Tuesday, 2013-10-29

*** malini_afk is now known as malini00:25
*** zyuan_ has quit IRC00:27
*** nosnos has joined #openstack-marconi00:34
*** oz_akan_ has joined #openstack-marconi01:37
*** amitgandhi has quit IRC02:08
*** amitgandhi has joined #openstack-marconi02:09
*** amitgandhi has quit IRC02:13
*** malini is now known as malini_afk02:13
*** vkmc has quit IRC02:20
*** oz_akan_ has quit IRC02:32
*** amitgandhi has joined #openstack-marconi03:09
*** amitgandhi has quit IRC03:13
*** amitgandhi has joined #openstack-marconi03:19
*** ayoung has quit IRC03:20
*** amitgandhi has quit IRC03:23
*** fifieldt_ has joined #openstack-marconi03:25
*** cpallares has left #openstack-marconi03:57
*** amitgandhi has joined #openstack-marconi04:09
*** amitgandhi has joined #openstack-marconi04:10
*** amitgandhi has quit IRC04:15
*** openstackgerrit has quit IRC04:36
*** openstackgerrit has joined #openstack-marconi04:37
*** reed has quit IRC05:02
*** amitgandhi has joined #openstack-marconi05:11
*** amitgandhi has quit IRC05:15
*** amitgandhi has joined #openstack-marconi05:21
*** amitgandhi has quit IRC05:25
*** amitgandhi has joined #openstack-marconi06:11
*** amitgandhi has joined #openstack-marconi06:12
*** amitgandhi has quit IRC06:16
*** ametts has quit IRC06:28
*** ametts has joined #openstack-marconi06:31
*** amitgandhi has joined #openstack-marconi07:12
*** amitgandhi has quit IRC07:16
*** EmilienM has quit IRC07:22
*** EmilienM has joined #openstack-marconi07:25
*** amitgandhi has joined #openstack-marconi08:12
*** nosnos_ has joined #openstack-marconi08:29
*** nosnos has quit IRC08:30
*** flaper87|afk is now known as flaper8708:34
*** nosnos has joined #openstack-marconi09:02
*** nosnos_ has quit IRC09:03
*** amitgandhi has joined #openstack-marconi09:13
*** amitgandhi has quit IRC09:17
*** amitgandhi has joined #openstack-marconi09:23
*** amitgandhi has quit IRC09:28
*** yassine has joined #openstack-marconi09:43
*** amitgandhi has joined #openstack-marconi10:14
*** amitgandhi has quit IRC10:14
openstackgerritFlavio Percoco proposed a change to openstack/python-marconiclient: Implement queue's API methods  https://review.openstack.org/5063810:28
openstackgerritFlavio Percoco proposed a change to openstack/python-marconiclient: Bootstrap Messages support  https://review.openstack.org/5238910:28
openstackgerritFlavio Percoco proposed a change to openstack/python-marconiclient: Add list of required fields to the API definition  https://review.openstack.org/5185010:28
*** nosnos has quit IRC10:31
*** nosnos has joined #openstack-marconi10:32
*** nosnos has quit IRC10:36
*** fifieldt_ has quit IRC10:52
*** amitgandhi has joined #openstack-marconi11:15
*** amitgandhi has quit IRC11:20
*** tedross has joined #openstack-marconi11:23
*** amitgandhi has joined #openstack-marconi11:25
flaper87al-maisan: ping11:27
*** amitgandhi has quit IRC11:29
*** flaper87 is now known as flaper87|afk11:30
*** jcru has joined #openstack-marconi13:08
*** yassine has quit IRC13:11
*** oz_akan_ has joined #openstack-marconi13:11
*** yassine has joined #openstack-marconi13:24
*** yassine has quit IRC13:25
*** yassine has joined #openstack-marconi13:26
*** alcabrera has joined #openstack-marconi13:36
alcabreraGood morning!13:37
openstackgerritAlejandro Cabrera proposed a change to openstack/marconi: feat: integrate shard storage with transport  https://review.openstack.org/5099813:38
*** malini_afk is now known as malini13:42
*** mpanetta has joined #openstack-marconi13:44
*** flaper87|afk is now known as flaper8713:50
flaper87alcabrera: GOOOOOD MORNING!13:51
alcabreraflaper87: Hey! :D13:55
alcabreraI was just making breakfast.13:55
flaper87alcabrera: enjoy your breakfast :D13:56
flaper87I just got to the office, we've PyMilano today13:56
flaper87and we'll discussing monkey patching13:57
alcabreraSounds like an exciting day. :)13:57
flaper87yeah, I just wish I didn't have so many things to do today :(13:57
flaper87I mean, this whole summit thing is consuming a lot of time13:57
flaper87and Friday is holiday here, which means my week will be shorter13:58
flaper87I think I'm the only one that complains because of holidays13:58
flaper87damnit! What's wrong with me!13:58
alcabreraI was rather silly about holidays last year.13:58
alcabreraI remember working right through Thanksgiving and such.13:59
alcabreraMaybe Christmas, too.13:59
alcabreraI worked way too much last year.13:59
alcabreraI'm totally taking off Halloween this week.13:59
flaper87that's the right thing to do, I mean, taking some time off14:01
flaper87I keep saying it but, you know!14:01
*** kgriffs_afk is now known as kgriffs14:01
*** amitgandhi has joined #openstack-marconi14:03
openstackgerritMalini Kamalambal proposed a change to openstack/marconi: Add Tests for non-existing resources  https://review.openstack.org/5369114:04
maliniGood Morning!!14:04
alcabreramalini: morning. :)14:04
*** ayoung has joined #openstack-marconi14:04
*** ayoung has quit IRC14:05
maliniflaper87: thanks for reviewing https://review.openstack.org/#/c/53691/..But I had to update the tests following our 404 vs 204 discussion on y'days meeting..I just submitted the updated patch14:05
*** acabrera has joined #openstack-marconi14:06
flaper87malini: ah see, I was a bit confused while reviewing it, TBH!14:06
flaper87malini: let me take a look, right a way!14:06
flaper87malini: and good morning :)14:07
maliniI totally forgot abt updating the patch after our discussion :-$14:07
flaper87malini: mmh, do we have unit tests for that patch?14:07
malinihmm..this is a test patch14:07
flaper87erm, I mean, for the same functionality14:08
maliniiirc alcabrera talked abt adding some unit tests..14:08
malinidont know if he added it.14:08
flaper87malini: it'd be really cool to have them within the same patch, in case he hasn't added them yet!14:08
malinione of these days I'll get my hands wet on the unit tests14:08
malinisounds like this might be the day :D14:09
acabreraI'm not sure if we have such unit tests. :P14:09
acabreraI'll need to look around.14:09
acabreralemme see...14:09
maliniaahh..u r acabrera today..14:09
*** alcabrera has quit IRC14:09
acabreraYup - because VPN kicked me off. -_-14:09
acabreratehre we go14:09
acabreraNow I can reclaim my name.14:09
*** acabrera is now known as alcabrera14:10
alcabreraAll better.14:10
zyuanalcabrera: https://review.openstack.org/#/c/54206/ do you mind me change logging in shards to use old style formatting?14:12
flaper87zyuan: we're moving everything to the new formatting :/14:13
flaper87or at least, that's the plan14:13
flaper87zyuan: btw, +1 for that patch!14:14
flaper87thanks for working on that!14:14
zyuani've found how logging support new .format() yet14:14
maliniI see a post messae to missing queue https://github.com/openstack/marconi/blob/master/tests/unit/queues/transport/wsgi/test_messages.py#L18314:15
alcabreraI don't expect logging to have a means externally to support .format14:15
zyuanme too...14:15
alcabreraI'm a little surprised about its %s-style feature.14:15
zyuandesigned for lazy logging14:16
alcabrerazyuan: I don't mind changing over to that style, given that it gives us lazy logging.14:16
zyuanok, i think i found how to make .format() work14:16
flaper87oh interesting issue14:17
alcabreramalini: thanks for looking around. :)14:17
flaper87didn't know that!14:17
flaper87zyuan: looking forward to see how that works14:17
alcabreramalini: we would benefit from unit tests that exercised GET on not_exists/messages and not_exists/claims.14:18
malinialcabrera: I just added tht now :D14:18
maliniI am looking around for other stuff tht is worth going in there14:18
alcabreramalini: awesome!14:20
alcabreraflaper87: could I get you to look into this one? https://review.openstack.org/#/c/50815/ (mongodb shard storage) :D14:21
flaper87alcabrera: looking14:21
flaper87alcabrera: Partially-Closes: 1241686 (missing the #)14:22
flaper87I think that's the reason there's no link there14:23
flaper87don't change it14:23
flaper87is not necessary14:23
flaper87I just wanted to mention it14:23
alcabreraI've been forgetting the # a lot lately. :P14:23
flaper87alcabrera: what did it do to you?14:23
alcabreraNo link~14:24
alcabreraI think it might not even trigger the 'Bug closing hook' thing without the #. :/14:24
zyuan.sethmm, no.  logging module only supports old style, because there is only one style can be used in config file, like "%(asctime)s - %(name)s".  the style in hard coded, logging/__init__.py line 46714:28
zyuani will go change the shards part14:28
alcabrerazyuan: thanks!14:28
*** jergerber has joined #openstack-marconi14:29
alcabrerakgriffs: thanks for the reviews! I'll fix up that catalogue driver with the latest feedback. :)14:33
*** ayoung has joined #openstack-marconi14:40
flaper87kgriffs: I addressed your comments in the client patch :)14:41
*** cpallares has joined #openstack-marconi14:43
cpallaresflaper87: hello!14:45
kgriffsgtk: "- Issue #18427: str.replace could crash the interpreter with huge strings."14:50
kgriffsthere is one example why it is important to validate input...14:50
kgriffsI'm not sure about the exact nature (*how* it crashes)14:51
kgriffscould be benign, but the point is, bugs like that happen14:51
*** fvollero has joined #openstack-marconi14:53
fvolleroflaper87: kgriffs : alcabrera : goooooood morninnnnng!14:53
alcabrerafvollero: hey hey!14:54
alcabreraHow go things? :D14:55
alcabrerakgriffs: very gtk14:55
fvolleroalcabrera: y0 y0 :)14:56
*** oz_akan_ has quit IRC14:57
fvolleroalcabrera: normal, if god will, I can go again and do some good @ marconi (client/server/whatever)14:57
*** oz_akan_ has joined #openstack-marconi14:57
alcabrerafvollero: sweet! How is the elasticsearch back end coming along? ;)14:58
*** vkmc has joined #openstack-marconi14:58
fvolleroalcabrera: i'm gonna start from scratch :D15:00
*** openstackgerrit has quit IRC15:00
*** openstackgerrit has joined #openstack-marconi15:01
openstackgerritMalini Kamalambal proposed a change to openstack/marconi: Add Tests for non-existing resources  https://review.openstack.org/5369115:03
openstackgerritAlejandro Cabrera proposed a change to openstack/marconi: feat: add catalogue storage driver for queues  https://review.openstack.org/5216015:04
alcabrerakgriffs: feedback addressed for ^^15:04
alcabrerafvollero: how come? I might have to do the same with my redis backend driver, admittedly. :P15:04
openstackgerritZhihao Yuan proposed a change to openstack/marconi: feat(logging): format msg when log level met  https://review.openstack.org/5420615:05
openstackgerritZhihao Yuan proposed a change to openstack/marconi: feat(logging): format msg when log level met  https://review.openstack.org/5420615:05
*** ayoung has quit IRC15:05
alcabreralots of patches.15:05
alcabreratoday's a patchy day.15:05
zyuanpatchy lol15:05
alcabrerakgriffs, flaper87: are we going to triage BPs come 1600 UTC? Or was it 1700 UTC? :)15:06
alcabreraI'm going to go review *all the patches* now.15:07
openstackgerritCindy Pallares proposed a change to openstack/marconi: Cleaned drivers in proxy and queues config files Closes Bug: 1231669  https://review.openstack.org/5435115:10
flaper87alcabrera: 16 UTC15:10
* flaper87 is back!15:11
flaper87I just did a puzzle (a real one)! :D15:11
flaper87fvollero: I got everything from Emanuela!15:11
alcabreraflaper87: welcome back!15:11
flaper87cpallares: around?15:11
cpallaresflaper87: yep!15:12
openstackgerritZhihao Yuan proposed a change to openstack/marconi: feat(logging): format msg when log level met  https://review.openstack.org/5420615:12
alcabreraflaper87: reviewing your client patches first. :)15:12
flaper87alcabrera: I'm reviewing your patch :D15:13
flaper87I got distracted by some red & white m&ms15:13
flaper87and then I thought about doing the puzzle that was right next to them15:14
flaper87cpallares: awesome, did my suggestion work ?15:14
cpallaresflaper87: I think they did15:14
alcabreraflaper87: mmmm, m&ms...15:16
cpallaresflaper87: They did work, but I was looking at the files and some of them look different... I forgot to pull the latest updates.15:18
* cpallares shakes head15:18
flaper87cpallares: hehee15:19
cpallareshi alcabrera :)15:19
flaper87cpallares: btw, you need to squash those commits15:19
cpallaresidk how to squash commits15:19
flaper87cpallares: otherwise it won't be possible to merge them since it still depends on the previous commits15:19
cpallaresflaper87: so.... it means merge them, right?15:19
flaper87cpallares: you've 2 options and both require rebasing15:19
flaper87cpallares: yeah!15:19
flaper87it's not difficult15:19
flaper87so, commit everything15:20
alcabreracpallares: hey! :)15:20
flaper87wait, before telling you how to squash them, let me tell you how to amend a commit15:20
flaper87unless you already know that15:20
flaper87cpallares: ^15:20
*** ayoung has joined #openstack-marconi15:20
cpallaresflaper87: I'm not sure so please do.15:20
flaper87alcabrera: the ones I just ate had 'Red Hat' written ontop :D15:20
flaper87cpallares: so, to amend a commit you just need: "git ci --amend"15:21
flaper87'git commit --amend'15:21
fvolleroflaper87: glad to hear that15:21
flaper87sorry, I'm so used to my aliases15:21
alcabreraflaper87: all client patches reviewed. Almost all +2'd, except for a typo in the queues one. :P15:21
openstackgerritZhihao Yuan proposed a change to openstack/marconi: feat(logging): format msg when log level met  https://review.openstack.org/5420615:21
flaper87alcabrera: damn, why that one? WHY?15:21
flaper87fvollero: I owe you one, buddy!15:22
fvolleroalcabrera: better to start fresh, in this way there are less probabilities that something get screwed :)15:22
fvolleroflaper87: yep! :P15:22
flaper87cpallares: that will amend the existing commit, which basically means, it'll re-use the last commit for your new changes15:23
alcabrerafvollero: good point, and it's esp. easy to do when it's still being experimentally developed. :D15:23
fvolleroalcabrera: exactly :)15:23
cpallaresflaper87: ooooooh okay, yes, I did not know that.15:23
flaper87cpallares: :)15:23
flaper87cpallares: if you already submitted something for review, and you're addressing the comments in the review, then you need to amend your commit and run git review again15:24
alcabreracpallares: amending commits is wonderful, and having awesome commit messages helps everyone keep track of why we changed things. :D15:24
flaper87it'll add a new patch set to the existing review instead of creating a new one15:24
flaper87cpallares: re commit messages, you may want to read this: https://wiki.openstack.org/wiki/GitCommitMessages15:24
flaper87alcabrera: https://review.openstack.org/#/c/50815/13/marconi/queues/storage/sharding.py15:25
flaper87alcabrera: isn't admin_mode already registered by the transport?15:26
alcabrerathat looks like an oops15:26
alcabreraflaper87: ahh, I see what I did.15:26
alcabreraI remember this from yesterday.15:26
alcabrerakgriffs asked the same question.15:26
alcabreraIt was an oops that I fix in the patch after. :P15:26
alcabreraflaper87: https://review.openstack.org/#/c/50998/12/marconi/queues/storage/sharding.py (fixed here)15:27
alcabreraflaper87: more puzzles and m&ms? :D15:35
flaper87alcabrera: just m&ms and ordering pizza for tonight's meetup15:35
flaper87sooo, glad to know that's already fixed in the next patch. I'm happy with letting it land as-is since that code doesn't break marconi15:35
alcabreracool, cool15:36
kgriffswhat do you think about this design? good? bad? ugly?15:36
flaper87kgriffs: MAN, I love it!15:37
openstackgerritMalini Kamalambal proposed a change to openstack/marconi: Add Tests for non-existing resources  https://review.openstack.org/5369115:37
kgriffsthat's an actual schematic (albeit doctored) from Marconi's 1896 radio receiver15:37
mpanettakgriffs: SWEET15:37
zyuanveeery cool15:37
alcabrerasimple, sweet, marconi15:38
malinikgriffs: tht is cool!15:38
zyuankgriffs: review https://review.openstack.org/#/c/54206/15:39
kgriffsI am having a real graphics guy spruce up the artwork for printing15:39
kgriffszyuan: looking15:39
flaper87alcabrera: +215:40
flaper87alcabrera: btw, I added some comments there that I'd like us to revisit15:40
flaper87they weren't worth a -115:40
alcabreraflaper87: just saw them pop up in my email.15:41
alcabreraI wonder if I've already addressed the AdminControllerBase thing yet...?15:42
openstackgerritA change was merged to openstack/marconi: feat: shards mongodb driver + tests  https://review.openstack.org/5081515:42
alcabreraall that's left before sharding work can continue is (https://review.openstack.org/#/c/50998/) and (https://review.openstack.org/#/c/52160/). Getting close!15:43
openstackgerritFlavio Percoco proposed a change to openstack/python-marconiclient: Implement queue's API methods  https://review.openstack.org/5063815:43
openstackgerritFlavio Percoco proposed a change to openstack/python-marconiclient: Bootstrap Messages support  https://review.openstack.org/5238915:43
openstackgerritFlavio Percoco proposed a change to openstack/python-marconiclient: Add list of required fields to the API definition  https://review.openstack.org/5185015:43
flaper87alcabrera: kgriffs ^^^^^^^15:43
flaper87alcabrera: no idea, re Admin! I'll get to the next shard patch in a bit!15:43
alcabreraoooohh, rebased patches. :D15:43
flaper87alcabrera: those are all brand new15:43
flaper87alcabrera: kididng15:43
* kgriffs eats rebased patches for breakfast15:43
alcabrerakgriffs: that *can't* be good for you. :P15:44
flaper87alcabrera: but it is for us15:44
flaper87alcabrera: so, SHHH!!!15:44
openstackgerritA change was merged to openstack/marconi: feat(logging): format msg when log level met  https://review.openstack.org/5420615:44
zyuanmalini: ping15:44
* alcabrera tastes a rebased patch - grimaces.15:44
* flaper87 watches kgriffs review rebased patches instead of having a real breakfast!15:44
kgriffsalcabrera: https://review.openstack.org/#/c/52160/7/marconi/queues/storage/mongodb/catalogue.py15:45
kgriffsyou are right, I must have been looking at it cross-eyed. :p15:45
flaper87alcabrera: puzzle: https://twitter.com/flaper87/status/395208060136083456/photo/115:46
kgriffsalcabrera: re https://review.openstack.org/#/c/52160/7/marconi/queues/storage/mongodb/utils.py15:46
kgriffsI don't mind globals so much when they are simple constants15:46
kgriffsand you access them using the module name15:46
alcabrerakgriffs: I'm with you on that. I forgot they existed when I wrote that function, heh. ;)15:47
alcabreraI think I've spent too much free time playing with Haskell. :P15:47
alcabreraflaper87: +2'd 50638 (client queues)15:48
kgriffsalcabrera: one more thought15:49
kgriffsjust realized my thought was invalid15:49
* kgriffs whistles nonchalantly15:49
kgriffsbad branch prediction15:49
alcabreralessee... puzzle...15:50
* kgriffs wonders if anyone gets CPU jokes15:50
alcabreraYou've stalled your pipeline, kgriffs15:50
* flaper87 does! :D15:50
* kgriffs is so glad he isn't alone in the world with his nerdy-a** jokes15:50
kgriffsalcabrera: https://review.openstack.org/#/c/5216015:50
openstackgerritMalini Kamalambal proposed a change to openstack/marconi: Add Tests for non-existing resources  https://review.openstack.org/5369115:50
kgriffsso, is that ready to rock?15:51
mpanettaNerdy jokes ++15:51
openstackgerritZhihao Yuan proposed a change to openstack/marconi: feat(mongo): starts to use aggregation for stats  https://review.openstack.org/5096915:51
zyuanjsut a rebase15:51
alcabrerakgriffs: yup, should be all ready.15:51
alcabreraEven has 20% more tests than last go around. ;)15:51
kgriffswaaaait a minute15:52
kgriffsdidn't you have 3 patches pending?15:52
mpanettaSomeone blew the patch stack!15:52
alcabrerakgriffs: one was merged a moment ago - mongodb impl. for admin API shard registry. :P15:53
* mpanetta is in a silly mood15:53
kgriffsmpanetta: nevermind that. I'm just hacking into the team core.15:53
* kgriffs goes back to smashing the stack15:53
mpanettaThe team isn't multi core?  Wow, time to upgrade. :P15:53
alcabrera(mdb) bt => No stack trace.15:53
kgriffsgdb marconi core15:55
kgriffsalcabrera: HECK JUST FROZE OVER15:55
alcabrera(2 to go - then **achievement unlocked**)15:56
kgriffsit's been a long road...gettin' from there to here.15:57
mpanettaDon't make me decrement you bro.16:00
alcabrerampanetta: lol, nice16:00
* flaper87 drops something here: https://wiki.openstack.org/wiki/Marconi/specs/api/next16:01
kgriffsah, it is 1600 UTC after all16:01
kgriffsmpanetta: do you prefer this instead? https://www.youtube.com/watch?v=p5kcBxL7-qI16:01
kgriffsok, lets spend a few minutes talking about api stuff16:02
alcabrerakgriffs: TNG was my favorite. :D16:03
kgriffsmine too16:03
openstackkgriffs: Error: A meeting name is required, e.g., '#startmeeting Marketing Committee'16:03
kgriffs#startmeeting marconi-api16:03
openstackMeeting started Tue Oct 29 16:03:43 2013 UTC and is due to finish in 60 minutes.  The chair is kgriffs. Information about MeetBot at http://wiki.debian.org/MeetBot.16:03
openstackUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.16:03
openstackThe meeting name has been set to 'marconi_api'16:03
kgriffs#link https://wiki.openstack.org/wiki/Marconi/specs/api/next16:04
mpanettakgriffs: Much better16:04
kgriffs#topic Auto-generate client UUID if not given16:04
kgriffswe discussed this briefly yesterday16:05
*** reed has joined #openstack-marconi16:05
kgriffsI propose giving this a thumbs-down16:05
kgriffsI could be convinced otherwise, if only because other projects don't require it.16:06
*** ayoung has quit IRC16:06
kgriffson the other hand, other projects don't have the echo cancellation thing16:06
kgriffswhich breaks if you auto-generate, so that may cause confusion from users anyway16:06
kgriffsflaper87: thoughts?16:06
flaper87sorry, I was writing an email :D16:07
* flaper87 is a slow email writer16:07
flaper87-1 for auto-generated UUIDs16:07
flaper87let me give my reasoning16:07
* alcabrera listens16:07
flaper871) Just 'some' endpoints will use that16:08
flaper872) I expect users to use marconi with the client library16:08
flaper873) If users are using marconi with CURL then I don't think it'll be hard for them to add an uuid to the call16:08
flaper874) I'm +1 for consistency in the API, as much as possible!16:08
flaper87that's it16:08
kgriffsmakes sense16:09
alcabreraworks for me.16:09
kgriffsI would just add to #1 that we want people to be able to mix and match messaging patterns without having to worry about when UUID matters16:10
*** ayoung has joined #openstack-marconi16:10
flaper87kgriffs: +116:11
kgriffs#info strike out auto-generate client UUID16:11
kgriffs#topic  Clearly define whether client ID is required for every request, and enforce it in the implementation16:11
kgriffscurrently we do not clearly define it in the spec, and it is only enforced in a few places16:11
alcabreraAh, I thought we were enforcing it every where... hmm...16:12
kgriffsI would like to propose always requiring it so that it can be logged and used for analytics16:12
flaper87kgriffs: +116:12
alcabreraYeah, I'm +1 here.16:12
flaper87btw, besides openstackgerrit, is anyone writing this down?16:12
kgriffsre logging, we will need a way to "bind" params to the logger, so it gets included in every subsequent log line. Can oslo.logging do that?16:12
flaper87I can do it if you guys want16:12
kgriffsflaper87: I am making notes on the wiki page16:13
flaper87kgriffs: you rock, I haven't told you that enough!16:13
kgriffsflaper87: we all rock in our own ways. :D16:14
kgriffsfor example, you are great at community building16:14
* flaper87 is great at eating m&ms16:14
alcabreracpallares: thanks for clearing out the duplicated patches from the review queue! Greatly appreciated. :)16:14
* kgriffs suddenly has a craving for chocolate16:14
mpanettakgriffs: The stuff you got me is almost gone, soo good, thank you again!16:15
cpallaresalcabrera: Just glad to help out :D16:15
kgriffs#info Always require client id, and log the id in every log line for tracing, analytics, support, etc.16:16
kgriffs#topic Consider allowing opaque string for client ID rather than UUID (will need to understand what else people want to use?)16:16
kgriffsso... I'm not sure on this one16:16
zyuankgriffs: the door is closed16:17
kgriffslike flaper87 said, it isn't *that* hard to generate one16:17
kgriffsand saying it has to be uuid makes validation easier, and reduces confusion IMO16:17
kgriffsany objections to striking this?16:17
* flaper87 listens to the silence16:18
kgriffs#info keep UUID requirement for Client-ID16:18
kgriffs#topic Remove deprecated "partial results" semantics from message posting16:18
kgriffsI added this one, since it is not longer possible to get partial results from the mongo driver16:19
flaper87you can get partial claims16:19
kgriffsthe only reason to keep it is if we expect other backends to need it16:19
kgriffsflaper87: true, this proposal is just when posting messages16:19
flaper87as for now, I'm not sure about other backends needing it!16:19
flaper87kgriffs: ah ok, sorry! I missed that16:19
zyuani would suggest try optional fields for "other backends"16:19
kgriffsmmm. makes me think of extensions16:20
kgriffsok, so get rid of the notion of a "partial" message post? Any objections?16:20
flaper87kgriffs: would it make sense to investigate other storage technologies?16:20
flaper87before taking a final decision16:21
flaper87mmh, actually16:21
flaper87I think the right quesiton here is: Do we want to allow this?16:21
kgriffsI guess that is what I was getting at as well16:22
kgriffsit does complicate clients16:22
flaper87I'm leaning towards removing partial message posting completely16:22
alcabreraI like the concept of all-or-nothing here.16:23
kgriffsalcabrera: do you think you will have any trouble implementing that with Redis, for example?16:23
flaper87it's different for claims, because one can re-claim stuff16:23
kgriffsSQL db's will have no problem, obviously16:23
flaper87but for messages there's too much to be aware off16:23
alcabrerait shouldn't be any problem with Redis. IIRC, posting all the messages is a single communication with Redis.16:24
flaper87btw, what are we returning if the max_retries is reached in mongodb's backend?16:24
kgriffsand redis will not ack on partial success?16:24
kgriffsflaper87: probably 50316:24
alcabrerakgriffs: turns out I'm doing one communication per message, so... it's not atomic with Redis, either. :P16:25
kgriffs#topic Include Client ID in claim data16:25
flaper87cool, so, voting16:25
flaper87erm, I mean, we all agreed16:26
alcabrerakgriffs: no partial ACK16:26
kgriffsalcabrera: oic. You would need to batch them, then16:26
kgriffsso, client ID in claim data16:26
kgriffsif we do that, then we can include claim info when listing messages16:26
kgriffssomeone mentioned that would be nice for auditing and stuff16:27
kgriffscons are extra storage space16:27
flaper87mmh, agreed, however, the claim does not belong to that specific client16:27
flaper87it was created by that client, though16:27
kgriffsyes, true16:27
flaper87any client holding that same claim id can still consume the claim16:27
kgriffsuuid is how many bytes, zyuan?16:28
zyuanis that helpful?16:28
flaper8734 ?16:28
kgriffszyuan: someone thought it would be16:28
alcabreralen(str(uuid.uuid4())) => 3616:28
zyuanan 128 bit integer16:29
kgriffsin bson it is 16, right?16:29
alcabreraahh, uuid as int16:29
kgriffsso, 16 bytes extra per claimed message. seems reasonable.16:29
alcabrerauuid.uuid1().int - gtk16:29
zyuanin message, it's hex form16:30
kgriffs(plus a little overhead for the field name)16:30
kgriffszyuan: only when serialized, right?16:30
zyuan16 in storage16:30
kgriffsi mean, in mongo we just stick it in the claim doc16:30
zyuan~36 in JSON16:30
kgriffsseems like space wouldn't be a big deal16:30
alcabreraalso gtk - uuid.UUID(uuid_as_str).int works16:31
flaper87and it is 3216:31
flaper87(whithout - )16:31
flaper87which is what we're using, IIRC16:31
kgriffsok, let's tentatively plan this for v1.116:32
flaper87alcabrera: str(uuid) adds -, uuid.hex gives you the real hex repr16:32
zyuanwe use str with -16:32
zyuanwe accept hex only as well16:33
flaper87zyuan: kk, thanks for the hint!16:33
kgriffstwo more things i'd like to try and triage today16:33
kgriffs#topic List claims for a given queue16:34
kgriffsuser request16:34
kgriffsnot sure how useful it is16:34
zyuan0 interests16:34
flaper870 interest16:34
flaper87-1 from me!16:34
zyuanthis is how the conversation wents16:34
flaper87that may work as an admin endpoint or something16:34
flaper87but I don't think it'll be of any help for real messaging systems16:35
alcabrerasounds useful in the context of /stats - # active claims on a given queue.16:35
flaper87also, that will make it difficult to support all kind of storage16:35
kgriffsalcabrera: +116:35
alcabreraso... +1 as an admin endpoint16:35
alcabrera-1 as a user endpoint16:35
kgriffsi think a user could find it useful16:35
kgriffsmaybe use it for autoscaling or something16:36
kgriffs(the count)16:36
kgriffs(not the list)16:36
zyuantoo much, too trikey16:36
flaper87I agree with zyuan here!16:36
mpanettaSounds like a way to introduce a race...16:36
kgriffsyes, the stat would be difficult to do and make fast16:36
kgriffsknowing claimed vs free should be sufficient16:37
flaper87I'd rather use N queues or N messages for autoscalling - or a ratio of those 2.16:37
kgriffsTBH, I think this request came from someone who just wanted to make sure their claim succeeded16:37
alcabrera+1 - message # sounds like it'd the core load metric16:37
zyuanif you get it from claim, then it *is* successed...16:38
kgriffsok, so strike this one?16:38
kgriffs#info do not implement listing claims per queue16:41
kgriffs#info existing stats should suffice for scaling metrics16:41
kgriffs#topic Automatically create queues the first time a message is posted to it16:41
kgriffslast one16:41
kgriffs(last topic)16:41
kgriffsflaper87: this one was your suggestion, right?16:41
flaper87kgriffs: correct!16:42
flaper87I think this one could be expanded a bit, TBH!16:42
zyuani think we are going to have control panel to create queue16:43
zyuannow we are going to see queues created by typos in the panel16:43
flaper87I mean, depending on the storage backend, we may want to stop treating queues as resources16:43
flaper87but that's up to the storage16:43
zyuani know16:43
alcabreraso pros - one less call to post a message (-PUT queue), and cons: a typo creates a queue.16:44
flaper87I just wanted to mention some of the things that are possible16:44
flaper87alcabrera: not sure about the type creates a queue16:44
zyuanand... it can actually decrease 204 vs 404 complains16:44
flaper87I mean, not sure what you mean16:44
zyuanso, if queue is nolonger regarded as a resource16:45
zyuanit's just fine if we don;t return 404 if it's not there16:45
zyuanit's just an attribute of message16:45
alcabreraTo clarify on that con -16:45
flaper87but, mmh, we won't be able to have metadata for that queue16:46
alcabreraIf someone develops an application where they have workers reading from queues/feed, and they have producers that POST to queues/feeed/messages... :x16:46
flaper87if we don't treat it as resource, I mean16:46
flaper87btw, this is v2 material!16:46
flaper87just want to make that point clear!16:46
zyuanyea. i can't quickly think of how this is being implemented, but i suspect that it has some value16:47
kgriffs#info lazy queues would be for v2 if we do it, not v1.116:47
alcabrera+1 for v2 feature16:48
flaper87the value I see in this is the lazyness of the API, it would work mostly as mongodb collections do!16:48
alcabreraI'm rather in favor of it16:48
flaper87plus other things, obviously16:48
flaper87but that's the one in terms of API semantic16:48
alcabreraooohh, idea for dealing with that whole typo issue... :D16:48
alcabreraSo, what if...16:49
alcabreraWe had a query flag16:49
alcabreraThat by derfault allowed lazy creation16:49
alcabreraBut if users preferred the checking behavior (maybe for debug dev?)16:49
alcabreraThat they could set to True16:49
flaper87alcabrera: that could work!16:49
alcabrerasomething like that16:49
flaper87however, I'd let the client and API discovery take care of that16:49
flaper87the API discovery should be ready before this work start16:50
kgriffs#info mitigate typos with a sticy/lazy flag on the operation16:50
flaper87and the client should fully support v116:50
flaper87which would make it easier to migrate it to v216:50
flaper87I hope16:50
* flaper87 hopes16:51
flaper87everybody hopes16:51
kgriffsthat's all folks!16:52
openstackMeeting ended Tue Oct 29 16:52:12 2013 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)16:52
openstackMinutes:        http://eavesdrop.openstack.org/meetings/marconi_api/2013/marconi_api.2013-10-29-16.03.html16:52
openstackMinutes (text): http://eavesdrop.openstack.org/meetings/marconi_api/2013/marconi_api.2013-10-29-16.03.txt16:52
openstackLog:            http://eavesdrop.openstack.org/meetings/marconi_api/2013/marconi_api.2013-10-29-16.03.log.html16:52
kgriffsI will add the minutes to the wiki16:52
* flaper87 just saw Bugs Bunny face!16:52
flaper87glad we did this meeting!16:52
alcabrerawith a daily meeting, we'll yet triage all the BPs and accumulate notes for rationale.16:53
zyuannot all, there are HTTP ones16:54
zyuanmaybe tomorrow16:54
alcabrerazyuan: there's also these: https://blueprints.launchpad.net/marconi16:54
kgriffsi want to let that stew for a bit before creating bps16:54
zyuani mean this https://wiki.openstack.org/wiki/Marconi/specs/api/next16:54
kgriffsin fact, maybe even just wait until after the design session16:55
kgriffsalcabrera: yes, we also need to triage blueprints and bugs after we get through the API things16:55
alcabreraflaper87: how do you feel about tackling another sharding patch today? :)16:56
alcabreraflaper87: oooohh, Tulip/asyncio video, HMMMM16:57
kgriffsflaper87: https://review.openstack.org/#/c/52160/16:59
*** tedross has quit IRC17:00
cpallaresalcabrera: how do I remove those items from the commit? (also I'm not sure how they got there in the first place)17:01
alcabreracpallares: hmmm...17:01
alcabrerafor 'marconi/queues/transport/wsgi/shards.py' and 'tests/unit/queues/transport/wsgi/test_shards.py', you can do 'git rm file'17:03
alcabreraI believe that should take care of that.17:03
alcabreraThen for...17:03
alcabrerait might be enough to do something like 'git mv marconi/common/transport/wsgi/utils.py marconi/proxy/transport/utils.py'17:04
alcabrera*'git rm <file>'17:04
*** alcabrera is now known as alcabrera|afk17:10
*** yassine has quit IRC17:13
*** amitgandhi has quit IRC17:17
*** amitgandhi has joined #openstack-marconi17:18
*** tedross has joined #openstack-marconi17:18
*** mpanetta is now known as mpanetta_lunch17:22
*** flaper87 is now known as flaper87|afk17:22
mpanetta_lunchHey guys, one enhancement, you may want top log any condition that would cause a 500 error...17:22
*** amitgandhi has quit IRC17:23
kgriffsmpanetta_lunch: can u create a bug? https://bugs.launchpad.net/marconi/+filebug17:26
*** amitgandhi has joined #openstack-marconi17:30
kgriffsamitgandhi: ^^^17:42
kgriffsif you guys aren't seeing 500's logged, feel free to create a bug report17:42
amitgandhii may be missing the discussion17:43
kgriffscheck the other channel17:43
*** alcabrera|afk is now known as alcabrera17:52
mpanetta_lunchkgriffs: Yep17:56
*** ayoung has quit IRC17:56
*** ayoung has joined #openstack-marconi17:57
*** mpanetta_lunch is now known as mpanetta18:00
mpanettakgriffs: bug created18:01
alcabrerampanetta: thanks!18:01
mpanettaalcabrera: No problem18:03
amettsalcabrera:  You live on #cloudqueues?18:03
mpanettakgriffs: it won't let me set priority :P18:05
mpanettaI would like to make it high...18:05
mpanettait is impacting debugging issues seen in LA.18:05
*** kgriffs is now known as kgriffs_afk18:08
openstackgerritCindy Pallares proposed a change to openstack/marconi: Renamed "exceptions" module to "errors"  https://review.openstack.org/5414118:28
*** tedross has quit IRC18:32
*** kgriffs_afk is now known as kgriffs18:33
mpanettawelcome back18:36
mpanettaYou are a wanted man18:36
mpanettakgriffs: ^^18:36
amettsTime for one or more of these:  http://3.bp.blogspot.com/-I9Jlv9oUd8I/TxyTGzDhvoI/AAAAAAAAANw/8ArcsgFoIi0/s400/the_greatest_american_hero-show.jpg18:46
mpanettaOh man I used to watch that show as a kid ametts18:50
*** tedross has joined #openstack-marconi18:56
*** tedross has quit IRC19:10
openstackgerritCindy Pallares proposed a change to openstack/marconi: Renamed "exceptions" module to "errors"  https://review.openstack.org/5414119:17
openstackgerritAlejandro Cabrera proposed a change to openstack/marconi: feat: add catalogue storage driver for queues  https://review.openstack.org/5216019:18
vkmcGo cpallares \o/19:18
* alcabrera hopes that Jenkins will approve of cpallares efforts19:19
* cpallares agrees with alcabrera, c'mon Jenkins19:20
vkmccpallares, Do you know about zuul? http://status.openstack.org/zuul/19:22
openstackgerritAlejandro Cabrera proposed a change to openstack/marconi: feat: integrate shard storage with transport  https://review.openstack.org/5099819:23
cpallaresvkcm: no, what is it?19:23
cpallaresvkmc: oooo graphs19:23
vkmcYou can check how is your patch doing with the tests... live! :)19:23
vkmcSo you start celebrating (or flipping tables...) earlier19:24
cpallaresvkmc: haha19:25
alcabreravkmc: hahaha, tables have been flipped for the sake of jenkins in the past. :P19:25
alcabreracpallares: do you know about how to run the checks (pep8, unit tests) locally? :)19:25
zyuancpallares: do you need my help?19:25
zyuani can fetch the patch down, fix it, send19:25
cpallaresalcabreara: no, how do i do that?19:27
alcabreracpallares: you can run 'tox -e pep8' to do style checks, 'tox -e py27' to run unit tests.19:27
alcabreraIf you have mongodb installed locally, as well, you can run a more complete unit test suite by issuing 'MARCONI_TEST_MONGODB=1 tox -e py27'19:28
alcabrerawhere that'll run through the full storage suite using mongodb in addition to sqlite.19:28
zyuanthat only works for `sh`-like shells19:28
zyuanuse `env VAR=1 cmd`19:28
cpallareswhat is pep8? Is py27, python 2.7?19:30
alcabreracpallares: pep8 is both a style standard for python (http://www.python.org/dev/peps/pep-0008/) and a style checking tool (https://pypi.python.org/pypi/flake8 - pyflakes + pep8 checker).19:31
alcabrerapy27 is python 2.7 - yes. :)19:31
kgriffsalcabrera: oooh, you rebased those?19:32
alcabrerakgriffs: yup! Conflicts resolved and all.19:32
kgriffscherry picking works w/o conflicts now19:32
alcabreranp. :D19:33
cpallaresoh and then the py33 is the newest one?19:33
alcabreracpallares: yup! The only downside is that py33 will always fail, because a few of our dependencies don'yt support python 3.3 yet. :(19:34
zyuanyes, but it's not required to pass (and it does not pass, but mostly not our fault)19:34
alcabrerawhich is fine by jenkins.19:34
alcabrerasince we have that gate set in non-voting mode (jenkins won't -1 you for failing py33 tests)19:34
alcabrerathere's also tox -e pypy, which runs against the latest pypy release (if you have it installed and on your path as pypy).19:35
*** amitgandhi has quit IRC19:36
zyuankgriffs: https://bugs.launchpad.net/marconi/+bug/1239834 what is the problem?19:36
alcabrerakgriffs: thanks for the +2s!19:36
*** malini is now known as malini_afk19:39
kgriffsalcabrera: I'm seeing all the shard tests fail19:56
kgriffsconnection error19:56
kgriffsany ideas?19:56
kgriffsI'm about to go spelunking19:56
openstackgerritChad Lung proposed a change to openstack/python-marconiclient: Updated the readme to have valid links  https://review.openstack.org/5440619:57
*** amitgandhi has joined #openstack-marconi19:58
alcabrerakgriffs: hmm... weird.19:58
kgriffsok, that's just odd. running just tests.unit.queues.transport.wsgi.test_shards:ShardsMongoDBTests is OK19:59
*** amit has joined #openstack-marconi19:59
*** amit is now known as Guest6211819:59
kgriffsbut running tests.unit.queues.transport.wsgi causes the thing to show up20:00
kgriffs"connection reset by peer"20:00
kgriffsmust be the mongo connection20:00
*** vkmc has quit IRC20:00
*** Guest62118 has quit IRC20:00
*** vkmc has joined #openstack-marconi20:00
*** vkmc has quit IRC20:00
*** vkmc has joined #openstack-marconi20:00
alcabreraI haven't run into that error before.20:00
alcabreraHowever, I did notice that test_lookup (sharding test, I think?) fails if you only run a subset of the unit test suite.20:01
openstackgerritChad Lung proposed a change to openstack/python-marconiclient: Updated the readme to have valid links  https://review.openstack.org/5440620:02
*** amitgandhi has quit IRC20:03
*** amitgandhi has joined #openstack-marconi20:04
kgriffsyeah, looks like a mongod connection error. I unset MARCONI_TEST_MONGODB for now20:12
alcabreraI'm also working on making the20:22
alcabreraunit test suite for sharding more thorough20:22
alcabreraIt'll probably be inheriting from the storage base, just ensuring that the sharding.DataDriver is used.20:23
alcabrera(and initializes the shard catalogue appropriately)20:23
alcabrerasometimes I look at our unit test suite, and I think, "The way we've put this together is pretty crazy/awesome!"20:23
kgriffsrock on20:29
alcabreraI'm out for the day.20:30
alcabreraMore sharding tomorrow~20:30
alcabreratake care, guys. :)20:30
*** alcabrera has quit IRC20:31
openstackgerritCindy Pallares proposed a change to openstack/marconi: Renamed "exceptions" module to "errors"  https://review.openstack.org/5414120:34
kgriffsfwiw, I'm working on adding cache support to the sharding driver20:36
openstackgerritCindy Pallares proposed a change to openstack/marconi: Renamed "exceptions" module to "errors"  https://review.openstack.org/5414120:40
openstackgerritCindy Pallares proposed a change to openstack/marconi: Renamed "exceptions" module to "errors"  https://review.openstack.org/5414120:45
mpanettakgriffs: Do you know what marconi will do if it can't talk to the backend DB during a queue operation?20:47
kgriffsshould return 50320:48
*** malini_afk is now known as malini20:55
malinimpanetta: ping20:55
mpanettaI'm here20:56
maliniJust saw ur mail..20:56
malinithe tests are in the same repo as marconi code20:56
mpanettaChecking it out now20:58
maliniDo you see my private chat?20:59
mpanettamalini: zyuan wanted me to put in more info in my bug20:59
mpanettaBut I am not quite sure what to put20:59
mpanettaLet me send you a link20:59
malinihmm..we cant put in more info becaues we dont know yet why the reqs are not showing up in the logs :(21:00
mpanettaThat is what I was thinking21:00
mpanettaHere is my bug, with comments: https://bugs.launchpad.net/marconi/+bug/124598721:00
openstackgerritAlejandro Cabrera proposed a change to openstack/marconi: Sync with global requirements  https://review.openstack.org/5441521:02
*** reed has quit IRC21:07
*** jergerber has quit IRC21:07
*** flaper87|afk has quit IRC21:07
*** westmaas has quit IRC21:07
*** cpallares has quit IRC21:08
mpanettamalini: Since you are working with both me and zyuan can you also comment on that bug please?21:12
malinisure..I am just starring at it right now figuring out what to comment :D21:13
zyuanmpanetta: i looked and commentted21:13
mpanettazyuan: Ok.  Ill take a look in a sec21:13
amettsmalini: Do you have a quick way to make a whole bunch of queue creation calls in IAD (see the email thread in the Queues list)21:13
maliniametts: sure..I can run a small tsung test21:14
zyuanmpanetta: starts gunicorn on a terminal, shown me the messy stuff left on the terminal (those did not go to log)21:14
maliniametts: create & delete queus21:14
*** jergerber has joined #openstack-marconi21:14
*** flaper87|afk has joined #openstack-marconi21:14
*** westmaas has joined #openstack-marconi21:14
*** hobana.freenode.net sets mode: +o flaper87|afk21:14
amettsSure, that'd work.  Just wondering if we have a bad node or something.21:14
*** ayoung has quit IRC21:14
mpanettazyuan: Unfortunately I can't do that on our servers :(21:15
mpanettaIf I break away from how prod does things I have no way to know if we are breaking things because we are doing it special or if something else is broken.21:15
*** reed has joined #openstack-marconi21:16
zyuanmpanetta: how the server is started?21:17
zyuanmpanetta: server should has its backup logging other than marconi logging21:18
zyuanotherwise, you just lose crash information21:18
mpanettaThe instance never crashes21:18
mpanettaWe are using uwsgi21:19
zyuanthe "backup logging" can be terminal, or see uwsgi's manual21:19
zyuanuwsgi is a container21:19
zyuanit restartes marconi app when it crashes21:19
mpanettaYeah it is logging to greylog at the moment.21:19
zyuanthen get that log to me21:19
mpanettaYeah, I have not seen any restart messages either :(21:19
mpanettaUm, will have to talk to oz about that21:19
zyuanotherwise, i have nothing to look at...21:20
mpanettaI know :(21:20
mpanettaThat is the problem, I don't think we have anything to show you21:20
mpanettaIcan show you the client side errors21:20
mpanettaBut I can't find any log entries that match the errors on the client in time.21:21
zyuanjust a 500 without traceback...21:21
mpanettaNot even a 50021:21
mpanettaThe 500 only gets logged on the client side21:21
zyuanyea, i mean the client error21:21
mpanettaMalini has that21:21
mpanettaI will in a few mins21:21
mpanettaAfter I get her test set up on my system21:21
*** ayoung has joined #openstack-marconi21:22
maliniametts: just sent the IAD results..It looks ok21:25
amettsGreat -- thanks!21:26
amettsAm I reading this right -- you created and deleted 17 queues?21:27
amettsQueue creation is pretty lightweight, isn't it?  What if we created and destroyed 1,000 queues?21:28
amettsqueues and messages are all just database entries.21:28
mpanettaNo problem :)21:33
mpanettazyuan: I should be thanking you :)21:35
zyuanhmm, not yet...21:36
mpanettahaha ok21:36
*** amitgandhi has quit IRC21:43
*** amitgandhi has joined #openstack-marconi21:44
*** amitgandhi has quit IRC21:48
mpanettazyuan: ping22:14
*** malini is now known as malini_afk22:26
*** oz_akan_ has quit IRC22:29
mpanettakgriffs: I have a surprise for you22:34
mpanettakgriffs: the verify_sha function you wrote uses the tag instead of the commit sha, if there is a tag...22:36
mpanettaJust FYI22:36
mpanettaI would consider this a feature BTW :)22:36
mpanettaIt also converts - to . apparently22:37
*** mpanetta has quit IRC22:41
openstackgerritKurt Griffiths proposed a change to openstack/marconi: chore: Pass cache into drivers  https://review.openstack.org/5444522:51
openstackgerritKurt Griffiths proposed a change to openstack/marconi: feat: integrate shard storage with transport  https://review.openstack.org/5099822:51
*** jcru has quit IRC22:52
*** fifieldt has joined #openstack-marconi23:28
*** jergerber has quit IRC23:31
*** kgriffs is now known as kgriffs_afk23:32
*** oz_akan_ has joined #openstack-marconi23:40
*** oz_akan_ has quit IRC23:44

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!