Friday, 2013-09-06

« Thursday, 2013-09-05 Index Saturday, 2013-09-07 »
*** oz_akan_ has joined #openstack-marconi00:51
*** nosnos has joined #openstack-marconi00:53
*** oz_akan_ has quit IRC00:55
*** amitgandhi has quit IRC01:00
*** oz_akan_ has joined #openstack-marconi01:02
*** amitgandhi has joined #openstack-marconi01:08
*** kgriffs_afk is now known as kgriffs01:33
*** kgriffs is now known as kgriffs_afk01:34
*** ayoung has joined #openstack-marconi01:51
*** oz_akan_ has quit IRC02:00
*** amitgandhi has quit IRC02:09
*** oz_akan_ has joined #openstack-marconi03:11
*** oz_akan_ has quit IRC03:15
*** torgomatic has quit IRC04:15
*** russell_h has quit IRC04:17
*** torgomatic has joined #openstack-marconi04:18
*** russell_h has joined #openstack-marconi04:20
*** jergerber has joined #openstack-marconi04:49
*** jergerber has quit IRC07:35
*** flaper87|afk is now known as flaper8707:38
openstackgerritFlavio Percoco proposed a change to stackforge/marconi: Pull actual tests out of marconi/tests  https://review.openstack.org/4447509:42
openstackgerritFlavio Percoco proposed a change to stackforge/marconi: Implement small http client for tests  https://review.openstack.org/4526709:42
openstackgerritFlavio Percoco proposed a change to stackforge/marconi: Implement embedded marconi-server execution  https://review.openstack.org/4474909:42
openstackgerritFlavio Percoco proposed a change to stackforge/marconi: Remove util packages  https://review.openstack.org/4447609:42
openstackgerritFlavio Percoco proposed a change to stackforge/marconi: Use oslo.config for functional tests  https://review.openstack.org/4447709:42
openstackgerritFlavio Percoco proposed a change to stackforge/marconi: Run functional tests under tox  https://review.openstack.org/4472309:42
flaper87erm :D09:46
flaper87gotta do something to keep this channel alive when Atlanta is sleeping09:46
*** vkmc has joined #openstack-marconi10:47
*** nosnos has quit IRC11:29
*** tedross has joined #openstack-marconi12:00
*** oz_akan_ has joined #openstack-marconi12:06
*** oz_akan_ has quit IRC12:13
openstackgerritA change was merged to stackforge/marconi: fix(transport.wsgi): JSON home doc missing GET and HEAD hints  https://review.openstack.org/4534112:28
openstackgerritFlavio Percoco proposed a change to stackforge/marconi: Move functional tests into wsgi/v1  https://review.openstack.org/4542313:02
openstackgerritFlavio Percoco proposed a change to stackforge/marconi: Move functional tests into wsgi/v1  https://review.openstack.org/4542313:03
*** oz_akan_ has joined #openstack-marconi13:11
*** oz_akan_ has quit IRC13:11
*** oz_akan_ has joined #openstack-marconi13:12
*** amitgandhi has joined #openstack-marconi13:48
*** kgriffs_afk is now known as kgriffs13:52
openstackgerritA change was merged to stackforge/python-marconiclient: Revert "Add apiclient library"  https://review.openstack.org/4297613:57
flaper87kgriffs: thnx for the review on the test patches14:03
flaper87kgriffs: I addressed your comments14:03
*** oz_akan_ has quit IRC14:03
flaper87kgriffs: btw, GOOOD MORNING!14:03
*** oz_akan_ has joined #openstack-marconi14:05
oz_akan_baack14:05
flaper87kgriffs: also, what do you think about letting this patch land https://review.openstack.org/#/c/37140/ ?14:05
flaper87oz_akan_: from where?14:05
flaper87:D14:05
flaper87oz_akan_: good morning14:05
*** kgriffs is now known as kgriffs_afk14:05
oz_akan_flaper87: hi14:05
flaper87kgriffs_afk: hey, don't you dare to AFK me on my face!!! JUM!14:06
oz_akan_flaper87: from the land of "disconnected for a short while"14:06
flaper87:D14:06
oz_akan_I was asking kgriffs_afk about large queue patch, then got disconnected, the felt the need to write I am back but wrote in the wrong challen14:07
oz_akan_channel14:07
oz_akan_how are you today flaper87?14:07
flaper87oz_akan_: I'm doing well, thanks. You?14:08
oz_akan_great thanks14:08
openstackgerritFlavio Percoco proposed a change to stackforge/python-marconiclient: Implement message controller.  https://review.openstack.org/3714014:09
flaper87oz_akan_: btw, I asked yday if we could have a mongod sharded environment for tests. I think kgriffs_afk spawned a new VM14:10
oz_akan_flaper87: we may, shared by?14:11
oz_akan_sharded by14:11
flaper87oz_akan_: mongodb ?14:11
flaper87the idea is to be able to test queries against both environments14:11
flaper87lemme give you more context14:11
flaper87oz_akan_: https://review.openstack.org/#/c/45299/1/marconi/storage/mongodb/messages.py14:12
oz_akan_my question was, what is your proposal on sharding14:12
oz_akan_hm I see14:12
flaper87Changes like that need to be tested in repl and sharded14:13
flaper87I'd say a 2 shard cluster14:13
flaper871 mongos, 1 config server 2 shards14:13
flaper87no repls on shards14:13
flaper87very simple14:13
oz_akan_I think there must be a switch then, checking marconi conf about mongodb environment then enabling that part of the code14:13
flaper87the purpose is not to test performance but stability14:13
oz_akan_otherwise, don't use read preference14:14
flaper87oz_akan_: there kinda is: https://github.com/stackforge/marconi/blob/master/marconi/storage/mongodb/driver.py#L3914:14
flaper87If that change is harmless in sharded envs, I'm cool with letting it there but, I don't know that and we need to test it14:14
flaper87oz_akan_: but yes, I agree w/ you14:14
*** kgriffs_afk is now known as kgriffs14:15
flaper87there are things that should be enabled depending on the mongodb deployment14:15
oz_akan_I got you so maybe, " conn = pymongo.MongoClient(options.CFG.uri)" case, it won't have any effect and will be harmless14:15
oz_akan_that we need to test14:15
oz_akan_I don't think I can provide that environment as quick as I would like to, we shall keep this as a todo14:16
oz_akan_should we create a bug report just to check this later on?14:16
flaper87oz_akan_: if the vm is up, I think I can help w/ that14:17
flaper87I mean, I don't think I can, I can.14:17
oz_akan_I am not aware of kgriffs VM14:17
flaper87:D14:17
flaper87oz_akan_: kk, lets wait for him to wake up14:17
flaper87he seems to be a bit unsure to whether be afk or not14:17
oz_akan_ok14:18
flaper87:P14:18
oz_akan_:)14:18
kgriffsflaper87: I just replied to ur comment here14:18
kgriffshttps://review.openstack.org/#/c/45299/14:18
flaper87kgriffs: ah, there you are14:18
flaper87:D14:18
kgriffsI'm at a hackday in San Antonio today, so kinda sorta here14:18
flaper87kgriffs: good morning14:18
kgriffs:p14:18
flaper87kgriffs: ah ok, what are you hacking on?14:19
flaper87kgriffs: thanks for the reply, sounds good.14:19
flaper87kgriffs: btw, if you can find a chance to approve, I mean, review the tests patches, It'd be really cool14:19
kgriffsFalcon14:19
kgriffsok, I reviewed some last night14:19
kgriffsI can keep an eye on reviews during the day14:19
oz_akan_guys, I desperately need large queue patch14:20
flaper87kgriffs: I replied to your comments14:20
kgriffskk14:20
oz_akan_https://bugs.launchpad.net/marconi/+bug/121695014:20
kgriffsoz_akan_: blocked by https://review.openstack.org/#/c/45299/14:20
flaper87How long does jury-duty last?14:20
oz_akan_https://review.openstack.org/#/c/44340/14:20
oz_akan_kgriffs: ok14:21
oz_akan_I thought it was merged, https://github.com/stackforge/marconi/commit/c1a564b4444a557cda037e57d7766b7b86661b4a14:22
oz_akan_found this is different14:22
kgriffsflaper87: maybe we can merge the claims race condition fix w/o Alej if we get a few +1s?14:22
amitgandhiflaper87: today should be his last day on the jury14:23
flaper87kgriffs: sounds good to me14:23
kgriffsametts: ^^14:23
flaper87amitgandhi: got it. Thanks14:23
amettsI'll take a look.14:24
kgriffsametts: cool, anyone else around that could also take a quick look?14:27
amettshttps://review.openstack.org/#/c/44340/ has a -1 from flaper87.  Are you two working through this?14:27
oz_akan_I am looking though I have never did this before14:27
kgriffsametts: so, here is the deal (thanks for reminding me discuss)14:28
oz_akan_my first +1, I feel good14:28
kgriffsre expiration filtering, it turns out it's a bad idea, since when I wrote it, I was thinking things would always be sorted in descending order (newest first)14:29
kgriffsbut then I realized (duh) that most of the time we actually sort ASC14:29
kgriffsso, a busy queue with lots of expired messages will end up really never returning anything14:29
kgriffs(since all the oldest messages up to the limit will be expired)14:29
kgriffsso, a few things14:29
kgriffsI commented that generator out and added back a range query on 'e', but kept indexing changes in. My thinking was, Mongo will do the filtering instead of the app, but still without use of the index - hopefully it will be about the same perf.14:31
kgriffs(C++ vs. Python filtering, heh)14:31
kgriffshttps://review.openstack.org/#/c/45307/14:31
ametts+1 on https://review.openstack.org/#/c/45299/   So 44340 will be abandoned, or are we still trying to get it through?14:34
flaper87kgriffs: sounds good to me14:34
flaper87ametts: that patch does other things, IIRC, so, I guess it needs to be updated14:34
kgriffsright, I was going to update it14:34
kgriffsI created the other one for discussion/testing14:34
kgriffs(it will be abandoned)14:35
amettsOkay, +1 on 45307.  I'll hold off on 44340.14:36
openstackgerritFlavio Percoco proposed a change to stackforge/python-marconiclient: Update tox.ini and setup.cfg  https://review.openstack.org/4543914:36
kgriffsflaper87: i can haz teh muhrg?14:40
kgriffshttps://review.openstack.org/#/c/45299/14:40
flaper87kgriffs: done14:41
openstackgerritFlavio Percoco proposed a change to stackforge/python-marconiclient: Implement message controller.  https://review.openstack.org/3714014:41
kgriffsFAN-FREAKIN-TASTIC14:41
* kgriffs is high on excedrin14:41
flaper87:P14:41
openstackgerritA change was merged to stackforge/marconi: fix(storage.mongodb): Race condition when creating a claim  https://review.openstack.org/4529914:44
flaper87Alex_Gaynor: ping14:46
Alex_Gaynorflaper87: pong14:46
flaper87Alex_Gaynor: hey, good morning14:46
flaper87Alex_Gaynor: https://review.openstack.org/#/c/37140/14:46
flaper87Alex_Gaynor: pypy tests didn't run, is there a gate for marconiclient ?14:46
flaper87pypy gate14:46
Alex_Gaynorflaper87: not yet, someone needs to approve https://review.openstack.org/#/c/44104/14:46
flaper87I thought there was14:46
flaper87Alex_Gaynor: cool, thanks14:47
Alex_Gaynorflaper87: and I need to send a review request for config to add it to the gate14:47
Alex_Gaynorflaper87: I can send that review up today if someone approves the review request for marconiclient14:47
amettskgriffs ^14:47
flaper87Alex_Gaynor: mmh, wait14:48
flaper87Alex_Gaynor: I have pypy in my local tox.ini14:48
flaper87why isn't it up there14:48
flaper87I swear I didn't add it manually14:48
flaper87:P14:48
Alex_Gaynor:)14:48
flaper87oh, erm, I know why :D14:48
flaper87Alex_Gaynor: lemme approve that14:48
flaper87Alex_Gaynor: done, thanks!14:49
Alex_Gaynorflaper87: awesome, will get a review request up for config immediately (I'll CC you on it)14:49
flaper87Alex_Gaynor: pls do, thanks a lot!14:50
openstackgerritA change was merged to stackforge/python-marconiclient: Added support for running the tests under PyPy with tox  https://review.openstack.org/4410414:50
Alex_Gaynorflaper87: https://review.openstack.org/#/c/45440/14:51
flaper87Alex_Gaynor: damn, that was fast! :P14:51
Alex_Gaynorflaper87: my local config repo has all the changes just waiting for the tox changes to me approvesd :)14:52
flaper87Alex_Gaynor: aahh, good one! +114:52
kgriffsoz_akan: that 404 patch has been merged, so feel free to pull and test or whatever14:56
kgriffsI will rebase the indexing perf WIP so you can benchmark that14:57
openstackgerritKurt Griffiths proposed a change to stackforge/marconi: fix: Requests get slower when queues have a lot of messages  https://review.openstack.org/4434015:21
openstackgerritKurt Griffiths proposed a change to stackforge/marconi: WIP: MongoDB indexing perf  https://review.openstack.org/4530715:21
kgriffsoz_akan: ^^15:24
kgriffsguys, pls. suggest design summit topics for Marconi here: http://summit.openstack.org/15:35
kgriffswe will have ~4 slots15:35
amettsI have a couple in mind, but I need time to write them up well enough to pass the PTL's scrutiny.15:36
amettsI hear he's really mean.15:37
flaper87ametts: LOL15:37
* kgriffs is very mean15:38
flaper87http://www.youtube.com/watch?v=7RE9iVY9beA15:44
ametts:)15:48
*** malini_afk is now known as malini15:48
amettsmalini is back!15:49
flaper87malini: WELCOME BAAAAAAAAAAAAAAAAAAAACK15:49
malinihelloo!!!!15:49
maliniI am working from orlando today15:50
malinithough not in disney world ;)15:50
maliniheard we got incubated !!!!!!!15:50
flaper87malini: YEEEEESSS!!!15:50
flaper87malini: https://review.openstack.org/#/q/status:open+project:stackforge/marconi+branch:master+topic:refactor-system-tests,n,z15:50
flaper87ops, did I just drop that in here?15:50
kgriffsmalini: w00t15:50
flaper87sooooorrryyyyyyy :D15:50
kgriffsmalini: I sent out a carrier pigeon with a critical message but it wasn't able to find you15:51
maliniflaper87: tht is a lot of reviews :)15:51
kgriffsmalini: I bet you had an awesome time15:51
* kgriffs is jealous15:51
malinikgriffs: I saw a bottle with a message in the beach :D15:51
kgriffssweet15:52
maliniI had a great time, but am glad to be back :)15:52
openstackgerritFlavio Percoco proposed a change to stackforge/python-marconiclient: Update tox.ini and setup.cfg  https://review.openstack.org/4543915:56
openstackgerritFlavio Percoco proposed a change to stackforge/python-marconiclient: Implement message controller.  https://review.openstack.org/3714015:56
kgriffsflaper87: do u need that mongo server today?15:57
kgriffsor is Monday OK?15:57
kgriffsI may get to it today, but no promises. :p15:58
flaper87kgriffs: Monday is fine15:59
flaper87no hurries at all15:59
oz_akan_malini: hi16:00
malinioz_akan_ : hello16:00
oz_akan_malini: how are you?16:03
malinigreat :)16:04
oz_akan_good to hear that16:04
maliniWhat abt you?16:04
oz_akan_great too :)16:04
oz_akan_I need some help16:04
oz_akan_with the tsung16:04
malinisure16:04
oz_akan_in tsung-simepl1.xml test, in scenario-5 we have  <transaction name="delete_message">16:04
oz_akan_I assume it is supposed to delete messages16:05
maliniyes16:05
oz_akan_but after each test run there are lots of messages staying in the queue16:05
malinilet me get to the server16:05
maliniI'll chk tht16:05
oz_akan_I think ti doesn't really delete messages16:05
oz_akan_I see it get 204 though16:05
malinihmmm..16:06
oz_akan_but I am not sure if it is a good 204 or not16:06
oz_akan_since we don't return a different code if message is not found16:06
malinimarconi returns 204 1) if the message is deleted 2)it cannot find the message to delete16:06
oz_akan_flaper87: kgriffs ^^ ;)16:06
kgriffsheh16:06
oz_akan_malini: right, so need tsung really delete messages16:07
maliniLet me chk if tsung scripts have a bug16:07
oz_akan_tahnks16:07
oz_akan_I will be out for lunch after a while for a while16:07
*** ayoung is now known as ayoung-afk16:08
oz_akan_malini:  I will run one last test16:10
oz_akan_pls don't change the xml file16:11
maliniok16:11
oz_akan_actually pls never change it yet, just pls don't run a a test now16:11
oz_akan_I need 5 mins16:11
maliniI wont do anything16:11
maliniwill just look at the xml :)16:11
oz_akan_pls don't even look16:11
oz_akan_kidding :)16:12
malini:D16:12
malinioz_akan_ : Scenario 5 posts 10 messages in each 'post_message' transaction, but we are claiming with '?limit=1' & deleting just the claimed message16:14
maliniThe post_message originally posted only 1 message, but we changed it to 10 to troubleshoot the 204 bug16:14
oz_akan_oh16:14
oz_akan_ohhh16:14
maliniWe can revert post_message to post just one, whenever you are ready16:15
oz_akan_ok test is running16:15
oz_akan_need a few mins16:15
maliniok16:15
oz_akan_you read through tsung16:15
oz_akan_such a wizdom16:15
oz_akan_s16:15
malini:D16:16
oz_akan_done16:17
oz_akan_loop covers delte as well16:17
oz_akan_so I guess you won't change tsung-simple1.xml but another file?16:17
maliniyes, just change the <option name="file_server" id='message' value="/root/.tsung/msgbody_batchpost10.csv"></option>16:18
malinimsgbody_batchpost10.csv has 10 messages in the list16:18
oz_akan_ok, I didn't understand, could u do it :)16:18
maliniUse msgbody.csv instead16:18
maliniboth of these files are in the .tsung directory16:19
oz_akan_got it16:20
oz_akan_let me do then16:20
maliniyeap16:20
oz_akan_despite having such a wisdom and you are also a great teacher16:21
malinihave you been reading art of war or something ?16:21
oz_akan_- know yourself and you will win all battles16:22
oz_akan_:D16:23
oz_akan_running the test again with new setting16:23
oz_akan_kgriffs: now we have test fixed, I am running agains latest master with no and lots of messages, then with your patch with no and lots of messages16:25
oz_akan_the initial testI ran with kind of wrong test, has awful results that I don't even want to share16:25
openstackgerritFlavio Percoco proposed a change to stackforge/marconi: Move Unit tests under a unit package  https://review.openstack.org/4504616:51
*** kgriffs is now known as kgriffs_afk17:03
*** ayoung-afk is now known as ayoung17:25
*** kgriffs_afk is now known as kgriffs17:55
*** russellb is now known as rustlebee18:43
kgriffshmmm pip install using crate.io seems to barf for d2to1, even after removing "restricted" subdomain18:54
*** vkmc has quit IRC19:00
kgriffsflaper87: updated the gist: https://blueprints.launchpad.net/marconi19:32
maliniflaper87: why do https://review.openstack.org/#/c/44475/4/tests/functional/test_claims.py etc. have the #noqa tag ?19:38
openstackgerritKurt Griffiths proposed a change to stackforge/marconi: fix: Requests get slower when queues have a lot of messages  https://review.openstack.org/4434019:51
kgriffsflaper87: https://review.openstack.org/#/c/44340/19:56
openstackgerritKurt Griffiths proposed a change to stackforge/marconi: fix: find_one() running slow (MongoDB)  https://review.openstack.org/4456020:24
openstackgerritKurt Griffiths proposed a change to stackforge/marconi: fix(storage.mongodb): find_one() running slow  https://review.openstack.org/4456020:25
flaper87back21:02
flaper87malini: TBH, I don't remember :(21:02
* flaper87 just saw Iron Man 3 and feels invincible21:03
amitgandhiflaper87: did you see elon musk's youtube video of the ironman like gesture control manufacturing?21:10
amitgandhihttp://www.youtube.com/watch?v=xNqs_S-zEBY21:11
flaper87amitgandhi: I don't think I did, do you happen to have the link?21:11
flaper87ah there it is21:11
flaper87:D21:11
flaper87thanks21:11
amitgandhithat dude is seriously gonna be ironman haha21:11
flaper87duuuuuuuuuuuuuuuuuuuuuuuuuuuude, awesome21:14
*** tedross has quit IRC21:55
malini'21:57
malini;21:57
malinisorry, tht was not really from malini21:59
malinimy kiddo decided to drop in to openstack-marconi22:00
*** oz_akan_ has quit IRC22:00
*** oz_akan_ has joined #openstack-marconi22:01
*** oz_akan_ has quit IRC22:05
flaper87malini: :D22:13
flaper87malini: thanks for those reviews, I'll update the README and submit a new patch22:13
*** amitgandhi has quit IRC22:17
*** kgriffs is now known as kgriffs_afk22:47
*** ametts has quit IRC22:49
*** vkmc has joined #openstack-marconi22:56
*** vkmc has quit IRC22:56
*** vkmc has joined #openstack-marconi22:56
*** oz_akan_ has joined #openstack-marconi23:11
*** oz_akan_ has quit IRC23:17
« Thursday, 2013-09-05 Index Saturday, 2013-09-07 »

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