Wednesday, 2014-06-18

*** pdmars has quit IRC00:01
*** rueben has joined #openstack-trove00:02
*** rueben has quit IRC00:02
*** achampion has quit IRC00:12
*** michael-yu has quit IRC00:35
*** achampion has joined #openstack-trove00:38
*** achampion has quit IRC00:43
openstackgerritOpenStack Proposal Bot proposed a change to openstack/python-troveclient: Updated from global requirements  https://review.openstack.org/9909500:47
openstackgerritOpenStack Proposal Bot proposed a change to openstack/trove: Updated from global requirements  https://review.openstack.org/9627500:48
*** ramashri has quit IRC00:52
*** makkie has joined #openstack-trove00:53
makkiehi00:57
makkieunfortunately, i encountered non-deterministic failures many times in trove-gate...00:57
makkiehttps://review.openstack.org/#/c/99001/00:57
makkiecan i request trove-gate re-check here?00:57
*** rwsu has quit IRC00:58
*** kevinconway has quit IRC01:07
*** mattgriffin has quit IRC01:11
*** grapex has joined #openstack-trove01:15
*** grapex has joined #openstack-trove01:15
*** matsuhashi has joined #openstack-trove01:17
*** mattgriffin has joined #openstack-trove01:22
*** mattgriffin has quit IRC01:22
*** saurabhs has left #openstack-trove01:22
*** michael-yu has joined #openstack-trove01:23
*** tvoran has joined #openstack-trove01:25
*** eghobo has quit IRC01:28
*** michael-yu has quit IRC01:30
*** nosnos has joined #openstack-trove01:31
SlickNikmakkie: Sure thing.01:32
SlickNikmakkie: are you on launchpad?01:32
makkieSlickNik: thank you!01:35
makkieSlickNik: i'm shuichiro-makigaki on launchpad01:35
*** ViswaV_ has joined #openstack-trove01:35
SlickNikmakkie: I'm going to add you to the trove-developers group, so you will be able to kick-off the gate checks yourself.01:36
makkieSlickNik: wow, sounds good, thank you very much!01:37
SlickNikNo problem at all. Thank you for taking the time to make trove better! :)01:38
*** ViswaV has quit IRC01:38
SlickNikDone01:38
makkiei'm interested in DBaaS as database administrator, and want to introduce trove to my company01:41
makkieanyway, thank you!01:41
SlickNikYou're welcome. That's awesome! We're here to help if you need any assistance with Trove-related questions.01:44
SlickNikIf you don't mind me asking, which company are you at?01:45
*** ramishra has joined #openstack-trove01:50
*** tvoran has quit IRC01:50
*** annashe__ has quit IRC01:54
makkieRakuten, which is japanese e-commerce web site and parent company of buy.com, kobo, priceminister and so on01:55
makkiei got emails from launchpad and gerrit. i'm happy to join trove developers group :-)01:57
fifieldtRakuten! すごい :)02:05
*** tvoran has joined #openstack-trove02:05
*** tvoran has quit IRC02:11
*** tvoran has joined #openstack-trove02:11
*** mattgriffin has joined #openstack-trove02:13
makkiehi, tom! basically it's a good company, but database infrastructure and management style are sometimes too legacy...02:19
makkiethen, i believe trove will save me02:20
*** coolsvap|afk is now known as coolsvap02:25
*** matsuhashi has quit IRC02:29
*** matsuhashi has joined #openstack-trove02:32
*** haomai___ has quit IRC02:36
*** achampion has joined #openstack-trove02:44
*** haomaiwa_ has joined #openstack-trove02:45
*** achampion has quit IRC02:54
*** matsuhashi has quit IRC03:00
*** mattgriffin has quit IRC03:14
*** mattgriffin has joined #openstack-trove03:16
*** nosnos has quit IRC03:27
*** michael-yu has joined #openstack-trove03:41
*** michael-yu has quit IRC03:44
*** annashe__ has joined #openstack-trove03:46
*** tvoran has quit IRC03:49
*** eghobo has joined #openstack-trove03:51
*** mattgriffin has quit IRC03:52
*** matsuhashi has joined #openstack-trove03:53
makkiehello again!03:55
makkiethank you for your support. it took long time, but i can pass trove-gate now.03:55
makkiecould you please review my changes?03:55
makkiehttps://review.openstack.org/#/c/99001/03:55
*** juantwo has quit IRC04:00
*** nosnos has joined #openstack-trove04:07
*** matsuhashi has quit IRC04:12
*** matsuhashi has joined #openstack-trove04:12
*** sgotliv has joined #openstack-trove04:15
*** matsuhashi has quit IRC04:17
*** haomaiwa_ has quit IRC04:17
*** haomaiwang has joined #openstack-trove04:18
*** matsuhashi has joined #openstack-trove04:21
*** Longgeek has joined #openstack-trove04:26
*** haomai___ has joined #openstack-trove04:27
*** haomaiwang has quit IRC04:31
*** denis_makogon has joined #openstack-trove04:35
*** matsuhashi has quit IRC04:36
*** matsuhashi has joined #openstack-trove04:36
*** rhodgin has quit IRC04:38
*** mattgriffin has joined #openstack-trove04:38
*** matsuhashi has quit IRC04:39
*** eghobo has quit IRC04:39
*** matsuhashi has joined #openstack-trove04:39
*** eghobo has joined #openstack-trove04:40
*** achampion has joined #openstack-trove04:44
*** achampion has quit IRC04:53
*** ramishra has quit IRC05:00
*** ramishra has joined #openstack-trove05:00
*** ramishra has quit IRC05:03
*** ramishra has joined #openstack-trove05:03
*** matsuhashi has quit IRC05:11
*** matsuhashi has joined #openstack-trove05:11
*** matsuhashi has quit IRC05:16
*** michael-yu has joined #openstack-trove05:16
*** matsuhashi has joined #openstack-trove05:17
*** matsuhashi has quit IRC05:24
*** matsuhashi has joined #openstack-trove05:24
*** grapex has quit IRC05:39
*** juantwo has joined #openstack-trove05:42
*** achampion has joined #openstack-trove05:42
*** juantwo has joined #openstack-trove05:42
*** achampion has quit IRC05:58
*** michael-yu has quit IRC06:03
*** michael-yu has joined #openstack-trove06:04
openstackgerritOpenStack Proposal Bot proposed a change to openstack/trove: Imported Translations from Transifex  https://review.openstack.org/10008306:07
*** michael-yu has quit IRC06:10
*** denis_makogon has quit IRC06:19
*** openstackgerrit has quit IRC06:19
*** cp16net has quit IRC06:19
*** denis_makogon has joined #openstack-trove06:23
*** openstackgerrit has joined #openstack-trove06:23
*** cp16net has joined #openstack-trove06:23
*** mattgriffin has quit IRC06:34
*** miqui has quit IRC06:36
*** grapex has joined #openstack-trove06:40
*** sgotliv has quit IRC06:41
*** juantwo_ has joined #openstack-trove06:48
*** ramishra has quit IRC06:48
*** grapex has quit IRC06:49
*** ramishra has joined #openstack-trove06:51
*** juantwo has quit IRC06:52
*** sgotliv has joined #openstack-trove07:00
*** juantwo has joined #openstack-trove07:06
*** matsuhashi has quit IRC07:08
*** matsuhas_ has joined #openstack-trove07:08
*** eguz has joined #openstack-trove07:08
*** juantwo_ has quit IRC07:09
*** sgotliv has quit IRC07:10
*** sgotliv has joined #openstack-trove07:10
*** eguz has quit IRC07:11
*** eghobo has quit IRC07:12
*** ramishra has quit IRC07:18
*** annashe__ has quit IRC07:18
openstackgerritSergey Gotliv proposed a change to openstack/trove: Fix enable on boot when working with systemd and symlinked units  https://review.openstack.org/10050407:22
*** julienvey has joined #openstack-trove07:25
*** flaper87|afk is now known as flaper8707:33
*** amytron has quit IRC07:38
*** grapex has joined #openstack-trove07:40
*** grapex has quit IRC07:45
*** denis_makogon has quit IRC07:51
*** coolsvap is now known as coolsvap|afk07:55
*** julienvey has quit IRC08:18
*** ramishra has joined #openstack-trove08:18
*** julienvey has joined #openstack-trove08:18
*** haomai___ has quit IRC08:28
makkieSlickNik: thank you for review!08:36
*** boden has joined #openstack-trove08:36
*** denis_makogon has joined #openstack-trove08:38
*** juantwo has quit IRC08:39
*** julienve_ has joined #openstack-trove08:45
*** Longgeek has quit IRC08:46
*** Longgeek has joined #openstack-trove08:46
*** Longgeek has quit IRC08:48
*** Longgeek has joined #openstack-trove08:49
*** julienvey has quit IRC08:49
*** Longgeek has quit IRC08:53
*** Longgeek_ has joined #openstack-trove08:53
openstackgerritDenis M. proposed a change to openstack/trove: Fix poll_until retriever when heat provisioning is enabled  https://review.openstack.org/9719408:58
openstackgerritDenis M. proposed a change to openstack/trove-integration: Add redstack command that enables heat support  https://review.openstack.org/10059809:01
*** matsuhas_ has quit IRC09:01
*** sgotliv has quit IRC09:05
*** matsuhashi has joined #openstack-trove09:09
openstackgerritDenis M. proposed a change to openstack/trove: Add datatore/version name into configuration response  https://review.openstack.org/9898509:11
*** isviridov|away is now known as isviridov09:13
*** ramishra has quit IRC09:16
*** matsuhashi has quit IRC09:38
*** julienve_ has quit IRC09:38
*** julienvey has joined #openstack-trove09:39
openstackgerritDenis M. proposed a change to openstack/trove: Added manager attribute to datastove-version-show response  https://review.openstack.org/10083609:39
*** grapex has joined #openstack-trove09:42
*** grapex has quit IRC09:46
*** ramishra has joined #openstack-trove09:57
*** matsuhashi has joined #openstack-trove10:01
*** matsuhashi has quit IRC10:06
*** annashe__ has joined #openstack-trove10:07
*** matsuhashi has joined #openstack-trove10:08
*** sgotliv has joined #openstack-trove10:08
*** ramishra has quit IRC10:09
*** matsuhashi has quit IRC10:17
*** matsuhashi has joined #openstack-trove10:18
openstackgerritA change was merged to openstack/trove: Reverting deleted tests  https://review.openstack.org/9935310:22
*** makkie_ has joined #openstack-trove10:29
*** makkie has quit IRC10:29
openstackgerritA change was merged to openstack/trove: Fix inheritance for ConfigurationParameterDeleted  https://review.openstack.org/9939110:34
*** Longgeek_ has quit IRC10:58
*** Longgeek has joined #openstack-trove10:59
*** ramishra has joined #openstack-trove11:10
*** Longgeek has quit IRC11:13
*** Longgeek has joined #openstack-trove11:14
*** ramishra has quit IRC11:14
*** denis_makogon has quit IRC11:17
*** setho2 has quit IRC11:17
*** Longgeek has quit IRC11:19
*** Longgeek has joined #openstack-trove11:19
*** sgotliv has quit IRC11:22
*** setho has joined #openstack-trove11:25
*** denis_makogon has joined #openstack-trove11:25
*** number80 has quit IRC11:32
*** annashe__ has quit IRC11:40
*** grapex has joined #openstack-trove11:43
*** grapex has quit IRC11:47
*** matsuhashi has quit IRC11:48
*** matsuhashi has joined #openstack-trove11:51
*** matsuhashi has quit IRC11:56
*** sgotliv has joined #openstack-trove11:56
*** matsuhashi has joined #openstack-trove11:56
*** Longgeek has quit IRC11:57
*** Longgeek has joined #openstack-trove11:58
*** matsuhashi has quit IRC12:01
*** setho2 has joined #openstack-trove12:02
*** setho has quit IRC12:03
*** setho2 has quit IRC12:07
*** denis_makogon has quit IRC12:08
*** ramishra has joined #openstack-trove12:11
*** setho has joined #openstack-trove12:11
*** denis_makogon has joined #openstack-trove12:12
*** ramishra has quit IRC12:15
*** grapex has joined #openstack-trove12:43
*** grapex has quit IRC12:47
*** pdmars has joined #openstack-trove12:48
*** Longgeek has quit IRC12:48
*** pdmars has quit IRC12:50
*** pdmars has joined #openstack-trove12:50
*** tvoran has joined #openstack-trove12:52
*** Longgeek has joined #openstack-trove12:57
openstackgerritTheron Voran proposed a change to openstack/trove: Corrects typo in instance models and service  https://review.openstack.org/10065413:01
*** makkie_ has quit IRC13:10
*** makkie has joined #openstack-trove13:24
*** jcru has joined #openstack-trove13:28
*** amytron has joined #openstack-trove13:29
*** jcru has quit IRC13:34
openstackgerritJulien Vey proposed a change to openstack/trove: Delete undeclared variable in guest-agent API class  https://review.openstack.org/9876613:34
openstackgerritJulien Vey proposed a change to openstack/trove: Delete undeclared variable in guest-agent API class  https://review.openstack.org/9876613:35
makkiehi13:43
peterstacAny core interested in reviewing the replication guestagent APIs?13:43
peterstachttps://review.openstack.org/#/c/91904/13:43
peterstac#link https://review.openstack.org/#/c/91904/13:43
makkieI also need one more core review +2.13:43
makkiehttps://review.openstack.org/#/c/99001/13:43
makkiecould you review my changes? thanks!13:43
makkiewith peterstac ;-)13:44
peterstacmakkie: Oh, it's not my review - just noticed it's been sitting there for almost a week13:45
*** tvoran has quit IRC13:45
peterstacmakkie: And I thought replication was needed quickly ;)13:45
*** miqui has joined #openstack-trove13:46
*** jcru has joined #openstack-trove13:46
*** ramishra has joined #openstack-trove13:51
makkiepeterstac: ah, i see. many need core reviewers... they are busy...13:52
*** ramishra has quit IRC13:56
*** Longgeek_ has joined #openstack-trove13:57
*** Longgeek has quit IRC13:59
*** edmund has joined #openstack-trove14:02
*** nosnos has quit IRC14:03
*** mattgriffin has joined #openstack-trove14:04
openstackgerritOpenStack Proposal Bot proposed a change to openstack/trove: Updated from global requirements  https://review.openstack.org/9627514:11
*** grapex has joined #openstack-trove14:18
*** rhodgin has joined #openstack-trove14:21
*** esmute has quit IRC14:22
*** haomaiwang has joined #openstack-trove14:23
*** esmute has joined #openstack-trove14:25
*** grapex has quit IRC14:25
*** tvoran has joined #openstack-trove14:25
*** grapex has joined #openstack-trove14:26
*** robertmyers has joined #openstack-trove14:26
*** rwsu has joined #openstack-trove14:29
*** ramishra has joined #openstack-trove14:32
*** kevinconway has joined #openstack-trove14:39
*** Longgeek_ has quit IRC14:43
*** Longgeek has joined #openstack-trove14:44
*** Longgeek has quit IRC14:45
*** Longgeek has joined #openstack-trove14:47
*** ramishra has quit IRC14:54
*** denis_makogon_ has joined #openstack-trove15:01
*** makkie has quit IRC15:03
*** juantwo has joined #openstack-trove15:12
*** annashe__ has joined #openstack-trove15:20
*** Longgeek has quit IRC15:20
*** denis_makogon has quit IRC15:28
*** denis_makogon_ is now known as denis_makogon15:28
denis_makogonamrith, ping15:28
*** dmakogon_ has joined #openstack-trove15:28
denis_makogonamrith, need to talk about https://review.openstack.org/#/c/99682/15:29
*** juantwo has quit IRC15:41
*** juantwo has joined #openstack-trove15:42
*** ViswaV has joined #openstack-trove15:45
*** ViswaV_ has quit IRC15:47
*** haomaiwang has quit IRC15:51
*** eghobo has joined #openstack-trove15:52
*** haomaiwang has joined #openstack-trove15:52
*** ViswaV has quit IRC16:07
*** haomaiwang has quit IRC16:08
*** ViswaV has joined #openstack-trove16:11
*** sgotliv has quit IRC16:12
amrithdenis_makogon, hi ...16:17
*** eghobo has quit IRC16:18
amrithalso https://review.openstack.org/#/c/100598/16:18
amrithwassup?16:18
*** eghobo has joined #openstack-trove16:18
*** jmontemayor has joined #openstack-trove16:23
*** jmontemayor has quit IRC16:23
denis_makogonamrith, those variables are shared but not the part of test config, thats why i left them there, but out of CONFIG object16:25
amrithso, I guess my question is this. the variable you are accessing are the same. The change really doesn't change anything16:26
amrithconfig.CONFIG is really the same as config16:27
amrithyes?16:27
denis_makogonand i agree, i can change name of method, https://review.openstack.org/#/c/100598/, but i disagree about disabling heat support seems strange because if user will have any instances, he should delete them before restaring services16:27
amrithlet's talk about one review at a time.16:27
denis_makogonamrith, config.CONFIG can be modified through test_begin.conf16:27
denis_makogonok16:27
amritheffectively, the entire change to run_tests.py changes nothing16:28
amrithyes?16:28
denis_makogonwe need to split up config for fake tests and real tests16:28
amrithso, trove/tests/config.py16:28
denis_makogonrun_tests.py was changed because config.py was changed16:29
denis_makogonthat's all16:29
amriththe last line16:30
amrithor almost the last line16:30
amrithCONFIG = TestConfig()16:30
amrithso, in reality, config.CONFIG is the same as config, yes?16:30
denis_makogonno16:30
denis_makogonconfig.py contains CONFIG object that represents test config that16:31
denis_makogonconfig != config.CONFIG16:31
denis_makogoni guess we have mismatch in terms for now, what do you mean by telling "config" ?16:32
denis_makogonis it service configuration file or test config ?16:32
denis_makogonamrith, are you there ?16:33
amrithyes I am16:33
denis_makogonso, config.py, as you can see, contains a CONFIG object that represents a configuration file, config module is a placeholder for that object and other things16:35
amrithso what you are changing in config.py is to move things from TestConfig.dbaas_datastore and make it just config.dbaas_datastore?16:35
denis_makogonamrith, that's why i left things like "dbaas_datastore_version_id" there as globals16:35
denis_makogonamrith, yes16:36
denis_makogonamrith, ther reason is to make tests datastore-independent16:36
denis_makogonamrith, and those variables are meaningful only for fake-mode tests16:36
amrithI saw that but assumed that this wasn't what you were doing because ...16:36
amrithinstance_create_time, I'd argue that it should be global as well16:37
amrithit does depend on the datastore16:37
amrithsome are fast, others are faster16:37
amrithor mysql_connection_method ...16:37
amrithetc.,16:37
amrithno?16:37
denis_makogonamrith, as you know, integration tests are made in the next way, we're not running all test cases for each datastore, those test groups that was proposed and merged for different datastores are not including those things that you mantioned16:38
denis_makogonamrith,16:39
denis_makogonhttps://github.com/openstack/trove-integration/blob/master/tests/integration/int_tests.py#L223-L24816:39
amrithok, I'll go and read it again16:40
amrithI don't quite understand in that case what you are trying to do in this change but I'll read it again.16:41
denis_makogonamrith, integration tests are also relaying on CONFIG object, that's why i removed those variables out of there, and proposed datastores/version API calls to retrieve datastores/versions, and it'll work for other datastores16:41
denis_makogonamrith, the main goal to break dependency between int-tests and config-defined datastores16:42
denis_makogonamrith, to be able to run tests with different datastores16:42
amrithok, what about https://review.openstack.org/#/c/100598/16:43
*** haomaiwa_ has joined #openstack-trove16:43
denis_makogonamrith, agreed with all comments, but how to act if user will have instances provisioned through heat ?16:44
denis_makogonamrith, or just say, no matter, let's disable heat support ?16:44
denis_makogonif this is the case, i'm fine with that16:45
amrithwhat's the question?16:45
*** ramashri has joined #openstack-trove16:45
amrithright now it appears that with heat the claim is that we should do more than just check instance status; we should check stack status. I'm merely pointing out that the code doesn't do that.16:45
denis_makogonamrith, why is that ?16:46
denis_makogonamrith, have you looked at heat codepath?16:46
amrithwhere?16:46
denis_makogonamrith,16:47
denis_makogonhttps://github.com/openstack/trove/blob/master/trove/taskmanager/models.py#L418-L42216:47
denis_makogonthis poll method will wait until stack is 100% ready16:47
amrithhuh?16:48
amrithare you looking at https://review.openstack.org/#/c/100598/ or something else?16:48
*** haomaiwa_ has quit IRC16:48
amrithyou seem to be talking about https://review.openstack.org/#/c/97194/16:48
amrithok, let's talk about that16:49
amrithmy comment earlier relates to this16:49
amrithright now it appears that with heat the claim is that we should do more than just check instance status; we should check stack status. I'm merely pointing out that the code doesn't do that.16:49
amrithbut I thought we were talking about something else16:49
amrithso ...https://review.openstack.org/#/c/97194/16:49
amrithif we use heat16:49
amriththen we call _check_trove_instance_is_active16:49
amrithdo you agree with the name change?16:50
denis_makogonamrith, status BUILD also valid16:51
denis_makogonamrith, ACTIVE is final status16:51
denis_makogonamrith, method name is 100% valid16:52
amrithI disagree16:52
denis_makogonamrith, why should it be changed?16:52
amrithyou ask, why should it be changed. I think it should be changed because I believe the name does not reflect what the method is doing. the method is checking whether the trove instance is in an active state or a running state. it returns true if it is in a running state16:53
amrithfor example,16:53
amriththe function check_if_a_number_is_even16:53
denis_makogonamrith, i don't see too much difference between valid/active word in function name16:53
amrithshould (one would assume) return true if the number is even16:53
amrithand false if the number is not even16:54
denis_makogonamrith, it also checks if instance is not ERROR/FAILE as well16:54
amrithwhy would I call this check_if_a_number_is_valid?16:54
amrithactually it does't. It checkes NOT NEW or BUILDING.16:54
amrithbut, be that as it might16:55
*** ramishra has joined #openstack-trove16:55
denis_makogonamrith, same words, but you talking nothing but Not New or Building, and avoiding other statuses16:55
amrithwhat do you mean by "same words"?16:55
*** isviridov is now known as isviridov|away16:56
denis_makogonamrith, you said  "It checkes NOT NEW or BUILDING.". I said - it checks that instance is not in ERROR/FAILED16:56
denis_makogonstick with two ends16:57
denis_makogonbut in general - same thing16:57
amrithsorry, error or failed is the check on the nova status. not new or building is the check for trove status.16:57
amrithI don't know what the "stick with two ends" means?16:58
denis_makogonfor short period of time, those method (..._is_valid) checks if instance is in one of three states not New, BUILD, ACTIVE16:58
amrithat any rate, the function_check_trove_instance_is_in_valid_state() is not named correctly; I think.16:58
*** ramishra has quit IRC16:59
denis_makogonfrom my perspective, that method checks if instance is not in invalid state(ERROR/FAILED)16:59
amrithok, we disagree.17:00
denis_makogonamrith, it checks if instance is not corrupted, it means instance is in one of the valid states (NEW, BUILD, ACTIVE)17:00
denis_makogonthat's why it called ..._is_valid17:00
amrithlike I said, we disagree ;)17:01
denis_makogonok, we need to bring it up on a meeting17:01
denis_makogonin open discussion17:01
amrithor someone else can post comments on the review17:01
denis_makogontoo long to wait17:01
denis_makogonneed to grab cores and let them decide17:02
amrithok, sure.17:02
denis_makogonSlickNik, cp16net, grapex, hub_cap, cp16net, amcrn, guys are you around?17:02
amrithI have no idea when they'll be around so I'll post my comments on the bug and they can read them there ...17:03
denis_makogonvipul, hey, are you the?17:03
vipulhey denis_makogon what's up17:03
denis_makogonvipul, have you got like 5 mins to review small patch?17:03
vipulsure shoot me the link17:03
amrithhttps://review.openstack.org/#/c/97194/317:04
denis_makogonvipul, we(i and amrith ) have small disagreement in method naming, https://review.openstack.org/#/c/97194/17:04
amrithvipul, ... https://review.openstack.org/#/c/97194/317:04
vipulyes please17:06
vipulwhoops wrong window17:06
denis_makogonamrith, https://review.openstack.org/#/c/98985/, let's talk about this one17:09
amrithok, let's17:11
denis_makogonamrith, do you agree about not asking backend for another copy of a configuration group?17:12
denis_makogonamrith, and just passing already existing copy17:13
amrithare you talking about models.py@12817:13
denis_makogonyes17:14
amrithwhat's the first parameter "context" being used for?17:14
amrithin this method?17:14
denis_makogonit's common for models methods17:14
denis_makogonso, i decided to keep it as it done in lots of models17:15
amrithis context used in load_configuration_datastore_and_version?17:16
amrithyou pass it as parameter 117:16
amrithI'm asking why17:16
amritham I missing something that needs context in that case?17:16
*** thedodd has joined #openstack-trove17:16
denis_makogonit's not used17:16
denis_makogonok, i'll remove it17:16
amrithno, not so fst17:17
amriths/fst/fast/17:17
amrithwhy not make load_configuration_datastore_version and and the other have simialr signatures17:17
amrithand you can even reuse some code17:17
amrithif you pass context and id, you can make the code a little cleaner17:17
amritheither would work for me.17:18
amrithI just didn't understand why someone would pass a parameter which isn't used; that's all.17:18
denis_makogoni would remove context since it's not used, but i woudn't change configuration into it's id17:18
amrithok, as you wish17:19
denis_makogonamrith, because i really don't like the idea of double asking backend for the same configuration object17:19
amrithdon't understand why but that's fine.17:19
*** saurabhs has joined #openstack-trove17:19
amrithas you wish17:20
denis_makogonamrith, same thing for services.py#5017:20
denis_makogondougshelley66, ping17:21
amrithlike I said, if you really didn't want to ask for config twice, there's a way to simplify the code. but that's your call.17:21
amrithare we all set? or should we wait for Vipul?17:22
denis_makogonamrith, i'll ping him later17:22
vipulSo amrith, denis_makogon -- In terms of logic change, it looks like the change is really to fast-fail the poll_until if the nova server status is already in a failed state17:23
dougshelley66denis_makogon: hi17:23
denis_makogonamrith, code simplification in this case will hurt perfomance  API service17:23
amrithvipul, I believe that is an unintended consequence17:23
vipulwhich i think is fine .. right now we'll just be waiting longer for the timeout17:23
vipuli think that part is a good change17:23
denis_makogondougshelley66, can we talk about https://review.openstack.org/#/c/98985/6/trove/configuration/views.py,cm ?17:24
vipuli do agree with amrith on the verbiage aorund the method name17:24
vipulit shouldn't be 'valid'17:24
dougshelley66denis_makogon: sure17:24
dougshelley66the bug report was pretty cryptic17:24
denis_makogonamrith, vipul, ok17:24
dougshelley66maybe we should fix that first17:24
dougshelley66it read it to mean we wanted those 2 pieces of data on list and show17:24
denis_makogondougshelley66, bug report tells to change only two types of views17:25
vipulthe goal is to find an active instance, so we should rename that to trove_instance_is_active or something17:25
dougshelley66denis_makogon, ok so i misunderstood what the second view was for17:25
denis_makogonvipul, i get that, i'll update it17:25
amrithvipul, do you believe the change in order or testing nova and trove is material? I wasn't sure so I just highlighted it. it could change the user visible experience.17:25
dougshelley66denis_makogon, however, the column headers did get chnaged on list17:25
dougshelley66denis_makogon, but contain no data17:25
SlickNikvipul / amrith / denis_makogon: Isn't the fast poll something that we could use for the non-heat case as well? I'm still looking at the code but is there a need to fork the logic for the two cases? (heat vs. non-heat)17:26
*** yidclare has joined #openstack-trove17:26
vipulamrith: that's a good point.. the main thing is the user will see thier instance go to Error with reason being BUILD_TIMEOUT17:26
amrithSlickNik, I don't know. One of my comments on the review is that the comments on the change imply that we'll be doing something more for heat but I don't see code for that.17:27
vipuleven when the reason really isn't because we timed out waiting for it go out of BUILDING.. it may have failed due to Nova failure first17:27
denis_makogonSlickNik, yes, before that poll_until method stack already completely ACTIVE17:27
amrithvipul, yes. but in the past the error they'd have got was that the SERVICE was in a bad state. Now it will say the SERVER is in a bad state17:28
amrithor some such17:28
denis_makogonSlickNik, no need to ask nova to verify server status17:28
amrithnotice that the two throw's have subtly different words.17:28
vipulSlickNik: yea according to the commit msg denis_makogon says it will already wait for stack to be completed, still need to verify that17:28
amrithI don't know if that could change for someone who was expecting a particular failure17:28
amrithlike, maybe a test?17:28
vipulamrith: this code is async, so that message wont' actually be sent back17:28
denis_makogonSlickNik, https://github.com/openstack/trove/blob/master/trove/taskmanager/models.py#L418-L42317:28
vipulit's really a loop the taskmanager does after it's already responded to the request17:28
vipulthe change would impact how quickly a instance changes to ERROR based on guest timeout17:29
amrithvipul, OK, I'll go read the code again and try to understand that.17:29
denis_makogondougshelley66, https://review.openstack.org/#/c/98988/2/troveclient/v1/shell.py,cm i guess you tested it with this patch ?17:31
denis_makogondougshelley66, if that's so, it means that i should revert change for list command17:31
dougshelley66denis_makogon, yes i believe i applied both patches and tried it17:31
denis_makogondougshelley66, [id, name, description] is enough for list17:31
dougshelley66denis_makogon, if you do configuration list what do you see for the result set?17:31
dougshelley66denis_makogon, before the patch is has those items + ds version id17:32
denis_makogondougshelley66, i should see the same attributes17:32
dougshelley66denis_makogon, after the patch is has name and version17:32
denis_makogondougshelley66, yes, i forgot about that17:32
dougshelley66denis_makogon, but they aren't populated17:32
*** michael-yu has joined #openstack-trove17:32
denis_makogondougshelley66, version id would be there17:33
denis_makogondougshelley66, i get that, just need to fix client code17:33
denis_makogondougshelley66, thanks for pointing out17:33
dougshelley66denis_makogon, ok - maybe the bug report should include a description of exactly what is expected for the result sets from the 2 commands?17:33
denis_makogondougshelley66, i guess no, since each view is pinned to exact shell command17:34
vipuldenis_makogon, amrith: i think a better thing to do here is throw a different exception when it's a Nova fast-fail Error, vs when it's a Trove Service status Error.  When you update the instance status, if it's a nova error, then we need to set it to     FAILED = ServiceStatus(0x08, 'failed to spawn', 'FAILED')17:34
denis_makogonvipul, it's already done17:35
vipulwhere17:35
denis_makogonvipul, https://github.com/openstack/trove/blob/master/trove/taskmanager/models.py#L26417:35
dougshelley66denis_makogon, so I assume that bug report and the resulting patch would confuse others. All I'm saying is we should clarify the intent/requirements in the bug report17:35
denis_makogonpoll_until will fail, and instance models will be updated17:35
vipuldenis_makogon: right -- that method assumes that we waiting 10 minutes and the error is because of a GUEST Timeout17:36
denis_makogondougshelley66, aren't they already there ?17:36
vipuls/waiting/waited17:36
denis_makogonvipul, poll_untill will fail faster if nova instance is in ERROR/FAILED state17:36
dougshelley66denis_makogon, so a bug report that starts out by dumping code fragments is misguided from the beginning17:37
denis_makogonit would not wait for guest17:37
vipuldenis_makogon: but the error message won't indicate that it failed because of Nova error -- it will only say that it failed because guest timed out17:37
dougshelley66denis_makogon, i would assume it should says something like "when we run trove configuration show" the output should look like this: etc17:37
denis_makogondougshelley66, let's ask SlickNik to update bug report17:37
dougshelley66denis_makogon, fine17:37
denis_makogondougshelley66, ok, i'm fine with that too17:37
denis_makogonvipul, it would be enough to add logging message here https://github.com/openstack/trove/blob/master/trove/taskmanager/models.py#L31717:38
denis_makogonvipul, and if it's not a guest timeout exception would be handled by https://github.com/openstack/trove/blob/master/trove/taskmanager/models.py#L266-L26717:39
denis_makogonvipul, so, i guess it'll be enough to add certain log message inside _service_is_active for that case if nova instance was corrupted17:40
denis_makogonvipul, and of course, self.update_statuses_on_time_out() would be added there too17:41
*** jmontemayor has joined #openstack-trove17:41
vipuldenis_makogon: it's not enough just to add a logging message.. you need to catch a diffrent exception, and based on the exception caught set the appropriate reason for failure17:41
*** ramashri has quit IRC17:42
*** cweidenkeller has joined #openstack-trove17:42
*** jmontemayor_ has joined #openstack-trove17:42
denis_makogonvipul, exception that is being threw is a TroveError, that is a parent of PollTimeout exception17:43
denis_makogonvipul, i agree that it should be fixed, but it's fit into given bug-report17:43
denis_makogonvipul, amrith or someone else can file a bug-report and fix described issue17:43
*** jcru has quit IRC17:44
denis_makogonvipul, do you agree?17:44
amrithhuh?17:44
*** haomaiwa_ has joined #openstack-trove17:44
denis_makogonamrith, if you don't want to do that, then ok17:44
SlickNikdenis_makogon: So, what's wrong with querying nova to ensure that the server status is ACTIVE even in the heat case (i.e. leaving the code as is)? I get that it will pass since the stack is ACTIVE.17:45
SlickNikBut there's an argument to be made for simplicity of code, and maintainability by not introducing a fork to do something different for the heat case when it's not absolutely necessary.17:45
amrithdenis_makogon, I believe that what vipul gave you was feedback on your fix and potentially the reason he would -1 the fix17:45
amriththis isn't a different bug that he's suggesting.17:45
*** jcru has joined #openstack-trove17:45
denis_makogonamrith, given issue doesn't fit into proposed bug report17:45
amrithI believe SlickNik's comments are in the same vein17:45
*** jcru has quit IRC17:46
*** jmontemayor has quit IRC17:46
amriththat he is proposing that your fix is not the right way (maybe). he's right here so I'll let him put words in his own mouth ;)17:46
*** jcru has joined #openstack-trove17:46
denis_makogonSlickNik, those nova calls for heat-provisioned resources are redurand17:46
denis_makogonSlickNik, because they are not needed since stack is already in ACTIVE state17:47
*** jmontemayor has joined #openstack-trove17:47
denis_makogonSlickNik, and it means that all resources are in ACTIVE/In-use state17:47
*** michael-yu has quit IRC17:47
denis_makogonSlickNik, taskmanager will do {usage_timeout}/sleet_timeout redurand calls to nova17:48
denis_makogonSlickNik, we should avoid such things17:49
*** haomaiwa_ has quit IRC17:49
*** jmontemayor_ has quit IRC17:49
*** grapex has quit IRC17:50
*** grapex has joined #openstack-trove17:51
amrithdenis_makogon, by that logic, we should not check nova state each time for non-heat instances either. just on the first time, maybe?17:51
amrithin other words, the calls to check nova status are redumbnant.17:52
vipullol @ amrith17:52
denis_makogonamrith, in this case we need to separate pollers17:52
denis_makogon*two17:52
denis_makogonone will finish as soon as nova instance is ACTIVE17:53
SlickNikdenis_makogon: Yes, I get that we're now making an extra call to nova to verify that the server is in an ACTIVE state (for the heat case).17:53
SlickNikFrom a code maintainability perspective though, the more branches your code has, the harder it gets to reason about it / maintain.17:53
SlickNikSo that tradeoff makes this a bit of a six-of-one, half-a-dozen of another issue to me. Therefore, my inclination is to preserve the status quo.17:53
*** kevinconway_ has joined #openstack-trove17:53
denis_makogonand other one will wait until guest is ACTIVE17:53
*** kevinconway has quit IRC17:53
*** kevinconway_ is now known as kevinconway17:53
SlickNikHowever, vipul does bring up a good point about the fail fast option, which I think would be beneficial  to both the heat / non-heat case.17:53
SlickNikSo I would not be opposed to fixing that (albeit as a separate bugfix).17:54
denis_makogonSlickNik, i get that, and will fix that too17:54
*** ramashri has joined #openstack-trove17:55
denis_makogonSlickNik, but i really think that we need to split out guest poller and use it while heat support is enabled17:55
vipulSlickNik: +1 if we can combine the two code paths into one i'm all for that17:55
*** ramishra has joined #openstack-trove17:56
amrithvipul, that's how the code was earlier17:56
amrithI'm getting the feeling that there is a question of whether we should split this code out in the heat case17:56
amriththat's above my level of understanding of the code/product so I'll put my nose elsewhere.17:56
vipulthen let's not do this change if we are unnecessarily optimizing17:57
vipuli think this change can morph into fast-fail if nova is already failed17:57
denis_makogonvipul, stack poller will fail even before guest poller17:58
vipulyea I get that. that's already in place17:58
denis_makogonvipul, so, i still thinks that we should not talk to nova inside poll method is heat is enabled17:58
*** NehaV3 has joined #openstack-trove17:59
*** ramishra has quit IRC18:00
vipuldenis_makogon: Yea i kinda of agree with that it's unnecessary to query Nova if the instance was provisioned by Heat -- but i think as SlickNik mentioned it introduces yet another code path.. so if it's not a big deal to query nova even in Heat case then we should keep doing it for simplicity18:04
denis_makogonvipul, what if we would split-up poller into to pollers, and move nova instance poller under _create_serve_... methods?18:10
denis_makogonvipul, than we would have an ability to fast-fail even before guest poller ?18:11
*** shayneburgess has joined #openstack-trove18:11
*** michael-yu has joined #openstack-trove18:11
denis_makogonvipul, lets say, poller will be placed at _create_server(...) method18:12
*** jasonb365 has joined #openstack-trove18:12
SlickNikglucas / vgnbkr / peterstac: Thanks for pushing up replication patches.18:12
denis_makogonvipul, and in this case we would not need to maintain two code paths18:13
SlickNikFYI I'm going to be taking it for a spin this week, and stepping up reviews on it.18:13
SlickNikWill keep you in the loop.18:13
SlickNik^ also dougshelley6618:13
dougshelley66Thanks SlickNik - that will be helpful18:14
glucasSlickNik: great, thanks!18:14
*** sgotliv has joined #openstack-trove18:21
peterstacSlickNik: Thx! (although I was just pushing the review, not actual patches ;) )18:23
*** ramishra has joined #openstack-trove18:23
*** jmontemayor has quit IRC18:24
SlickNikpeterstac: I know. Included you on the list, since you mentioned it earlier this morning ;)18:24
peterstacSlickNik: np18:25
*** ramishra has quit IRC18:27
vipuldenis_makogon: It's still two code paths, but push up an update to that review.. we can see how that looks18:28
denis_makogonvipul, of course18:31
*** yidclare has quit IRC18:31
*** NehaV3 has quit IRC18:31
amrithdenis_makogon, are we all set? you are going to push new code, yes?18:40
denis_makogonamrith, yes18:43
amriththx18:43
denis_makogonamrith, thanks to you too18:50
amrithdenis_makogon, np.18:51
*** ViswaV has quit IRC19:05
openstackgerritMichael Yu proposed a change to openstack/trove: Adds backup/restore support for couchbase  https://review.openstack.org/8673119:06
*** openstackgerrit has quit IRC19:06
*** openstackgerrit_ has joined #openstack-trove19:07
*** openstackgerrit_ is now known as openstackgerrit19:08
openstackgerritMichael Yu proposed a change to openstack/trove: Adds backup/restore support for Couchbase  https://review.openstack.org/8673119:10
*** yidclare has joined #openstack-trove19:11
*** ViswaV has joined #openstack-trove19:22
*** yidclare has quit IRC19:23
*** ramishra has joined #openstack-trove19:24
*** ramishra has quit IRC19:29
*** yidclare has joined #openstack-trove19:36
*** ramashri has quit IRC19:45
*** yidclare has quit IRC19:47
*** pdmars has quit IRC19:50
*** pdmars has joined #openstack-trove19:55
*** shayneburgess has quit IRC19:57
*** julienvey has quit IRC19:57
*** shayneburgess has joined #openstack-trove19:59
*** michael-yu has quit IRC20:05
*** boden has quit IRC20:15
*** michael-yu has joined #openstack-trove20:19
*** ramishra has joined #openstack-trove20:25
*** michael-yu has quit IRC20:27
*** julienvey has joined #openstack-trove20:28
*** ramishra has quit IRC20:29
*** michael-yu has joined #openstack-trove20:29
*** julienve_ has joined #openstack-trove20:29
*** ramashri has joined #openstack-trove20:30
*** jasonb365 has quit IRC20:30
*** julienvey has quit IRC20:32
*** julienve_ has quit IRC20:35
*** ramashri has quit IRC20:35
*** shayneburgess_ has joined #openstack-trove20:36
*** ramashri has joined #openstack-trove20:36
*** shayneburgess has quit IRC20:36
*** shayneburgess_ is now known as shayneburgess20:36
*** ramashri has quit IRC20:37
*** ramashri has joined #openstack-trove20:37
*** edmund has quit IRC20:38
*** edmund has joined #openstack-trove20:40
*** yidclare has joined #openstack-trove20:41
*** yidclare has quit IRC20:42
*** juantwo has quit IRC20:46
*** juantwo has joined #openstack-trove20:52
*** miqui has quit IRC20:54
*** juantwo has quit IRC20:56
*** juantwo has joined #openstack-trove20:57
*** juantwo has quit IRC21:01
*** jmontemayor has joined #openstack-trove21:04
*** jmontemayor has quit IRC21:06
*** denis_makogon has quit IRC21:06
*** jmontemayor has joined #openstack-trove21:07
*** ramishra has joined #openstack-trove21:25
*** fifieldt has quit IRC21:27
*** mattgriffin has quit IRC21:28
*** ramishra has quit IRC21:30
*** julienvey has joined #openstack-trove21:30
*** julienvey has quit IRC21:31
*** julienvey has joined #openstack-trove21:31
*** dkehn_ has joined #openstack-trove21:33
*** dkehnx has quit IRC21:34
*** jmontemayor has quit IRC21:34
*** julienvey has quit IRC21:36
*** fifieldt has joined #openstack-trove21:40
*** yidclare has joined #openstack-trove21:51
*** ranjitha has joined #openstack-trove21:59
*** shayneburgess has quit IRC22:10
*** robertmy_ has joined #openstack-trove22:12
*** robertmy_ has quit IRC22:12
*** michael-yu has quit IRC22:12
*** robertmyers has quit IRC22:13
*** robertmy_ has joined #openstack-trove22:13
*** ramashri has quit IRC22:15
*** ramashri has joined #openstack-trove22:15
*** dkehn_ is now known as dkehnx22:17
*** robertmy_ has quit IRC22:18
*** mattgriffin has joined #openstack-trove22:20
*** juantwo has joined #openstack-trove22:23
*** juantwo has quit IRC22:23
*** juantwo has joined #openstack-trove22:24
*** ramishra has joined #openstack-trove22:26
*** kevinconway has quit IRC22:27
*** michael-yu has joined #openstack-trove22:28
*** thedodd has quit IRC22:29
*** ramishra has quit IRC22:31
*** shayneburgess has joined #openstack-trove22:33
*** edmund has quit IRC22:35
*** flaper87 is now known as flaper87|afk22:41
*** cweidenkeller has quit IRC22:44
*** sgotliv has quit IRC22:44
*** rhodgin has quit IRC22:54
*** mattgriffin has quit IRC23:02
*** jmontemayor has joined #openstack-trove23:05
*** juantwo_ has joined #openstack-trove23:06
*** juantwo_ has joined #openstack-trove23:06
*** saurabhs has left #openstack-trove23:08
*** juantwo has quit IRC23:09
*** pdmars has quit IRC23:12
*** jcru has quit IRC23:14
*** Isotopp has quit IRC23:17
*** tvoran has quit IRC23:18
*** mordred has quit IRC23:24
*** Isotopp has joined #openstack-trove23:25
*** Isotopp has joined #openstack-trove23:25
*** ramishra has joined #openstack-trove23:27
*** mordred has joined #openstack-trove23:27
*** mordred has quit IRC23:30
*** mordred has joined #openstack-trove23:30
*** ramishra has quit IRC23:31
*** michael-yu has quit IRC23:33
*** shayneburgess has quit IRC23:44
*** ranjitha has quit IRC23:48
*** ranjitha has joined #openstack-trove23:51
*** amytron has quit IRC23:52
*** jmontemayor has quit IRC23:54
*** achampion has joined #openstack-trove23:55

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