has quit IRC18:14
*** cp16net is now known as cp16net|away18:16
*** jcru is now known as jcru|away18:17
*** garyk has quit IRC18:23
*** eglynn has joined #openstack-meeting18:26
*** jrd has joined #openstack-meeting18:27
*** Mandell has joined #openstack-meeting18:29
*** colinmcnamara has quit IRC18:31
*** colinmcnamara has joined #openstack-meeting18:32
*** sarob has joined #openstack-meeting18:36
*** ryanpetrello has joined #openstack-meeting18:40
*** dhellmann has joined #openstack-meeting18:41
*** esp1 has joined #openstack-meeting18:41
*** esp1 has left #openstack-meeting18:41
*** vishy_zz is now known as vishy18:42
*** devananda has quit IRC18:42
*** devananda has joined #openstack-meeting18:43
*** markmcclain has joined #openstack-meeting18:44
*** milner has joined #openstack-meeting18:45
*** darraghb has quit IRC18:53
*** colinmcnamara has quit IRC18:57
*** colinmcnamara has joined #openstack-meeting18:58
*** colinmcnamara1 has joined #openstack-meeting18:59
*** rkukura has quit IRC19:00
*** cp16net|away is now known as cp16net19:00
*** adjohn has joined #openstack-meeting19:01
*** dhellmann_ has joined #openstack-meeting19:01
*** colinmcnamara2 has joined #openstack-meeting19:02
*** colinmcnamara has quit IRC19:02
*** colinmcnamara1 has quit IRC19:04
*** jrd has quit IRC19:05
*** dhellmann has quit IRC19:05
*** dhellmann_ is now known as dhellmann19:05
*** colinmcnamara2 has quit IRC19:06
*** colinmcnamara has joined #openstack-meeting19:07
*** andrewbogott_afk is now known as andrewbogott19:09
*** milner has quit IRC19:09
*** colinmcnamara has quit IRC19:13
*** novas0x2a|laptop has quit IRC19:13
*** colinmcnamara has joined #openstack-meeting19:14
*** jcru|away is now known as jcru19:16
*** devananda has quit IRC19:16
*** TheEstablishment has joined #openstack-meeting19:16
*** devananda has joined #openstack-meeting19:16
*** jrodom has quit IRC19:19
*** colinmcnamara1 has joined #openstack-meeting19:20
*** jrd has joined #openstack-meeting19:20
*** colinmcnamara1 has quit IRC19:21
*** colinmcnamara has quit IRC19:21
*** jog0 has joined #openstack-meeting19:23
*** markvoelker has quit IRC19:26
*** annegentle_ has joined #openstack-meeting19:31
*** devananda has quit IRC19:31
*** vipul is now known as vipul|away19:31
*** devananda has joined #openstack-meeting19:31
*** mdomsch has joined #openstack-meeting19:32
*** lloydde has joined #openstack-meeting19:33
*** dwcramer has quit IRC19:35
*** TheEstablishment has quit IRC19:37
*** jrodom has joined #openstack-meeting19:41
*** stevebake has quit IRC19:44
*** dwcramer has joined #openstack-meeting19:47
*** novas0x2a|laptop has joined #openstack-meeting19:49
*** jrd has quit IRC19:56
*** stevebake has joined #openstack-meeting19:57
*** dani4571 has joined #openstack-meeting20:01
*** zbitter has left #openstack-meeting20:02
*** cp16net is now known as cp16net|away20:04
*** cp16net|away is now known as cp16net20:05
*** annegentle__ has joined #openstack-meeting20:05
*** krtaylor has quit IRC20:07
*** jrodom has quit IRC20:09
*** jrodom has joined #openstack-meeting20:09
*** annegentle_ has quit IRC20:09
*** markmcclain1 has joined #openstack-meeting20:11
*** cp16net is now known as cp16net|away20:12
*** markmcclain2 has joined #openstack-meeting20:14
*** markmcclain2 has joined #openstack-meeting20:14
*** markmcclain2 has joined #openstack-meeting20:15
*** markmcclain has quit IRC20:15
*** alexpilotti has quit IRC20:16
*** markmcclain1 has quit IRC20:17
*** garyk has joined #openstack-meeting20:25
*** salv-orlando_ has joined #openstack-meeting20:26
*** salv-orlando has quit IRC20:26
*** salv-orlando_ is now known as salv-orlando20:26
*** dprince has quit IRC20:26
*** jrd has joined #openstack-meeting20:30
*** alexpilotti has joined #openstack-meeting20:37
*** garyk has quit IRC20:39
*** jcru has left #openstack-meeting20:44
*** dwcramer has quit IRC20:44
*** dprince has joined #openstack-meeting20:49
*** annegentle__ has quit IRC20:50
*** annegentle_ has joined #openstack-meeting20:55
*** annegentle__ has joined #openstack-meeting20:56
*** dwcramer has joined #openstack-meeting20:57
*** Vek has joined #openstack-meeting20:57
*** ayoung has quit IRC20:57
*** vishy is now known as vishy_zz20:58
*** eglynn has quit IRC20:58
*** annegentle_ has quit IRC21:00
*** dkehn has joined #openstack-meeting21:00
russellbnnnnnnnnnnnova meeting time21:01
*** markmc has joined #openstack-meeting21:01
russellb#startmeeting nova21:01
openstackMeeting started Thu Nov 29 21:01:42 2012 UTC.  The chair is russellb. Information about MeetBot at
openstackUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.21:01
*** danwent has quit IRC21:01
*** openstack changes topic to " (Meeting topic: nova)"21:01
openstackThe meeting name has been set to 'nova'21:01
* Vek yawns21:01
* dansmith snorts21:01
russellbHere's the agenda:
russellbeveryone here and awake?!21:02
russellbcool, comstud, around?21:02
*** afazekas has joined #openstack-meeting21:02
russellbor how about devananda ?21:02
sdaguehe and vishy_zz were road tripping, but it sounded like they'd be back on by now21:03
* Vek is trying to grab comstud's attention21:03
dansmithreally? road tripping?21:03
*** carlp-away has quit IRC21:03
russellbwell we'll skip to bugs for now21:03
russellb#topic bugs21:03
*** openstack changes topic to "bugs (Meeting topic: nova)"21:03
russellbnot terrible, 25 not touched yet21:04
russellbif everyone could triage a couple that'd help get us back down really low, so please do21:04
dansmithI did a couple last week I think21:04
russellbi don't have mikal's nifty script so we can give out gold stars for triage21:04
dansmither, I mean I did 2021:04
russellbdansmith gets a gold star!21:05
russellbanyway, any other bugs we need to cover?21:05
russellbreally important stuff that has popped up?21:05
*** eglynn has joined #openstack-meeting21:05
*** carlp-away has joined #openstack-meeting21:05
*** vishy_zz is now known as vishy21:05
* russellb would know had he been doing more triage before 20 minutes ago :-/21:05
alexpilottirussellb: I have a question for a bug  that I still have to file up21:05
alexpilottiit's related to porting the full nova stack to Windows21:06
russellbalexpilotti: and we can do your topic next21:06
dansmithoh, sorry, go on21:06
russellb#topic hyper-v testing, windows support21:06
*** openstack changes topic to "hyper-v testing, windows support (Meeting topic: nova)"21:06
alexpilottisimply put: in nova-api there'a large use of fork()s21:06
russellbgo ahead, and then we can talk hyper-v unit tests next21:06
russellbso windows should add fork() ?21:06
alexpilottias you probably know fork() is not implemented on Windows21:07
*** krtaylor has joined #openstack-meeting21:07
russellbis it just for the multi-process API stuff?  or is it always required?21:07
alexpilottiI'm pretty close to have a working implementation on fork(9 but I still need some low level things sorted out21:07
alexpilottirussellb: yep21:07
markmcalexpilotti, btw, that code is moving to oslo-incubator21:07
alexpilottithe problem is that cpython dosn't really offer alternatives due to the lock21:07
markmcalexpilotti, so might be better fixing it there21:08
*** salv-orlando has quit IRC21:08
vishyrussellb: you want to lead this?21:08
alexpilottianyway, the 2 way out are: rewriting it to avoid locks21:08
vishymy connection is being flaky21:08
russellbvishy: sure21:08
alexpilottian the other is spawning a process instead of a fork21:08
russellbvishy: we hit bugs already, not much there, now talking windows support / hyper-v21:09
comstudwoot, i'm here21:09
alexpilottiwhat do you guys think?21:09
*** danwent has joined #openstack-meeting21:09
markmcalexpilotti, which lock is this?21:09
russellbalexpilotti: might be worth an openstack-dev thread so we can dive into details21:09
*** eglynn has quit IRC21:09
markmcyeah, agree21:09
alexpilotti ;-)21:09
*** eglynn has joined #openstack-meeting21:10
*** dwcramer has quit IRC21:10
russellb#action alexpilotti to start a -dev list thread on dealing with usage of fork() in nova for Windows21:10
russellbso the other topic we wanted to cover was the hyper-v unit tests21:10
alexpilottirussellb: ok21:10
russellbthere seems to have been confusion around how they work, and what impact they have on development21:11
*** ttrifonov is now known as ttrifonov_zZzz21:11
russellbso i wanted alexpilotti to give an overview of what it's doing, and what the plans are for improving it21:11
alexpilottifirst, I still don't  get what issue came up with the latest live migration tests fix21:12
alexpilottiit's a 2 lines fix, just adding a lambda and that's it21:12
dansmithalexpilotti: there was no issue with the patch itself21:12
dansmithalexpilotti: sdague just raised the issue there, which is part of the confusion21:12
alexpilottidansmith: no, that's the patch to the bug21:13
uvirtbotLaunchpad bug 1083018 in nova "HyperV compute "resume" tests need to be fixed" [High,Fix committed]21:13
alexpilottinot the sdague comment21:13
vishyalexpilotti: I think the problem is the fix isn't obvious21:14
russellbso I think the stubs just scare people off and we've been assuming that we can't fix your unit tests, heh21:14
alexpilottianyway, the hyper-v tests have nothing to do with the driver's interface changes21:14
alexpilottiI understand that21:14
alexpilottiwe also want to get rid of them21:15
alexpilottithe reason we have them is that this was the only way to get the test the pre-Folsom code21:15
russellbalso worth noting nova/tests/hyperv/README.rst (which I just now discovered)21:15
*** markvan has quit IRC21:15
alexpilottiyes, I added that readme to explain the architecture21:16
russellbso sdague did you have any other questions or issues with it you wanted to discuss?21:16
dansmithwell, I don't feel as strongly as he does I think,21:17
sdaguealexpilotti: ok, so you said you are getting rid of the tests, what's coming to replace them21:17
vishymaybe a simple example in the readme involving adding a parameter to an interface21:17
vishywould help21:17
sdagueand when? it's definitely a little weird to be constantly trying to grok reviews that have 50 changed .gz files in them :)21:17
alexpilottivishy: adding a parameter to the interface doesn't need any changes in the stubs21:17
dansmithbut I really would like to see a different implementation for those so we don't have 20 changed pickle files for small fixes21:17
alexpilottisdague: I agree on that21:18
sdagueI think it was rmk that hit the actual issue21:18
alexpilottithis happens when we change code that affect all the tests internally21:18
alexpilottiit happened quite often lately due to the type of features that we are adding21:18
russellbso nobody else can fix actual driver code then?21:18
vishyalexpilotti: why did his change break the hyperv tests then?21:18
*** Mr_T has joined #openstack-meeting21:18
alexpilottivishy: I don't think that it did. that's the point21:19
sdaguealexpilotti: it did, the unit tests wouldn't pass21:19
alexpilottivishy: if you look here:
alexpilottiyou'll see that the only code that I changed is adding a couple of lambdas in the tests to handle that extra parameter in the tests21:20
dansmithalexpilotti: are you saying that changing the actual driver code would have broken the sub matching?21:20
alexpilottiin and that's it21:20
russellbdansmith: that's what i'm getting ...21:20
*** eglynn_ has joined #openstack-meeting21:20
dansmithalexpilotti: meaning, if rmk had changed it to match the other drivers he was changing, would the stubs fail?21:21
alexpilottirussellb: only when it affects the WMI calls21:21
*** eglynn has quit IRC21:21
*** ndipanov has quit IRC21:21
dansmithI don't understand why he would have broken the stubs with that change, is my point21:21
dansmithso I wonder if it was a knee-jerk reaction, thinking he couldn't without updating the gz files21:21
vishyalexpilotti: this was the original patch:
russellbok, so stubs only have to change when WMI calls change?21:21
dansmithrussellb: right21:22
alexpilottisince rmk didn't change anything except the driver's interface, I don't see how could that be possible21:22
vishyalexpilotti: right but that isn't obvious21:22
russellbwell that sounds better than what i thought21:22
dansmithrussellb: that's what *I* have been trying to get at :)21:22
vishyalexpilotti: I'm saying just add a section to the doc showing what needs to be done21:22
alexpilottiso, here's the patch:
alexpilottihe changed only the line with the parameter21:22
russellbi honestly think this was a case of not everyone understanding what layer the stubs came in21:23
vishyalexpilotti: so you added a lambda to only pass the first param21:23
alexpilottivishy: there's nothing that needs to be done, as those patches are showing21:23
alexpilottivishy: correct21:23
sdagueso I think the issue here is that the hyperv test layer is very different because it doesn't use the stubs model that other drivers use21:23
sdagueand instead uses a different indirection mechanism21:23
alexpilottivishy: to pass the lambda to a utility method in the tests that expects a callable21:23
vishyalexpilotti, sdague: which means an example of that would be useful21:24
sdaguewhich makes it harder for people to match a change in it, because it's different.21:24
alexpilottisdague: it uses also those stubs21:24
russellbso, anything in the works to change how this works?21:24
alexpilottisdague: I understand21:24
alexpilottirussellb: it is21:24
dansmithbut just so we're clear,21:24
dansmiththis stub approach is still required at some layer to verify anything that would normally be calling WMI stuff21:25
russellbmakes sense21:25
sdagueright, that's actually what I'm most interested in, how do we make it better, and not so different21:25
dansmiththe improvement that could/should be made, is so that not all of the testing depends on that being there21:25
*** salv-orlando has joined #openstack-meeting21:25
russellb#note stubs only have to change if WMI interaction changes (driver internals)21:25
sdaguebecause there is a huge mental cost in it being different in the tree21:25
alexpilottiI discussed this in SD and we came up with the idea to replace the stubs with serialized json instead of pickled files as a first step21:25
russellb#action alexpilotti to add an example to the README of making a driver interface change and updating the tests to reflect it21:26
*** eglynn__ has joined #openstack-meeting21:26
dansmithrussellb: I don't think that makes sense21:26
alexpilottirussellb: but, if there' snothing to change, what example can I add? :-)21:26
openstackRemoving item from minutes: <ircmeeting.items.Action object at 0x2179a90>21:26
*** eglynn_ has quit IRC21:26
dansmithrussellb: there's nothing different about making that sort of change in hyperv vs. libvirt21:26
dansmithwhen you think about this,21:26
dansmiththey are testing *more* than the other drivers,21:26
dansmithbecause they're actually simulating the _libvirt.so21:27
dansmithit's just that they don't have a FakeHyperV to sit between there,21:27
dansmithso that most stuff gets tested against FakeHyperV (like FakeLibvirt)21:27
*** ndipanov has joined #openstack-meeting21:27
dansmithand then just use the stubs to test the interaction with WMI21:27
russellbyeah, ok, i at least understand what's happening in the tests better now21:28
dansmithI've modified their driver interface several times,21:28
dansmithwith no ill-effects, or hand-editing WMI pickled gz files :)21:28
alexpilottithis type of tests have some great advantages21:28
russellbany other questions/actions/notes before we wrap this topic up?  have a number of other things to hit21:28
russellbi'm not sure that we came to any sort of conclusion necessarily, but hopefully some folks understand it better (I do)21:29
vishyit appears a little different to me21:29
alexpilottiadding the json files will let anybody edit manually the stubs21:29
dansmithalexpilotti: that is a good thing, IMHO21:29
vishyusually with the other drivers you go in and add the extra param to the interface21:29
dansmithnot enough, because I think you still need a FakeHyperV, but better21:30
sdaguealexpilotti: yeh, I agree with dansmith, we really need a FakeHyperV layer in the tree21:30
alexpilottivishy: that's what has been done here as well!21:30
sdaguebecause I don't think it's obvious to anyone that adding lambdas would have been the fix :)21:30
vishyalexpilotti: fair enough, i think it is just the _test method that is confusing21:31
alexpilottisdague: adding the lambda was just a "styly" way to solve this with two lines21:31
dansmithalexpilotti: it was the wrong way, IMHO21:31
alexpilottithe alternative, would have been an "if" in the tests21:31
*** ewindisch has joined #openstack-meeting21:31
dansmithalexpilotti: and I know that, because I don't understand what the right way is yet21:31
dansmithbut regardless,21:31
dansmiththis is really not indicative of the WMI snapshots keeping people from making interface changes21:32
vishyok it isn't as confusing as i expected21:32
alexpilottivishy: the "_" was marking it as private21:32
vishy_test_vm_state_change doesn't accept extra parameters21:32
alexpilottivishy: teh last parameter is a callable21:32
alexpilottithat method is there because there are a gazillion of tests that look the same:21:32
alexpilottisuspend, resume, start, etc etc21:33
vishythey all just pass the instance data21:33
alexpilottithey all have the same signature21:33
alexpilottithey all HAD the same signature :-)21:33
alexpilottiuntil the resume got that extra param21:33
dansmithso, CELLS, huh?21:33
alexpilottiso, since tha actual method to test was passed as a callable21:34
russellbcomstud: still around?21:34
russellb#topic cells21:34
*** openstack changes topic to "cells (Meeting topic: nova)"21:34
russellbcomstud: what's up21:34
comstudhad some other things to take care of the last week... but I'm about done moving some stuff around with the main cells code21:34
*** devananda has quit IRC21:34
comstudI'm hoping tomorrow is the day for updated reviews21:34
russellbcool, i'm going to make myself go heads down in that code once you update it ...21:34
comstudyep, appreciate it21:35
*** devananda has joined #openstack-meeting21:35
comstudshould be less ugly21:35
russellbwho else is going to review?21:35
comstudand i've noticed some things that weren't structured correctly with respect to pluggable communication that I meant to have21:35
comstud(rpc vs something else021:35
comstudso that'll all be fixed and should hopefully be easier to understand21:35
*** cp16net|away is now known as cp16net21:35
russellbwell i'll be watching for the updates ...21:36
comstudmight be late tomorrow :)  or Saturday.. depends on how much I partcipate in this meetup:)21:36
*** torandu has quit IRC21:36
sdagueso there was still the open discussion of the cells adds to nova-manage21:36
comstudah right21:36
*** vipul|away is now known as vipul21:36
*** enikanorov has quit IRC21:36
comstudmy argument is that it seems silly to have to start an unconfigured service to configure it.21:36
comstudi wouldn't want to start something that's unconfigured21:37
russellbs/nova-manage/nova-bootstrap/  ?  :)21:37
comstudbut honestly I don't really care.21:37
comstudI have the code in an extension as well21:37
russellb(i'm not really suggesting renaming it)21:37
sdaguecomstud: could things go in a config file? or are they too dynamic?21:37
comstudwell, i mean, i do care.  i'm just not going to fight hard over it21:37
comstudbetter things to worrya bout :)21:38
comstudsdague: it's a lot to configure for a .conf file21:38
comstudeach parent cell21:38
comstudeach child cell21:38
*** enikanorov has joined #openstack-meeting21:38
comstudthe Rabiit broker credentials for each21:38
sdaguecomstud: does it tend to be dynamic? or staticly defined?21:38
russellbhow is it too much for a config file, but not too much for a bunch of calls to nova-manage ?21:38
comstudConfigParser doesn't really work well for this21:38
vishycomstud: you don't have to start the cells service do you? You can just start nova-api and configure first?21:39
comstudi'm not sure what i'd name the config options21:39
comstudyeah, nova-api21:39
vishycomstud: I have secret plans to make db-sync into an extension21:39
comstudnot a secret anymore!21:40
vishynova-api can currently start without the db migrations21:40
vishycomstud: doh21:40
comstudcan/can't ?21:40
*** eglynn__ has quit IRC21:40
russellbdid we lose you guys?21:41
sdaguecomstud: if it's mostly static, what about a json file to config it?21:41
comstudi guess it can.. it's just hosed until the DB is upgraded21:41
vishy* can21:41
rmksorry guys I am here now21:41
rmkI was tied up in meetings21:42
comstudif you add a new column in the model code21:42
comstudbut don't update the DB...21:42
dansmithrmk: we voted you as the new hyperv maintainer21:42
comstudnova-api starts spewing failures21:42
comstuduntil the DB is upgraded21:42
comstudcuz sqlalchemy starts querying those new columns21:42
*** ayoung has joined #openstack-meeting21:42
rmkdansmith: Oh cool!21:42
*** eglynn__ has joined #openstack-meeting21:43
comstud(which is annoying that sqlalchemy explicity asks for the columns by name for *)21:43
rmkI feel bad for people using my newly inherited driver21:43
russellb#note give your opinion on the cells nova-manage additions on the associated -dev list thread21:43
russellbsdague: thanks21:43
russellbanything else, or can we move on?21:44
comstudi'm done21:44
russellbdevananda: around for a quick baremetal chat?21:44
russellbor anyone else that has been working on it21:44
*** sandywalsh has quit IRC21:44
* dansmith resists the urge to say something sarcastic21:44
russellbskip it then21:45
russellb#topic project name mapping21:45
*** openstack changes topic to "project name mapping (Meeting topic: nova)"21:45
russellbvishy: take it away sir21:45
*** dolphm has quit IRC21:45
vishythere are many cases in nova where we know the project_id / tenant_id of an object21:46
vishybut we have no way of telling the name21:46
vishywhich forces users / services to go look them up in keystone21:46
vishysince humans generally use names vs long uuid strings21:46
vishyI wanted to see how people felt about keeping a mapping of names to ids in nova21:46
russellbwhere do we need the name in nova?21:46
vishywell there are a number of places where it would be useful21:47
vishynova list --all-tenants21:47
vishynova list --all-tenants --tenant= <name>21:47
vishyfor example21:47
dansmithwhere are the names kept, keystone?21:48
vishyI would also like our default dhcp hostnames to include tenant name21:48
jog0what about doing the lookups in the nova client?21:48
vishywe get the name in the context for every api command21:48
dansmithcouldn't the cli query the list first?21:48
russellbi was about to ask about novaclient, yeah21:48
vishydansmith: assuming that the nova admin is also a keystone admin yeah21:48
comstudshould a nova admin who is not a keystone admin see the mapping tho?21:48
vishyunless keystone decides to make the mapping public21:48
dansmithis it an information leak if they're not and we give them that info?21:48
comstuddansmith: +121:48
vishyI'm mainly coming from a usability perspective21:49
*** dwcramer has joined #openstack-meeting21:49
vishythe name is far more useful imo21:49
dansmithdefinitely nice from the dhcpd hostnames thing tho21:49
rmkThis is a general problem with how decoupled keystone is from everything.21:49
rmkIt's not just tenant names.21:49
sdagueit seems a little weird to dup the data though21:49
comstudi don't really mind about the idea in general.. but it can be a decent sized table for large OS deployments.21:50
vishyif we add it to nova it would just be more like a cache.21:50
rmkOne problem we frequently encounter and have built custom scripts around is reconciling tenant deletion with reclaiming resources21:50
rmkIt's probably a different class of issue than this though21:50
sdaguevishy: right, but then we need a cache management for it21:50
*** eglynn__ has quit IRC21:50
*** eglynn has joined #openstack-meeting21:50
vishyi recognize that it isn't exactly an easy addition21:50
rmkAre we talking about regularly pulling tenant lists from keystone to have an id:name mapping?21:51
comstudplease no21:51
vishythat's why I'm bringing it up21:51
comstudnot a regular pulling21:51
comstudof everything21:51
vishyno, cache would be populated via normal nova requests21:51
comstud(that would not scale)21:51
rmkYeah I wasn't proposing an architecture21:51
vishyyou could theoretically have an external service to resync every so often (or if a tenant name changes in keystone)21:51
sdaguecan we solve the information leak with getting keystone to return only the mapping data the user could see? I guess I don't see how it's an information leak if we can get it from keystone, but not if we can get it from nova.21:52
comstudyou'd have to cache permissions in table along with the tenant name21:52
rmkI've got an environment with over a hundred tenants and I can't tell you what a nightmare it is to deal with the lack of keystone association21:52
sdaguevishy: sweet, you created active directory!21:52
comstudi'd think21:52
dansmithsdague: that doesn't help the --all-tenants thing21:52
vishyif the solution here is to just make keystone expose an api for getting a mapping21:53
rmkvishy: yes21:53
vishyi guess i can handle that21:53
jog0what about bringing this to the -dev ML so the keystone guys can get involved in the discussion?21:53
rmkIt would be nice to make a single bulk request of which IDs need to resolve, and get a single response back with the mapping21:53
vishyhorizon must be doing this already21:53
dansmithjog0: easier to make the decision with out them,... duh21:53
rmkvishy: It does it very inefficiently21:53
sdaguermk: agreed, a bulk operation would be a must21:53
rmkEach mapping is a request and it's terrible21:54
sdagueso it's one wire burst21:54
*** cbaeseman has quit IRC21:54
rmkHorizon as it exists today does not scale21:54
vishyok lets take that to keystone21:54
sdagueso it would help horizon as well, which should make it even more valuable21:54
comstud+1 on bulk21:54
rmkThere's a lot of requests which need to be made bulkable in keystone, this is one of them21:54
russellbvishy: you going to start a thread on it?21:54
russellb#action vishy to start a thread on this topic on the -dev ML so that keystone devs can get involved, too21:54
russellb#topic grizzly-2 status21:55
*** openstack changes topic to "grizzly-2 status (Meeting topic: nova)"21:55
vishyi expect I will need to cache for performance, but I can just do that in memory21:55
russellbgrizzly-2 is still a ways out, but it's our next milestone21:55
jog0if nova is going to cache the mapping how will it handle deletions?21:55
russellbscheduled for January 10th21:55
vishyjog0: just a performance cache, deletions shouldn't matter21:55
rmkvishy: I don't know if you'd cache the response in Nova or just expect keystone to be sane about caching itself so you can just ask it again21:55
russellbvishy: any grizzly-2 planning stuff you want to hit?21:55
vishyrussellb: not really21:56
rmkHaving every service maintain its own cache of the others data seems like a recipe for pain.21:56
jog0vishy: if a tenant is renamed or deleted  nova won't know right away right?21:56
russellbcool ... so ... keep hacking!21:56
rmkjog0: By right away, you mean never21:56
russellb#topic Open Discussion21:56
*** openstack changes topic to "Open Discussion (Meeting topic: nova)"21:56
russellb4 minutes left21:56
jog0rmk: yes21:56
rmkNothing gets reclaimed21:56
vishyjog0: The cache would be temporary just to avoid making requests for every instance launch21:56
rmkSo if you have network associated to deleted tenants, its up to you to clean them up21:56
vishyjog0: like only one request every 5 minutes per node or something21:57
jog0vishy: I like rmk 's idea about keystone doing sane caching itself instead of us dealing with  it21:57
alexpilottisdague: can I ask you to remove the -1 from ? unless there are other reasons for teh -1 of course :-)21:58
sdaguealexpilotti: I'll do you one better21:58
dansmithsdague: technically that's three better21:58
sdagueoh, last item for open discussion21:59
alexpilottisdague: tx ;-)21:59
sdaguetempest gate is close to being ready, but even when it's successful there are a lot of nova stack traces21:59
uvirtbotLaunchpad bug 1079210 in nova "Successful full gate jobs show ERRORs and stacktraces" [Medium,Confirmed]21:59
russellbyeah i was just looking at those21:59
sdaguethose are mostly real bugs, the 413 explosion that I fixed a couple weeks ago was one of those22:00
russellbthere are some that are just noise in the log that we need to silence22:00
russellbwe often pass back exceptions over rpc to the caller22:00
russellband we log every single one on the manager (server) side22:00
sdagueso more eyes on those would be great, as the hope was to get the tempest gate to also check for exceptions in the logs, and fail if they were found22:00
sdaguebut obviously can't do that until those get cleaned up22:00
russellbbut for most cases, we really shouldn't, and we should let the client side decide what to do with it22:01
*** patelna has joined #openstack-meeting22:01
russellbmakes sense22:01
sdagueanyway, more eyes would help22:01
*** jhenner has joined #openstack-meeting22:01
*** danwent has quit IRC22:01
russellbalright, we're a bit over22:02
russellbthanks everyone!22:02
*** openstack changes topic to "OpenStack meetings || Development in #openstack-dev || Help in #openstack"22:02
openstackMeeting ended Thu Nov 29 22:02:06 2012 UTC.  Information about MeetBot at . (v 0.1.4)22:02
openstackMinutes (text):
*** markmc has left #openstack-meeting22:02
*** Mr_T has left #openstack-meeting22:02
jog0for anyone staying around for the db meeting22:03
jog0lets take a 7 min break22:03
*** Vek has left #openstack-meeting22:03
jog0so start at :10 after22:03
russellbdripton_: hey22:03
russellbdripton_: didn't get a chance to sync up with you before that meeting22:03
russellbdidn't see the patch up though, right?22:03
*** jkyle has joined #openstack-meeting22:03
*** jrodom has quit IRC22:04
jkyleis there an announce list for team meetings?22:05
dripton_russellb: my alembic branch is still buggy so wasn't ready to drop on the nova team22:06
russellbdripton_: cool no worries, just wanted to check in22:06
russellbjkyle: that page also has a link to an ical feed22:06
russellbjkyle: and some teams post reminders to the mailing list(s)22:06
jkyleah, must have missed the feed22:06
russellbbut best not to rely on it22:06
russellb(rely on ML reminders i mean)22:07
*** dkehn has quit IRC22:07
jkylecalendar feed sounds perfect22:07
jog0ok lets start22:11
jkyleI must be blind, missing the calendar subscribe22:11
jog0#startmeeting db22:11
openstackMeeting started Thu Nov 29 22:11:11 2012 UTC.  The chair is jog0. Information about MeetBot at
openstackUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.22:11
*** openstack changes topic to " (Meeting topic: db)"22:11
openstackThe meeting name has been set to 'db'22:11
*** krtaylor has quit IRC22:11
russellbjkyle: very top of that wiki page22:11
jog0I am sitting in for devananda who is on a plane22:11
* jkyle got it22:11
jog0who's here?22:12
russellb\o sorta22:12
*** vishy is now known as vishy_zz22:13
jog0thats the agenda22:14
jog0lets start with last weeks action items22:14
jog0were any of the HP guys able to release the db-archive code for mysql?22:15
jog0dripton_: ?22:15
dripton_I didn't hear anything from devananda about that.22:16
*** danwent has joined #openstack-meeting22:16
jog0ok, we will put that back up for next week22:16
jog0#action devananda to find and post existing db-archive code for MySQL22:16
*** dripton_ is now known as dripton22:16
jog0russellb: it looks like you had some more blueprints to create22:17
jog0db blueprints #8 and 9 it looks like  :)22:17
russellboh yes22:17
russellband in fact the patch already went in22:17
russellbcomstud wrote it and i think it's merged22:18
jog0link to the merge?22:18
jog0#topic thread pool22:19
*** openstack changes topic to "thread pool (Meeting topic: db)"22:19
russellbdone!  \o/22:20
russellbso yeah, i'll put a blueprint up for it22:20
russellbso it's on the record22:20
jog0#info done!22:20
jog0#action russellb put up blueprint for record keeping22:20
jog0#topic db-unique-key22:20
*** openstack changes topic to "db-unique-key (Meeting topic: db)"22:20
jog0you had this blueprint creation as well22:21
driptonthat's devananda's BP, right?22:21
jog0I think so, but russellb  volunteered to create it22:21
russellbi did?22:21
* russellb doesn't know what that is22:21
driptonI thought devananda volunteered to create it.22:22
* jog0 going over again22:22
jog0ohh my mistake22:22
russellbbtw, link to non-blocking-db blueprint -
jog0#action devananda to post db-unique-key blueprint22:23
sdagueso the existing db-unique-key review I've seen has issues - because mysql is sloppy :)22:23
sdaguebut postgresql isn't22:23
sdaguedeleted=False is a type error in postgresql unless the column is a boolean22:24
driptonSo that non-blocking-db blueprint is only for mysql. Do we need to add a parallel one for postgres?22:25
jog0ahh devananda did create the blueprint, just hasn't been approved22:25
*** dwcramer has quit IRC22:25
jog0dripton:  lets come back to that in a minute, and finish unique-keys first22:26
jog0sdague: it sounds like we need to keep working on this one22:27
jog0sdague: has no assignee, do you want to take it on?22:28
*** annegentle__ has quit IRC22:28
jog0it also needs approval, russellb22:28
sdaguejog0: yeh, the reality is that False != 0 thing actually caught us other places in tempest trying to get postgresql up22:28
russellbjog0: ack22:28
sdaguejog0: I wasn't actually trying to take it on, I was hoping who ever was working the review would be on so I could help explain things22:28
russellbthoughts on the priority?22:29
driptonit's a prereq for db-archiving, so fairly high priority22:29
*** reed has quit IRC22:29
jog0dripton: I thought doing the unique delete column gets around the prereq part22:30
driptonjog0: ok, I must be mixing up the blueprints22:30
*** reed has joined #openstack-meeting22:30
jog0dripton:  I may be as well, we do have something like 9 bps now22:30
jog0sdague: do know if boris is around?22:31
sdagueI do not22:31
jog0so should we just give him the bp anyay?22:32
driptonI agree with jog0; db-unique-keys is no longer a prereq for db-archiving.22:32
jog0so Devananda, has the bp for now.  Lets move on… or rather back to db pool22:33
jog0#topic db pool22:33
*** openstack changes topic to "db pool (Meeting topic: db)"22:33
jog0dripton: you pointed out the current patch only works for mysql22:33
jog0so do we still need a postrgres solution?22:34
*** eglynn has quit IRC22:34
driptonThe current blueprint only mentions mysql. I haven't tested the patch on postgres.22:34
russellbi looked at the patch, it is indeed mysql specific22:34
*** eglynn has joined #openstack-meeting22:34
jog0would a postgres solution be completely different or can it reuse part of the code?22:35
*** vipul is now known as vipul|away22:36
*** vipul|away is now known as vipul22:36
sdagueI think at this point as long as we don't break postgresql it's ok, people that care can optimize there22:36
*** jrd has quit IRC22:37
jog0sdague:  works for me22:37
*** reed has quit IRC22:37
driptonok, I'll take a look at whether it's easy to port to postgres22:37
driptonIt's hacking around in the mysql driver, and it's possible that the parallel hack is possible in psycopg222:38
jog0#action dripton explore dbp pool for postgres22:38
jog0if you want to pursue that option we can just generalize the current blueprint to cover postgres as well22:38
jog0moving on ...22:38
driptonLet me look first, and if it's practical we can edit the BP22:38
jog0anyone here from db-common?22:39
jog0ewindisch: ?22:39
jog0#topic no-db-compute22:40
*** openstack changes topic to "no-db-compute (Meeting topic: db)"22:40
jog0russellb: it looks like things are moving along nicely22:40
*** afazekas_ has joined #openstack-meeting22:40
russellbmaking good progress, hard to say what percentage we're at22:41
russellbwe've started moving things to nova-conductor, and we're exposing some issues here and there along the way22:41
*** reed has joined #openstack-meeting22:41
russellbhopefully it will speed up once we understand the types of issues we're going to hit while doing it22:41
russellbwe need to be on the lookout for property-style access for db models in the compute code22:41
russellbi know there's more still in there22:42
jog0yeah, I am working on attribute access for my bp as well22:42
russellband they're going to turn into bugs either when we move db access, or you stop model leaks22:42
russellbyeah, saw a patch for that22:42
russellbi guess we just hope for good test coverage to expose it22:42
jog0I am using my PoC  no sqlalchemy leaks to detect the attribute  accesses22:43
russellbvery helpful22:43
jog0I have another patch I am working  right now22:43
jog0what happens to nova-conductor if its not finished by Grizzly release?22:44
jog0as in still db writes from nova-compute?22:44
*** sarob has quit IRC22:44
russellbdepends how far along we are22:45
russellbif we stopped right where we are today22:45
russellbwe'd probably just change the default config so [conductor] use_local=True is the default22:45
russellbwhich would make it so you don't have to run nova-conductor22:45
russellbsince it wouldn't provide much value22:45
jog0with 15 min to go  we have a bunch of BPs to cover so lets move on22:46
jog0#topic db-api-cleanup22:46
*** openstack changes topic to "db-api-cleanup (Meeting topic: db)"22:46
jog0as I just said, this one is moving along nicely22:46
jog0hope to have the whole thing done by g3 but targeting nova.db.api only returning primitives by Grizzly-222:46
jog0leaving db.api consolidation for G322:47
jog0#topic db-archive22:47
*** openstack changes topic to "db-archive (Meeting topic: db)"22:47
jog0dripton: take it away22:47
driptonNot started on db-archive yet.22:47
driptonI'll have something for you next week22:47
jog0can you target it to a milestone22:47
jog0is this G2 or G3?22:47
driptonG3 is safer for now.  I think it's easy enough that we'll probably hit G2, but I don't want to promise yet.22:48
jog0sounds good22:49
*** alexpilotti has quit IRC22:49
jog0#action dripton  target db-archive for G3, with goal of hitting G222:49
jog0#topic db-reconnect22:49
*** openstack changes topic to "db-reconnect (Meeting topic: db)"22:49
jog0it looks like this one hasn't been started either22:49
jog0#action devananda target db-reconnect22:50
jog0#topic db-session-cleanup22:50
*** openstack changes topic to "db-session-cleanup (Meeting topic: db)"22:50
jog0looks like this one is moving along nicely, anyone have any comments?22:50
*** vipul is now known as vipul|away22:51
jog0#action devananda target db-session-cleanup to milestone22:51
*** sarob has joined #openstack-meeting22:51
jog0#topic backportable-db-migrations22:51
*** openstack changes topic to "backportable-db-migrations (Meeting topic: db)"22:51
jog0dripton: any comments on this one22:51
jog0you talked about working on it and alembic last time22:52
driptonI have a work in progress alembic conversion but have not submitted to the dev list yet.22:52
driptonI want to make sure it passes all tests first since it will be controversial.22:52
driptonIt's currently failing a couple.  When it works I will send mail to openstack-dev asking for comments.22:52
jog0ETA on that?22:52
driptonMonday of next week22:53
jog0#action dripton prepare alembic patch to discuss on openstack-dev22:53
*** mtreinish has quit IRC22:53
jog0I think that brings us to the end22:55
jog0#open discussion22:55
jog0#action open discussion22:55
openstackRemoving item from minutes: <ircmeeting.items.Action object at 0x2351dd0>22:55
jog0#topic open discussion22:55
*** openstack changes topic to "open discussion (Meeting topic: db)"22:55
jog0anything else?22:55
driptonI guess we're done.22:57
jog0thank you everyone22:57
driptonthanks dog022:57
*** openstack changes topic to "OpenStack meetings || Development in #openstack-dev || Help in #openstack"22:57
openstackMeeting ended Thu Nov 29 22:57:24 2012 UTC.  Information about MeetBot at . (v 0.1.4)22:57
openstackMinutes (text):
*** dripton has quit IRC22:57
*** dripton has joined #openstack-meeting22:58
*** almaisan-away has quit IRC22:59
*** almaisan-away has joined #openstack-meeting23:01
*** almaisan-away is now known as al-maisan23:01
*** vishy_zz is now known as vishy23:02
*** colinmcnamara has joined #openstack-meeting23:02
*** vipul|away is now known as vipul23:03
*** newptone has joined #openstack-meeting23:04
*** devananda has quit IRC23:07
*** devananda has joined #openstack-meeting23:08
*** dhellmann has quit IRC23:12
*** markmcclain2 has quit IRC23:16
*** andrewbogott is now known as andrewbogott_afk23:17
*** al-maisan is now known as almaisan-away23:18
*** vipul is now known as vipul|away23:18
*** eglynn has quit IRC23:23
*** vipul|away is now known as vipul23:23
*** annegentle_ has joined #openstack-meeting23:25
*** reed has quit IRC23:26
*** annegentle_ has quit IRC23:26
*** lifeless has quit IRC23:27
*** ryanpetrello has quit IRC23:29
*** ryanpetrello has joined #openstack-meeting23:30
*** jrodom has joined #openstack-meeting23:32
*** jrodom has quit IRC23:35
*** reed has joined #openstack-meeting23:38
*** mattray has quit IRC23:40
*** Gordonz has quit IRC23:41
*** newptone has quit IRC23:42
*** cp16net is now known as cp16net|away23:44
*** dkehn has joined #openstack-meeting23:46
*** ryanpetrello has quit IRC23:46

Generated by 2.14.0 by Marius Gedminas - find it at!