Thursday, 2017-04-06

*** wanghao has joined #openstack-meeting-300:03
*** pvaneck has quit IRC00:03
*** gouthamr has joined #openstack-meeting-300:03
*** neiljerram has quit IRC00:03
*** donghao has joined #openstack-meeting-300:08
*** dimtruck is now known as zz_dimtruck00:09
*** mtanino has quit IRC00:14
*** shuyingya has joined #openstack-meeting-300:14
*** donghao has quit IRC00:14
*** zz_dimtruck is now known as dimtruck00:16
*** shuyingya has quit IRC00:19
*** shuyingya has joined #openstack-meeting-300:20
*** diablo_rojo has quit IRC00:21
*** stanzgy has joined #openstack-meeting-300:21
*** huanxuan has joined #openstack-meeting-300:23
*** lucasxu has joined #openstack-meeting-300:24
*** shuyingya has quit IRC00:24
*** armstrong has quit IRC00:28
*** zhurong has joined #openstack-meeting-300:32
*** tellesnobrega has quit IRC00:32
*** tellesnobrega has joined #openstack-meeting-300:34
*** harlowja has quit IRC00:39
*** mariannelm has joined #openstack-meeting-300:40
*** baoli has quit IRC00:49
*** ediardo has quit IRC00:52
*** eliqiao has quit IRC00:58
*** lucasxu has quit IRC01:01
*** Sukhdev has quit IRC01:10
*** shuyingya has joined #openstack-meeting-301:21
*** shuyingya has quit IRC01:21
*** shuyingya has joined #openstack-meeting-301:21
*** mrunge_ has quit IRC01:23
*** harlowja has joined #openstack-meeting-301:26
*** harlowja has quit IRC01:26
*** mariannelm has quit IRC01:27
*** mrunge_ has joined #openstack-meeting-301:33
*** ediardo has joined #openstack-meeting-301:34
*** Sukhdev has joined #openstack-meeting-301:47
*** shuyingya has quit IRC01:59
*** s3wong has quit IRC01:59
*** shuyingya has joined #openstack-meeting-301:59
*** gcb has joined #openstack-meeting-302:04
*** seanatcisco has quit IRC02:08
*** Sukhdev has quit IRC02:21
*** reedip has quit IRC02:25
*** wanghao has quit IRC02:31
*** wanghao has joined #openstack-meeting-302:32
*** reedip has joined #openstack-meeting-303:09
*** VW has joined #openstack-meeting-303:11
*** lamt has joined #openstack-meeting-303:13
*** baoli has joined #openstack-meeting-303:16
*** VW has quit IRC03:23
*** scottda has quit IRC03:35
*** sdague has quit IRC03:37
*** baoli has quit IRC03:37
*** reedip has quit IRC03:44
*** reedip has joined #openstack-meeting-303:45
*** Sukhdev has joined #openstack-meeting-303:53
*** seanatcisco has joined #openstack-meeting-303:54
*** qwebirc5777 has joined #openstack-meeting-303:59
*** qwebirc5777 has quit IRC04:00
*** bobmel_ has joined #openstack-meeting-304:02
*** anilvenkata has joined #openstack-meeting-304:10
*** xavierr has quit IRC04:10
*** bobmel has quit IRC04:10
*** markmcclain has quit IRC04:12
*** markmcclain has joined #openstack-meeting-304:13
*** xavierr has joined #openstack-meeting-304:17
*** amotoki has joined #openstack-meeting-304:37
*** wanghao has quit IRC04:38
*** seanatcisco has quit IRC04:38
*** wanghao has joined #openstack-meeting-304:39
*** ediardo has quit IRC04:42
*** lamt has quit IRC04:45
*** psachin_ has joined #openstack-meeting-304:47
*** gouthamr has quit IRC04:47
*** stanzgy has quit IRC04:59
*** stanzgy has joined #openstack-meeting-305:05
*** iyamahat has joined #openstack-meeting-305:17
*** pgadiya has joined #openstack-meeting-305:41
*** pgadiya has quit IRC05:41
*** cshastri has joined #openstack-meeting-306:08
*** makowals has joined #openstack-meeting-306:08
*** andreas_s has joined #openstack-meeting-306:11
*** reedip has quit IRC06:17
*** jtomasek has joined #openstack-meeting-306:19
*** dimtruck is now known as zz_dimtruck06:56
*** matrohon has joined #openstack-meeting-306:56
*** pcaruana has joined #openstack-meeting-307:02
*** reedip has joined #openstack-meeting-307:10
*** belmoreira has joined #openstack-meeting-307:13
*** aarefiev_afk is now known as aarefiev07:17
*** egallen has joined #openstack-meeting-307:25
*** egallen has quit IRC07:26
*** Sukhdev has quit IRC07:26
*** yamamoto has quit IRC07:33
*** yamamoto has joined #openstack-meeting-307:34
*** rmart04 has joined #openstack-meeting-307:43
*** shuyingya has quit IRC07:50
*** shuyingya has joined #openstack-meeting-307:51
*** ccamacho has joined #openstack-meeting-307:55
*** shuyingya has quit IRC07:56
*** shuyingya has joined #openstack-meeting-307:57
*** shuyingy_ has joined #openstack-meeting-308:05
*** ralonsoh has joined #openstack-meeting-308:06
*** ralonsoh_ has joined #openstack-meeting-308:06
*** shuyingya has quit IRC08:08
*** ralonsoh has quit IRC08:10
*** ralonsoh_ is now known as ralonsoh08:12
*** iyamahat has quit IRC08:20
*** ltomasbo|away is now known as ltomasbo08:22
*** stevejims has joined #openstack-meeting-308:23
*** rossella_s has joined #openstack-meeting-308:24
*** markvoelker has quit IRC08:26
*** jjung has joined #openstack-meeting-308:31
*** stendulker has joined #openstack-meeting-308:32
*** abalutoiu has quit IRC08:39
*** alex_xu has quit IRC08:41
*** gcb has quit IRC08:42
*** alex_xu has joined #openstack-meeting-308:42
*** gcb has joined #openstack-meeting-308:43
*** yamamoto has quit IRC08:48
*** yamamoto has joined #openstack-meeting-308:48
*** jtomasek has quit IRC08:49
*** jtomasek_ has joined #openstack-meeting-308:49
*** abalutoiu has joined #openstack-meeting-309:11
*** mickeys has quit IRC09:12
*** apetrich has quit IRC09:14
*** abalutoiu_ has joined #openstack-meeting-309:15
*** abalutoiu has quit IRC09:17
*** reedip has quit IRC09:22
*** abalutoiu_ has quit IRC09:34
*** jtomasek_ has quit IRC09:46
*** neiljerram has joined #openstack-meeting-309:49
*** yamamoto has quit IRC09:57
*** apetrich has joined #openstack-meeting-309:57
*** yamamoto has joined #openstack-meeting-309:57
*** mickeys has joined #openstack-meeting-310:12
*** yamamoto has quit IRC10:13
*** yamamoto has joined #openstack-meeting-310:15
*** lpetrut1 has joined #openstack-meeting-310:16
*** mickeys has quit IRC10:18
*** lpetrut1 has quit IRC10:21
*** markvoelker has joined #openstack-meeting-310:27
*** markvoelker has quit IRC10:31
*** abalutoiu has joined #openstack-meeting-310:44
*** stevejims has quit IRC10:48
*** shuyingy_ has quit IRC10:57
*** shuyingya has joined #openstack-meeting-310:57
*** reedip has joined #openstack-meeting-310:58
*** lpetrut has joined #openstack-meeting-311:00
*** apetrich has quit IRC11:04
*** jtomasek has joined #openstack-meeting-311:04
*** sdague has joined #openstack-meeting-311:05
*** xavierr has quit IRC11:08
*** qwebirc80427 has joined #openstack-meeting-311:10
*** mariannelm has joined #openstack-meeting-311:13
*** mickeys has joined #openstack-meeting-311:15
*** ccamacho is now known as ccamacho|lunch11:15
*** stendulker has quit IRC11:16
*** mickeys has quit IRC11:19
*** alexismonville has joined #openstack-meeting-311:30
*** rossella_s has quit IRC11:36
*** alexismonville has quit IRC11:44
*** elmiko has joined #openstack-meeting-311:54
*** scottda has joined #openstack-meeting-312:00
*** ccamacho|lunch is now known as ccamacho12:05
*** alexismonville has joined #openstack-meeting-312:07
*** mickeys has joined #openstack-meeting-312:16
*** mickeys has quit IRC12:20
*** shuyingya has quit IRC12:21
*** shuyingya has joined #openstack-meeting-312:21
*** shuyingy_ has joined #openstack-meeting-312:25
*** qwebirc80427 has quit IRC12:25
*** markvoelker has joined #openstack-meeting-312:28
*** shuyingya has quit IRC12:28
*** shuyingy_ has quit IRC12:32
*** markvoelker has quit IRC12:32
*** apetrich has joined #openstack-meeting-312:37
*** gouthamr has joined #openstack-meeting-312:37
*** rossella_s has joined #openstack-meeting-312:39
*** lamt has joined #openstack-meeting-312:41
*** zhurong has quit IRC12:42
*** alexismonville has quit IRC12:52
*** wanghao has quit IRC12:56
*** lyan has joined #openstack-meeting-312:56
*** psachin_ has quit IRC12:56
*** rabel_webclient has joined #openstack-meeting-312:57
*** rabel_webclient has quit IRC12:57
*** tobberydberg has joined #openstack-meeting-312:59
*** rabel has joined #openstack-meeting-313:00
*** stanzgy has quit IRC13:02
*** sdague has quit IRC13:03
*** tobberydberg has quit IRC13:03
*** ralonsoh has quit IRC13:05
*** haleyb has joined #openstack-meeting-313:05
*** ralonsoh has joined #openstack-meeting-313:06
*** cleong has joined #openstack-meeting-313:06
*** srobert has joined #openstack-meeting-313:09
*** srobert has quit IRC13:09
*** srobert has joined #openstack-meeting-313:09
*** mickeys has joined #openstack-meeting-313:16
*** shuyingya has joined #openstack-meeting-313:21
*** mickeys has quit IRC13:21
*** shuyingya has quit IRC13:25
*** baoli has joined #openstack-meeting-313:28
*** shu-mutou-AWAY has quit IRC13:29
*** spzala has joined #openstack-meeting-313:30
*** shuyingya has joined #openstack-meeting-313:32
*** huanxuan has quit IRC13:34
*** lpetrut has quit IRC13:36
*** lpetrut has joined #openstack-meeting-313:36
*** baoli has quit IRC13:39
*** lhx_ has joined #openstack-meeting-313:41
*** jeremyfreudberg has joined #openstack-meeting-313:41
*** VW has joined #openstack-meeting-313:41
*** lpetrut has quit IRC13:42
*** VW has quit IRC13:42
*** lpetrut has joined #openstack-meeting-313:42
*** VW has joined #openstack-meeting-313:42
*** armstrong has joined #openstack-meeting-313:44
*** tosky has joined #openstack-meeting-313:44
*** zz_dimtruck is now known as dimtruck13:45
*** sdague has joined #openstack-meeting-313:47
*** baoli has joined #openstack-meeting-313:47
*** rbak has joined #openstack-meeting-313:50
*** esikachev has joined #openstack-meeting-313:51
*** vgridnev has joined #openstack-meeting-313:51
*** tenobreg has joined #openstack-meeting-313:58
*** tellesnobrega has quit IRC13:59
*** armstrong has quit IRC13:59
*** tenobreg is now known as tellesnobrega13:59
tellesnobrega#startmeeting sahara14:00
openstackMeeting started Thu Apr  6 14:00:03 2017 UTC and is due to finish in 60 minutes.  The chair is tellesnobrega. Information about MeetBot at
openstackUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.14:00
*** openstack changes topic to " (Meeting topic: sahara)"14:00
openstackThe meeting name has been set to 'sahara'14:00
elmikoᶘ ᵒᴥᵒᶅ14:00
tellesnobregaelmiko, that's new14:00
*** lamt has quit IRC14:00
elmikogotta keep it fresh14:01
*** rabel has left #openstack-meeting-314:01
tellesnobrega#topic News/Updates14:01
*** openstack changes topic to "News/Updates (Meeting topic: sahara)"14:01
vgridnevnothing new from me, worked a little bit on job updates in project config14:02
tellesnobregaI've been working on image gen for CDH, I'm almost done with that image, testing it but it fails with Cant Format NameNode, I'm working on it, hopefully we will have something by next week14:02
esikachevi am working on sahara ci patches14:02
tellesnobregaother than that I updated versions of spark and storm, so if you guys have time, please review those14:03
*** armstron_ has joined #openstack-meeting-314:03
vgridnevyou are welcome to review those14:03
shuyingyaI am thinking about how to refactor CDH plugin and going to update CDH 5.1014:03
tellesnobregavgridnev, will do14:03
tellesnobregashuyingya, nice14:04
vgridnevthe general idea is to replace common methods into base classes14:04
vgridnevnothing new, I guess14:05
tellesnobregayes, should be "simple"14:05
tellesnobreganot fast or easy, but simple to do it14:05
*** melwitt has quit IRC14:05
*** melwitt has joined #openstack-meeting-314:05
shuyingyaOK, let's keep it simple.  will post few patch latter14:06
*** melwitt is now known as Guest4891014:06
shuyingyathanks vgridnev tellesnobrega14:06
*** jraim has quit IRC14:06
*** tonytan4ever has joined #openstack-meeting-314:06
*** jraim_ has joined #openstack-meeting-314:06
jeremyfreudbergshuyingya, also good to take a look how this was done for vanilla (common utils in "hadoop2" folder, only the version-specific stuff separate)14:06
tellesnobregaabout sahara CI, the problem was resources, but I'm seeing a lot of failure on the patches, is that expected and is being worked on? or the errors are "random" failures14:07
*** spzala has quit IRC14:07
shuyingyayes,  I have thought about mkdir "cdh5" like "hadoop2". but looks like unecessary14:08
toskyI'm just reviewing stuff for now, but preparing for future changes14:08
*** armstrong has joined #openstack-meeting-314:09
tellesnobregamariannelm, what have you been working on lately?14:09
*** bobh has joined #openstack-meeting-314:10
tellesnobregaabout outreachy, there were two official applications to the Sahara project, our goal is to allow import and export of templates into sahara, it is something simple, but I would love to see that in, specially for moving around clouds, that makes things easier14:10
tellesnobregathere will be more mention of it in the future if we do get someone14:11
*** armstron_ has quit IRC14:11
*** rickflare2 has joined #openstack-meeting-314:11
*** rickflare has quit IRC14:11
*** rpodolyaka is now known as rpodolyaka_afk14:11
*** toan has quit IRC14:11
*** sindhu has quit IRC14:11
tellesnobregado we have any more news or updates?14:12
*** rickflare2 is now known as rickflare14:12
*** sindhu has joined #openstack-meeting-314:13
tellesnobregalets move on them14:13
mariannelmI didn't have much time this week, but I'm currently working on JSON changes of APIv2, and planning to fix some bugs this week14:13
*** armstrong has quit IRC14:13
*** toan has joined #openstack-meeting-314:13
tellesnobregamariannelm, nice, if you have some questions about api v2, stay tunned14:13
tellesnobrega#topic APIv214:13
*** openstack changes topic to "APIv2 (Meeting topic: sahara)"14:13
tellesnobregaI wanted to add this topic so we can bring our expert on the matter elmiko14:14
*** MarkAtwood has quit IRC14:14
*** kiseok7 has quit IRC14:14
*** johndperkins has joined #openstack-meeting-314:14
shuyingyawelcome back ~  elmiko14:14
*** davidlenwell has quit IRC14:14
*** baoli has quit IRC14:14
elmikoso happy to see some movement on this!14:14
tellesnobregaas you've probably seen, I sent the email about microversions on sahara and I was talking with michael yesteday and asked him to come over and gives a little overview on that, and why this should be done and when so we are all on the same page on it14:15
*** rpodolyaka_afk has quit IRC14:15
tellesnobregaelmiko, shuyingya and mariannelm are the ones doing the hard work on api v214:15
elmikotellesnobrega: i did not see the email on the list, do you have a link to the archive?14:15
tellesnobregaI'm kinda of organizing the effort, but they are the brains and hands on it14:15
tellesnobregayes, just a second14:16
*** MarkAtwood has joined #openstack-meeting-314:16
*** baoli has joined #openstack-meeting-314:16
*** davidlenwell has joined #openstack-meeting-314:16
*** alex_didenko has quit IRC14:17
elmikoshuyingya mariannelm i am usually in the sahara channel but mostly quiet, if you run into issues or have questions feel free to ping me14:17
toskyand we have a slot reserved into the API WG meeting in less than two hours14:17
tosky(about microversioning)14:17
*** rpodolyaka_afk has joined #openstack-meeting-314:17
*** armstrong has joined #openstack-meeting-314:17
shuyingyagreat elmiko. thanks14:17
mariannelmelmiko, cool, thanks!14:18
tellesnobregaelmiko, this is the title [Sahara][APIv2] Microversions looking ofr a link14:18
elmikotellesnobrega: ok, i'll search again. i have it on digest now14:18
tellesnobregathanks jeremyfreudberg14:18
*** alex_didenko has joined #openstack-meeting-314:19
elmikofor future reference, if you want to attract the api-wg, put [api] in the subject14:19
*** mickeys has joined #openstack-meeting-314:19
tellesnobregaelmiko, will do14:19
*** esikache1 has joined #openstack-meeting-314:19
* elmiko reads14:19
*** raissa has joined #openstack-meeting-314:19
elmikoso, in answer to the question from the email14:20
elmikoi would say prioritize the other work before microversions14:20
elmikoyou will need to have a solid foundation to build the microversion stuff on top of (ie making sure you can support older code paths), but so long as you are aware of that i think you can wait until the new api looks stable to add them14:20
tellesnobregathat makes total sense14:21
elmikoin the case of sahara it probably won't be that big a jump14:21
elmikoespecially since the v1 v1.1 and v2 apis all have different top level URIs14:21
elmikoonce v2 is rolling, you will need to be aware that any changes which bump the microversion _may_ have impact on the code paths14:22
elmikofor example14:22
*** esikachev has quit IRC14:22
elmikoreturning a resource payload on a request for a version X.Y should ensure that it doesn't include payload information from X.Y+114:22
elmikobut that will largely depend on how the api grows after release14:23
*** mickeys has quit IRC14:23
elmikoand of course, try to follow
elmikoer oops, wrong link14:24
elmikothat's the correct one14:24
*** abalutoiu has quit IRC14:24
tellesnobregathanks elmiko14:24
tellesnobregadoes anyone have any concerns on API microversions?14:24
* tosky has14:24
tellesnobregaI know of your concern tosky :)14:25
toskyok, apart from me then :)14:25
tellesnobregaI was going to ask you to bring it up, just wanted to know if anyone else had anything to discuss on the topic14:25
toskyjust to summarize: I'd like to have a "sane" use of microversioning, avoiding incompatible changes in the API methods, even if microversioning allows for it14:25
elmikoimo, tosky's concern can definitely be handled appropriately by the team14:26
elmikoeven if the governance doesn't strictly support that stance14:26
elmikoi would document your usage heavily though if you intend to go down that road14:26
toskythat's for sure, let's see14:26
elmikoand honestly, i'm guessing that most folks would say that if you are going to introduce backward in-compat changes, you should really consider bumping major rev.14:27
elmikoso DONT DO THAT!14:27
* elmiko mostly agrees with tosky vision on this14:27
tellesnobregatosky, awesome, whoever will work on this, will have to write a spec on it, and we will keep an eye to have that part documented in the spec and on sahara docs14:27
toskyit seems a bit more of work, but it reduces the test matrix later14:28
tellesnobregathat is the general assumption, if it breaks bump major version14:28
*** dimtruck is now known as zz_dimtruck14:28
*** mtanino has joined #openstack-meeting-314:28
*** markvoelker has joined #openstack-meeting-314:29
tellesnobregado we anything else on APIv2?14:29
tellesnobregamariannelm, shuyingya any concerns or doubts you want to clear out while elmiko is around?14:29
mariannelmsmall thing about json changes, that we were talking about on sahara channel...14:30
tellesnobregago ahead14:30
shuyingyaelmiko  what do you think about this
* elmiko looks14:31
elmikoi think that for the internal job binaries, the big trick about using the common job endpoints will be allowing the client to upload the data14:32
*** galstrom_zzz is now known as galstrom14:32
elmikothat is the only sticking point i see, but in general my preference is to coalesce all jobs to a single group of endpoints14:32
elmikoso, even if you needed to make /jobs/internal that would be preferable to having /jobs and /jobs-internal or similar14:32
*** rossella_s has quit IRC14:33
toskywasn't the idea to kill the internal job binaries?14:33
elmikothat was my preference14:33
mariannelmabout renaming hadoop_version -> plugin_version, I'm creating a new schema for the APIv2 with this change, and dealing with it manually in validationcoode, does it makes sense?14:33
elmikoi think storing th binaries in the database is non-ideal14:33
mariannelmI'm following:
*** rossella_s has joined #openstack-meeting-314:33
mariannelm*validation code14:33
shuyingyaBut about the /jobs/internal, should we implement all of CRUD method14:34
shuyingyaI think it is unecessary14:34
tellesnobregashuyingya, I don't think we need jobs/internal at all14:34
elmikomariannelm: yes, that makes sense. a new schema is required14:34
shuyingyaoops, misunderstanding14:34
elmikoshuyingya: i agree with tellesnobrega, if the team can drop internal jobs then so much the better14:35
*** markvoelker has quit IRC14:35
shuyingyayeah, make sense.14:35
tellesnobregawe shouldn't store binaries on our internal db, we can use swift and bring stuff from there, or manila or whatever14:35
elmikoimo, jobs should leverage cloud storage not sahara's internal database. that is just begging for problems14:35
mariannelmthanks elmiko14:35
*** abalutoiu has joined #openstack-meeting-314:35
shuyingyagot it.14:35
tellesnobregain our db we keep the link to the binary and that is all :)14:35
tellesnobregaas elmiko said, it is trouble, and we don't want any trouble14:36
elmikoremember, if sahara starts managing job storage, then it is now in the storage business as well. there are plenty of cloud services which provide much better storage options than sahara can create.14:36
elmikoit sounds to me like you all have this covered nicely14:37
elmikokudos to the team14:37
tellesnobregaelmiko, as I said to you, we are planning on finishing it in Pike and hopefully have it stable to release in Queen, but lets see how that goes14:37
elmikothat will be awesome, good luck!14:38
tellesnobregaand thanks for dropping by and helping out clarify this14:38
tellesnobregalets move on14:38
elmikoman, this started like back in icehouse...14:38
*** armstrong has quit IRC14:38
elmikono problem, thanks for inviting me =)14:38
mariannelman spec will be needed to deprecate job binary internals? How this will exactly be handled?14:38
tellesnobregamariannelm, yes, I think that is the best way to go14:39
*** diablo_rojo has joined #openstack-meeting-314:40
*** armstrong has joined #openstack-meeting-314:40
tellesnobregajust in case someone haven't seen it, this is the outline of apiv2 work14:40
tellesnobregawe are trying to keep it up to date, and please look at the ones that are started and take some time to review it14:41
*** Guest48910 is now known as melwitt14:42
tellesnobrega#topic OpenDiscussion14:42
*** openstack changes topic to "OpenDiscussion (Meeting topic: sahara)"14:42
* elmiko sinks back into the shadows14:42
tellesnobregaelmiko, thanks again for being here14:42
tellesnobregajeremyfreudberg, did you take a look into vanilla 2.8 and if you will be able to work on that?14:43
tellesnobregaI think you said you would like to14:43
*** anilvenkata has quit IRC14:43
jeremyfreudbergI want to, but haven't had time yet. I can look into it this week, but I probably can't get the work done until after the summit14:43
*** armstrong has quit IRC14:44
toskyvgridnev: just a reminder about - also because I'd like to backport it to Ocata14:44
tellesnobregaok, I don't think that it will be an issue, just take a look and see how complicated it will be and let us know if you want to take that14:44
jeremyfreudbergtellesnobrega, sure14:44
vgridnevwaiting for jenkins to pass14:44
tellesnobregajeremyfreudberg, thanks14:45
*** lyan has quit IRC14:45
toskyvgridnev: right; the jobs mostly passed the build phase, but there was a failure in uploading the images, if I read the logs correctly from sahara-ci14:45
*** zz_dimtruck is now known as dimtruck14:45
toskyvgridnev: but let's see now :)14:45
vgridnevno, I'm saying about just jenkins, which should be able to build all images14:46
tellesnobregajeremyfreudberg, about the Pig job issue, I did take a quick look on it, didn't see any reason why it should be integrated with oozie14:46
*** galstrom is now known as galstrom_zzz14:46
*** lpetrut has quit IRC14:47
tellesnobregamy opinion is that if you want you can move it out, but it is up to you14:47
tellesnobregaI won't have the time to do it now14:47
jeremyfreudbergI definitely want to move it out. I will publish a spec when I figure out how to do it (I have a few questions for the oozie team first)14:48
tellesnobregajeremyfreudberg, great14:48
jeremyfreudbergthat's it for me, gtg14:49
*** jeremyfreudberg has quit IRC14:49
tellesnobregaanyone else has anything to discuss?14:49
tellesnobregaSotK, please keep up with the reviews, we have quite a bit of patches to do so, and lets keep up the work14:51
tellesnobregathanks all14:51
*** spzala has joined #openstack-meeting-314:51
tellesnobregayou all have 9 minutes back14:51
*** openstack changes topic to "OpenStack Meetings ||"14:51
openstackMeeting ended Thu Apr  6 14:51:43 2017 UTC.  Information about MeetBot at . (v 0.1.4)14:51
openstackMinutes (text):
*** tosky has left #openstack-meeting-314:51
*** vgridnev has quit IRC14:52
*** mlavalle has joined #openstack-meeting-314:52
*** john-dav_ has joined #openstack-meeting-314:55
*** cshastri has quit IRC14:56
*** mickeys has joined #openstack-meeting-314:57
*** janzian has joined #openstack-meeting-314:58
*** markvoelker has joined #openstack-meeting-314:59
*** raissa has left #openstack-meeting-314:59
mlavalle#startmeeting neutron_l315:00
openstackMeeting started Thu Apr  6 15:00:27 2017 UTC and is due to finish in 60 minutes.  The chair is mlavalle. Information about MeetBot at
openstackUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.15:00
*** openstack changes topic to " (Meeting topic: neutron_l3)"15:00
openstackThe meeting name has been set to 'neutron_l3'15:00
john-dav_hi o/15:00
*** john-dav_ is now known as john-davidge15:01
mlavallewhy the john-dav_ nick today?15:01
mlavalle#topic Announcements15:01
*** openstack changes topic to "Announcements (Meeting topic: neutron_l3)"15:01
*** marst has joined #openstack-meeting-315:02
mlavallePike-1 is almost here. Next week15:02
mlavalleTime flies as usual!15:02
*** lyan has joined #openstack-meeting-315:02
john-davidgemlavalle: variety is the spice of life15:02
*** galstrom_zzz is now known as galstrom15:03
mlavalleWe are a month away from the Summit in Boston15:03
mlavalleI know I will see haleyb. Hoping to see john-davidge!15:03
*** matrohon has quit IRC15:03
john-davidgemlavalle: Still hopeful! Rackspace have still not approved any travel15:04
mlavallegood luck with that, john-davidge!15:04
john-davidgemlavalle: haha, thanks15:04
*** galstrom is now known as galstrom_zzz15:04
*** esikache1 has quit IRC15:04
*** esikache1 has joined #openstack-meeting-315:05
mlavalleFinally, the venue / date for the Queens PTG has been announced. September 11th - 15th in Denver15:05
mlavalleI'll start bugging my manager to get a travel approval :-)15:06
*** galstrom_zzz is now known as galstrom15:06
mlavalleAny other annoucements!15:06
john-davidgemlavalle: I can say for certain I will not be there as I leave for my honeymoon on the 16th :)15:06
mlavalleThat's a good reason, Orlando, right?15:07
haleybjohn-davidge: congrats!15:07
john-davidgemlavalle: Yup, we're big kids15:07
john-davidgehaleyb: Thanks!15:07
mlavalle#topic Bugs15:08
haleybmlavalle: i will start asking for approval (and check the calendar)15:08
*** openstack changes topic to "Bugs (Meeting topic: neutron_l3)"15:08
mlavalleFirst one is
openstackLaunchpad bug 1627424 in neutron "FlushError on IPAllocation" [High,In progress] - Assigned to Miguel Lavalle (minsel)15:08
*** rmart04 has quit IRC15:08
mlavalleI proposed a patchset that I think will have impact on this bug15:09
mlavalleIt changes the way IPAllocations are associated to the port:15:09
*** ibmko has joined #openstack-meeting-315:10
mlavalleIt was simpler that originally thought, thanks to the get() method in the sqlalchemy session object15:10
mlavalleTak a look and see what you think15:10
* john-davidge looking15:11
mlavalleI am hoping kevinbenton will have some time soon to review it15:11
*** shuyingya has quit IRC15:11
mlavalleok, moving on15:12
mlavalleNext one is
openstackLaunchpad bug 1610483 in neutron "Pluggable IPAM rollback mechanism is not robust" [High,In progress] - Assigned to Aliaksandr Dziarkach (aliaksandr-dziarkach)15:12
mlavalleAs agreed last week during this meeting, I tried to ping johnbelamaric without success. He doesn't seem to hang out in the neutron channel regularly anymore15:13
*** esikache1 has quit IRC15:14
mlavalleSo last night I sent him an email asking his opinion on the next step15:14
mlavalleHe hasn't responded yet15:14
mlavalleMoving on15:16
mlavalleNext up is
openstackLaunchpad bug 1627480 in neutron "create_port can succeed without returning fixed_ips on all requested subnets" [High,Confirmed] - Assigned to James Anziano (janzian)15:16
mlavalleI spent time in Kibana looking for occurences of this bug15:16
mlavalleAs of an hour ago, we had 3 hits over the past 7 days15:16
mlavalleAll of them were with the same patchset, 45343115:17
haleybwas that patch tweaking ports?15:18
mlavalleWhich is a cherry pick to stable Newton in project openstack/puppet-aodh15:18
mlavalleNope, it uses Neutron stable/Newton15:18
haleyband it's not a neutron patch either15:19
mlavalleso my guess is that this bug has been fixed kevinbenton's refactoring of the of subnets deletes in ml2 and DB plugin15:19
mlavallebecause there are DHCP ports creations and subnets deletes involved in the failure15:20
haleybyes, i think we can close it15:20
mlavalleok cool15:21
mlavalleI'll make a note explaining how we reached the conclusion of closing it and then I'll close it15:21
mlavallesounds good haleyb?15:22
mlavalleThe last one for today is
openstackLaunchpad bug 1509004 in neutron ""test_dualnet_dhcp6_stateless_from_os" failures seen in the gate" [High,Confirmed] - Assigned to James Anziano (janzian)15:22
mlavalleDidn't have time to look at this one. Since I will be closing the previous one soon, I'll give this one some TLC next15:23
mlavalleAny other comments?15:23
*** armstrong has joined #openstack-meeting-315:24
mlavalleAny other bugs from the team?15:24
mlavalleok, moving on15:25
mlavalle#topic DVR non-voting job15:25
*** openstack changes topic to "DVR non-voting job (Meeting topic: neutron_l3)"15:25
haleybi just added an item for that on the etherpad15:26
mlavallehaleyb: this is the topic about the non-voting job that ihrachys brought up15:26
mlavalleso how do we move forward with this?15:26
haleybyes.  there is actually an experimental job i had forgotten about15:26
haleybi think the plan can be move that one to take the place of the dvr-multinode one - they are both non-voting now anyways15:28
*** pcaruana has quit IRC15:28
haleybthen update grafana dash, etc, so we can see how it goes15:28
mlavalleis that something you will do or you need some help?15:28
haleybi can give it a try, of course i might be missing some step that i will learn about along the way15:29
mlavalleok, cool15:29
mlavalleand we can report back to the Neutron CI team during the Tuesday meeting15:30
haleybyes. i'll get a review out to swap the two, i can't imagine we'd want to have both non-voting due to resource contraints15:30
haleybactually, we can probably just retire the existing one15:31
mlavalleyeah, that makes sense15:31
*** marst_ has joined #openstack-meeting-315:31
*** marst has quit IRC15:32
haleybi'll ping ihar about it once i get the patch together15:32
mlavalleMoving on15:32
mlavalle#topic Routed Networks15:32
*** openstack changes topic to "Routed Networks (Meeting topic: neutron_l3)"15:32
mlavallejohn-davidge: we didn't get to the RFE last week during the drivers meeting15:33
*** stevejims has joined #openstack-meeting-315:33
john-davidgemlavalle: Do you think there would be some value in getting started on a WIP before it's been approved?15:33
john-davidgemlavalle: I have the cycles15:33
mlavalleYes, I think so15:34
*** mjturek has joined #openstack-meeting-315:34
john-davidgemlavalle: Ok, I'll do that then15:34
john-davidgemlavalle: thanks15:34
mlavalleI also want to contribute to that, so maybe we can coordinate outside this meeting15:34
mlavallejohn-davidge ^^^15:34
john-davidgemlavalle: Sure, perhaps I'll work on a first draft and ping you if I hit any roadblocks?15:35
mlavalle]That makes sense15:35
mlavalleMoving on15:35
mlavalle#topic Open Agenda15:36
*** openstack changes topic to "Open Agenda (Meeting topic: neutron_l3)"15:36
mlavalleAny other topics we should discuss today?15:36
ibmkohello, if no topics, I would have one general question about L3 in neutron15:37
*** tobberydberg has joined #openstack-meeting-315:37
mlavalleibmko: sure15:37
*** seanatcisco has joined #openstack-meeting-315:38
ibmkois it possible with current neutron API to develop neutron routers backed by nova instances ?15:38
mlavalleDo you mean the router implemented by a Nova instance?15:38
ibmkoI am looing at it a bit and it seems rarther problematic given that both the router object as well as the nova instance has to have neutron ports allocated and connected to desired networks15:39
*** belmoreira has quit IRC15:39
mlavalleFrom the purely API point of view, it might be possible15:39
mlavalleBut you will have to completely change what is behind the API15:40
haleybnot via the API, but you can deploy an openstack cloud on an openstack cloud if that's what you want15:40
ibmkoand these ports are supposed to be "same" ports with same IP adressess15:40
haleybmlavalle: what i'm saying is you can create a neutron router that happens to be inside a nova instance, but neutron doesn't know that15:41
ibmkohaleyb, thats probably not what we need, we just need routers (or generally L3 devices) which are normal virtual routers and not namespaces on network node nor DVR15:41
*** tobberydberg has quit IRC15:41
mlavallehaleyb: yeah, I see what you say15:41
mlavalleBut I think that what ibmko is saying is that he wants to re-use the API calls to implement routers in a completely different way15:42
*** esikache1 has joined #openstack-meeting-315:43
mlavalleso he needs to change what lies behind the API15:43
mlavalleibmko: did I get you right?15:43
ibmkoyes, well, I was hoping that neutron just defines L3 API and lets people to implement it whatever the way they need, but it doesn't seem to provide this flexibility15:43
*** vgadiraj_ has joined #openstack-meeting-315:43
mlavalleibmko: that is the intent of the plugin architecture15:44
mlavallein fact, in the reference implementation, L# is implemented by a service plugin15:45
ibmkoand given our specific implementation desire - to use nova instances as routers - do you see some viable way witnout breaking the API ?15:45
ibmkoI think from the API point of view, following thing is problematic:15:45
*** sdague has quit IRC15:45
mlavalleibmko: see here
ibmkothe api defines "add_router_interface"15:46
ibmkoand this api says: If the port is already in use, this operation returns the Conflict (409) response code.15:46
ibmkothis is I think what prevents us from doing nova based routers15:46
mlavalleibmko: keep in mind that Neutron's routers are not general purpose L3 routers15:46
*** zhurong has joined #openstack-meeting-315:46
ibmkomlavalle, yeah, this is something I am starting to realize15:47
ibmkobut even with limited scope of what  the neutron router is, it might make sense to implement it using nova instance15:47
*** andreas_s has quit IRC15:48
mlavalleyeah, in principle15:48
*** iyamahat has joined #openstack-meeting-315:48
*** zhurong has quit IRC15:49
mlavalleanything else?15:49
haleybi think that's possible, but you would have to manage it yourself.  in the "old days" my company had notes on creating such things (two ports, no security groups), and having it be the router, but neutron router* commands didn't manage it15:49
*** numans has quit IRC15:49
haleybfor example, the old vpn code was like that15:49
mlavallebecause the semantics change15:49
ibmkoso probably there is no way now of having nova instance under semantics of neutron router API ?15:50
ibmkobtw brocade vyatta did that two years ago15:50
mlavalleI don't think the problem is the Nova implementation itself15:51
ibmkobut I think it is abandoned and they must have hacked something in too in order to make it work15:51
mlavalleThe problem is what you want that router to do for you aqnd that, with the current API, you don't have the semantics to manage that15:51
mlavalleOf course you are going to have to coordinate with Nova in your backend (what lies below the API)15:52
mlavallesend requests to NOva to create the instance for the router15:53
mlavalleYou can also explore the possibility of an API extension15:53
ibmkomlavalle, I can imagine that the management of it can be doable, but where I see the main problem is that current neutron L3 API requires neutron port to be connected to the router object but if the router is represented by nova instance, this nova instances needs the very same neutron port connected too otherwise it cannot communicate into the netwrok15:53
*** donghao has joined #openstack-meeting-315:54
mlavalleIn Neutron, to give a port to an instance, you don't need L315:54
*** donghao has quit IRC15:55
mlavallea port is a core resource and is really managed at L215:55
ibmkomlavalle, uh not sure i understand15:55
*** donghao has joined #openstack-meeting-315:55
*** rosmaita has joined #openstack-meeting-315:55
*** rbak has quit IRC15:55
mlavalleL3 comes into play when you want routers, floating ip's and snat15:55
*** Patifa has joined #openstack-meeting-315:56
ibmkoI know that15:56
mlavalleso you could propose an extension API15:56
mlavalleLet's call it Real Routers extension API15:56
ibmkobut the router object requires a neutron port when calling add_router_interface, and if router is implemented by nova instance, this nova instance needs a port (with same IP adress) too15:57
ibmkoand that is a conflic15:57
mlavallewhere you define your API calls with the semantics you need15:57
*** cdent has joined #openstack-meeting-315:57
mlavalleand then you implement the real router anyway you want15:58
mlavalleCurrent Neutron's L# API is itself an extension API15:58
mlavalleTake a look:
mlavalleAnd after saying this, I will close the meeting because in 1 minute another one is going to start15:59
ibmkothank you very much15:59
mlavalleI don't want to piss-off cdent or edleafe15:59
*** openstack changes topic to "OpenStack Meetings ||"15:59
john-davidgemlavalle: ibmko: let's move into openstack-neutron channel15:59
openstackMeeting ended Thu Apr  6 15:59:40 2017 UTC.  Information about MeetBot at . (v 0.1.4)15:59
openstackMinutes (text):
* edleafe is watching mlavalle 15:59
* cdent waves15:59
* mlavalle waves at cdent and edleafe :-)16:00
*** janzian has left #openstack-meeting-316:00
cdent#startmeeting api_wg16:00
*** dtantsur has joined #openstack-meeting-316:00
cdent#chairs cdent edleafe elmiko16:00
openstackMeeting started Thu Apr  6 16:00:14 2017 UTC and is due to finish in 60 minutes.  The chair is cdent. Information about MeetBot at
openstackUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.16:00
*** openstack changes topic to " (Meeting topic: api_wg)"16:00
openstackThe meeting name has been set to 'api_wg'16:00
cdent#link agenda
*** tosky has joined #openstack-meeting-316:00
cdent#link last meeting
elmikowe may have a few saharans join us16:00
*** john-davidge has left #openstack-meeting-316:00
cdent#topic action items16:01
*** openstack changes topic to "action items (Meeting topic: api_wg)"16:01
cdentelmiko: you were going to make an agenda item about what we do16:01
rosmaitai'm just doing a drive-by, want to thank cdent for looking over and comenting on some glance specs16:01
elmikocdent: doh, did i miss that?16:01
cdentrosmaita: my pleasure16:01
*** stevelle has joined #openstack-meeting-316:01
elmikosorry, i totally missed that16:02
cdentelmiko: if I recall correctly it was in response to suggestions that our job is guidelines, whereas I was asserting that we do other stuff that happens to result in guidelines16:02
cdentwe can punt that until next time or even the summit?16:02
elmikocdent: that sounds familiar, thanks for jarring the dust16:03
elmikoi'm ok with that16:03
cdent#action elmiko to continue thinking about how to discuss and formulate what we do16:03
elmikoit seems like a perfect topic *for* summit16:03
cdentthat was the only action item16:03
cdent#topic new biz16:03
*** openstack changes topic to "new biz (Meeting topic: api_wg)"16:03
cdentit turns out that due to the good graces of the foundation, elmiko and I will be at summit and there will be at least one session (in the usual roundup sense, but with a focus on the interoperability debate)16:04
cdentedleafe: are you still a no?16:04
edleafeyeah, that's not going to change16:04
elmikodepending on how our conversations here go about the purview of the wg, i think we should have an agenda item at summit to discuss it too16:05
cdentare we happy with just one session and a strong presence elsewhere, or should we book a second session (we're allowed up to two)16:05
*** aarefiev is now known as aarefiev_afk16:05
elmikoone on interop, one on the wg scope?16:05
dtantsurs/summit/forum/ right?16:05
elmikois that what it's called now?16:05
cdentdtantsur: heh, yeah, sorry, habit.16:06
dtantsurwe have forums and PTGs :)16:06
elmikoahh, cool. thanks for that dtantsur16:06
cdentelmiko: it's forum for the interact with people bits16:06
elmikothe expo is still the summit?16:06
cdentthe overarching thing is still summit16:06
* dtantsur reminds that people can already start planning for the PTG16:06
elmikoand the ptg is only 3 months away?16:07
cdent#action elmiko and cdent to decide (with input from everyone) about how to do summit16:07
cdent#topic sarah and microversions16:08
*** openstack changes topic to "sarah and microversions (Meeting topic: api_wg)"16:08
cdenttosky and tellesnobrega ?16:08
toskysahara* :)16:08
cdentaw, that's habit too, it's my wife's name :)16:08
cdentmuscle memory16:08
toskyhi, the long standing work to Sahara API v2 is finally moving16:09
*** marst_ has quit IRC16:09
cdentbut I suspect she is not too concerned with microversions (except listening to me curse about them)16:09
edleafeYour wife is in the Big Tent?16:09
*** marst has joined #openstack-meeting-316:09
tosky(work formalized by elmiko some time ago)16:09
cdentprogress \o/16:09
toskyso one of the questions was how to use microversioning; speaking with elmiko the impression is that, until we declare the API as stable,16:09
elmikocdent: indeed!16:09
* dtantsur has hard time writing "Iron" without ending up with "Ironic" (and I made this mistake in this sentence too!)16:09
toskywe don't really need microversioning, as we can tune it on the way slightly16:09
*** marst_ has joined #openstack-meeting-316:10
toskyand that the microversioning part, if needed, would start after stabilizing the API16:10
dtantsurtosky, what's your definition of "stable" here?16:10
toskydtantsur: there is an official definition, iirc16:10
cdenttosky there are a lot of different perspectives on that and it depends quite a bit on whether you consider yourself a CD project16:10
dtantsurthere are many, I suspect, I wonder which one you use ;)16:10
elmikoi was taking "stable" to mean when the sahara team actually released it as official to users16:11
dtantsurelmiko, I suspect it's not a common definition (though I like it)16:11
elmikodtantsur: yeah, i could be off in left field here16:11
toskywhat are the possible states for an API? Only STABLE and DEPRECATED?16:11
cdentIf sahara does not consider it self continuously deployed, then until there is a release (at the end of the cycle) then while building the API versions can be a hassle16:11
elmikoi thought we had EXPERIMENTAL too16:12
dtantsurelmiko, we should, at least16:12
elmikomaybe we need to adjust that guideline16:12
* dtantsur wonders if wording like "EXPERIMENTAL" or "UNSTABLE" may frighten people16:13
toskydtantsur: they should!16:13
*** kiseok7 has joined #openstack-meeting-316:13
cdentthe microversions guidelines which describes version states is explicitly for microversions and makes the assumption that the thing, whatever it is, is already released16:13
dtantsurprobably you're right, yes :)16:13
toskyif you use something EXPERIMENTAL or UNSTABLE, you are on your own (Debian teached that for decades)16:13
*** marst has quit IRC16:13
elmikoi totally agree, people should take notice at those terms16:14
toskyok, so let's go back16:14
cdenttosky: the trouble with that for an API is that seeing that experimental-ness is not always assured16:14
toskycdent: well, I mean16:14
toskycaveat emptor16:14
elmikofrom , the statuses are CURRENT, SUPPORTED, DEPRECATED, and EXPERIMENTAL16:14
elmikoedleafe: jinx!16:14
dtantsurfrom all my experience, if people want something in production, they'll use it no matter what and complain if you break it16:14
toskydtantsur: we will not enable it by default16:15
toskyI'm pretty sure we are not big enough for people to start using it16:15
dtantsuryou bet :)16:15
elmikoyeah, i agree tosky16:15
toskyand well, about using new API and switching to them, we have many examples were we can't kill the old APIs :)16:15
elmikothis is a good problem to study with sahara as a pattern for other projects given its size16:15
cdenttosky: can we role back to whatever the original question is? What are you trying to decide?16:15
toskyso, two things:16:16
edleafedtantsur: true, but at some point we can justify not wasting effort for those cases16:16
dtantsuredleafe, this is also true, yeah16:16
tosky- EXPERIMENTAL state, then CURRENT/SUPPORTED and only at that point we can really need microversioning, correct?16:16
toskybut now I have a more personal question16:16
toskyabout microversioning, as I'm not one of the people exactly happy about microversioning as it stands16:17
edleafetosky: you should lay the groundwork for microversions ASAP, but you would never worry about bumping an experiemental version16:17
elmikoagreed with edleafe16:17
cdentagreed with elmiko agreeing with edleafe16:17
toskyedleafe: just to clarify: do we need to bump the version even experimental APIs?16:18
edleafeagreed with anyone agreeing with me16:18
toskyif it's experimental, I thought that it can break anytime16:18
elmikotosky: no, no need to bump experimental16:18
dtantsurtosky, my understanding is "no". just make sure you provide the correct information in the root and /v{MAJOR} endpoints16:18
cdenttosky: no, that's what he's trying to say: put in the infrastructure for handling microversions but don't bump the version16:18
edleafetosky: no - experiemental is the equivalent of alpha software16:18
*** numans has joined #openstack-meeting-316:18
cdentonce it is not experimental, then you start bumping16:18
dtantsurtosky, i.e. when I started ironic-inspector, I did not think about it, and we had to fix / as we went16:18
toskyok, we are on the same page - sorry for asking again16:19
* dtantsur did not think about many useful things back then, to be fair16:19
edleafetosky: no worries16:19
toskyback to microversioning in general16:19
elmikotosky: i said something similar in the sahara meeting, you should lay the foundation for microversions now, but don't do anything to advance them until the api is released as current/supported16:19
toskyright now the microversioning allows you to create incompatible changes, which are hidden until you request that specific microversion16:19
cdentpretty much, yes16:20
dtantsur*any* changes are hidden, modulo maybe critical bug fixes16:20
*** iyamahat has quit IRC16:20
toskywhat I would like to try to implement for Sahara (but I'm only a QA, so it's not up to me too), given that we have time to stabilize it, is to try to use microversioning with non-breaking changes16:20
toskymore or less the equivalent of "keep the ABI" in C/C++ world16:21
edleafetosky: any change is considered "breaking" in microversion land16:21
cdentthat's pretty much how it's intended, already16:21
cdentmake any change, new version16:21
toskythis is how it is now16:22
toskybut let's say we implement it in a way to keep always a theretical compatibility with old clients16:22
dtantsurtosky, actually, we do it for ironic-inspector (more or less historically, but I doubt anyone cares enough to change it).16:22
dtantsurI still prefer it to the official api-wg way, but I'm badly outnumbered in it16:23
toskyfor example introducing only changes which do not change the signature of methods, or only add new parameters in the end with default values16:23
edleafetosky: are we talking code methods, or API methods?16:23
cdentdoing that is good behavior, but if you're doing microversions, those types of changes are still another microversion16:23
toskyAPI methods16:23
toskyyes, it is16:24
*** ralonsoh has quit IRC16:24
toskybut would it be possible to have a recognition, going forward, for this kind of practice?16:24
toskyI'm not suggesting to change anything for people who want to use microversioning as they are now16:24
toskyI'm just asking for the possibility to recognize this behavior going forward16:24
cdenttosky: I think mordred is thinking about something along those lines16:24
elmikoimo, i see this as a social contract at the project level and it sounds like tosky would like something in the guidelines acknowledging this behavior. is that accurate tosky?16:25
cdentwith a "never breaks compatibility" tag of some kind16:25
toskyyeah, that's the idea16:25
toskyno one harmed, just something more for people who want to try to be more... conservative16:25
cdentI think in the guidelines is the wrong place, as we get combinatorial explosions of explanation in the guidance16:25
elmikoi like the "never breaks compat." tag, that speaks volumes to the work the team is doing16:25
cdentbut a tag that a project could assert would be nice16:25
toskyfrom my QA point of view, keeping the compatibility is a bit more costly when you plan it, but it reduces the test matrix16:26
toskywhich I really want to keep simple :)16:26
edleafeMost people are assuming this when talking about "nnon-breaking" changes:
cdenthere's the tag:16:26
toskyas I said, my background is more this:
edleafeThat doesn't cover multi-cloud interoperability, though16:27
cdent#link never break compat:
edleafe#link Tolerant reader
toskythat spec seems exactly what I was thinking16:28
cdenttosky: I'd recommend you comment on there if you'd like to see such a thing exist16:28
toskyoh, sure16:29
cdentthere was some discussion at a TC meeting but the reaction was mixed16:29
toskybut why? It does not take away anything for people who don't want to implement it16:29
toskyI see it as something more16:29
elmikowelcome to governance?16:29
cdentthere was concern that it was hard to intepret16:29
*** Patifa has quit IRC16:29
dtantsurplaying devil's advocate a bit: too many ways to version API make it harder to consume various services via some common code16:29
cdentdtantsur: don't think tosky is really talking about versioning here, something next to it16:30
edleafeYeah, this is the "till death do us part" mindset to APIs16:30
toskyversioning come close, but technically with such rules implemented, a client for x.y would work with any future x.y+n16:30
toskyif it could ask for that x.y+n16:31
dtantsurplaying devil's advocate again: a badly written client can break on new fields introduced in a response16:32
dtantsur(I've seen such examples)16:32
toskybut that's a bug in the client16:32
edleafeNot all agree that the problem is the client16:33
elmikoideally, wouldn't the client be requesting x.y?16:33
toskyit depends on how much you are clear on your rules16:33
*** ibmko has quit IRC16:33
edleafeAdding a field to a response means a client can't use a "for each" pattern in processing the response16:33
elmikoi mean, once you introduce microversions, won't all clients need to use them or default to the minimum version provided?16:33
*** iyamahat has joined #openstack-meeting-316:33
elmikoi thought there was guidance on this16:34
*** vgadiraj_ has quit IRC16:34
toskyright, but if you say it from the beginning (you will get *at least* fields), you should plan accordingly16:34
cdent(or default to latest and run with fire)16:34
elmikocdent: yeah, one or the other, i can't remember what the guide saysa16:34
toskyelmiko: yes, with microversioning that's the case; I was talking about the technical possibility if you use stricter rules16:34
elmikotosky: i'm not following how this would be possible, if you release v2 with microversions then by definition all sahara v2 clients will need to use them16:34
toskyelmiko: and (possible) evolution for people using this in a far future :)16:35
cdenttosky: it would be nice if the world was like that, but when you say "you say it" who is listening? can you guarantee people will listen? no. do you still want to be useful and usable16:35
edleafeelmiko: the whole point of microversions is you don't get a changed response unless you specifically ask for it16:35
*** abalutoiu has quit IRC16:35
elmikoedleafe: right16:35
toskycdent: but if you start from this point, then we can close everything and go work on agriculture16:35
elmikowhich is why i'm stumped how a client expecting x.y would choke on x.y+1, it should never request x.y+116:35
cdenttosky: I totally hear what you are saying. It is an argument I have made myself. But there's been a lot of discussion which has ended up being pretty resigned to the fact that we should make the HTTP APIs robust in the face of people not paying attention.16:36
dtantsurelmiko, this is all, however, assuming that no abstractions are leaky16:37
toskycdent: that's fine; that's why having at least recognition about this tag could help at least, even if there is microversioning16:37
toskywhich hides the strictness implemented16:37
dtantsurif you have abstractions leaking so heavily as Ironic's ones, it can never be strictly implemented16:37
cdenttosky: Yeah, I think you're right. I think doing both microversions and what you are describing (in some form) is a good thing. Overall we want to minimize changes to APIs once they are published.16:37
toskyso don't implement it :)16:37
edleafedtantsur: well, in that case, no versioning approach will work16:38
toskyand then, in few years, we can discuss again wheater we can allow a "free for all" for strict clients :)16:38
elmikodtantsur: good point16:38
cdenttosky: and of course you are always welcome to not do microversions, or put off using them for a long time16:38
cdentI think it is fine to take a long time to stabilize an API16:38
dtantsuredleafe, s/work/be perfect/16:38
dtantsureven leaky abstractions can be useful ;)16:39
elmikoi think that's a strong point cdent , especially for sahara16:39
cdent(but that's my personal opinion, not official guidance, etc)16:39
edleafedtantsur: "work" in the sense of keeping every consumer happy16:39
*** iyamahat has quit IRC16:39
elmikodoes the complexity of microversions really buy anything for the sahara project, i tend to think no16:39
toskycdent: that's my hope; the API has been redesigned from scratch, it is taking some time, I hope that version bumps won't really be needed16:39
edleafecdent: that's my opinion, too16:40
elmikowell, i layed the foundation, so of course it won't need changing! /s16:40
* cdent proposes we rename the group from api-wg to microversion-support-group16:40
* dtantsur +116:40
*** tonytan4ever has quit IRC16:40
elmikocdent: maybe that should be our second session at the forum ;)16:40
edleafecdent: nah. MSG will always be Madison Square Garden to this New Yorker16:40
*** tonytan4ever has joined #openstack-meeting-316:40
elmikoi'll bring the tissue16:40
cdenttypical new yorker, thinks they are the center of the universe ;)16:41
toskyI think it's all from me for this point, at least for now (and probably for the next two or three cycles :)16:41
* edleafe notes that we actually *are* the universe16:41
cdenttosky: feel free to post the mailing list with questions and comments and what not as you progress. I think having more conversation in email about this stuff would draw out lots of ideas16:42
cdentmy only other open biz topic is that I renamed the compability guideline to interoperability because I felt like that was how it had to go:16:42
elmikocdent: +116:42
cdent#link interop:
edleafe+1 on the name change16:42
toskyyep, at least for this cycle we will mostly work on the EXPERIMENTAL side16:42
cdentwe need to bring that one to a close and perhaps also start a new one for the issue that edleafe brought up16:42
cdentbut we need some completion/closure/motion16:43
cdentif the people here can review it, yet again, maybe by next week we can freeze it16:43
elmikoack, added to my queue16:44
* dtantsur too16:44
cdent#topic guidelines16:44
*** openstack changes topic to "guidelines (Meeting topic: api_wg)"16:44
cdentlooks like everything needs either more review, or tuning16:45
cdentthe http tags thing is close, but hasn't actually had that many eyes on it16:45
cdentanyone want to highlight an issue with a guideline?16:46
cdent#topic bug review16:46
*** openstack changes topic to "bug review (Meeting topic: api_wg)"16:46
edleafeThe pagination issue you commented on16:46
cdentno new bugs16:46
edleafewow, you're too fast16:47
cdentedleafe: see, I have expectations about your typing. I've been primed16:47
cdentoh the 400 or 404 thing on pagination, yeah. your thoughts?16:47
edleafeWell, to be fair, I was looking at the tags first16:47
edleafeOn 400 vs. 404, neither is correct16:47
edleafewe have to decide which is less wrong16:48
cdent404 is more wrong because of why I said :)16:48
cdentand 400 is the fallback when nothing else seems to fit16:48
edleafeWell, I'd appreciate some others to act as tie-breakers16:48
cdentin case it is not clear we are back on16:48
elmikotechnically it is a client side error16:49
cdent#topic guidlines16:49
*** openstack changes topic to "guidlines (Meeting topic: api_wg)"16:49
openstackRemoving item from minutes: #topic guidlines16:49
cdent#topic guidelines16:49
*** openstack changes topic to "guidelines (Meeting topic: api_wg)"16:49
*** VW has quit IRC16:50
cdent#link pagination guideline
*** VW has joined #openstack-meeting-316:50
cdentanyone else besides elmiko want to weigh  in?16:50
* dtantsur will, but he did not read yet16:51
elmikoyeah, guess i'm in the 400 camp now16:51
cdentcool, thanks16:51
edleafeelmiko: all 4xx are client-side errors16:51
edleafe400 is simply the fallback16:51
elmikoedleafe: right, it just seemed like there was no more specific error16:52
elmikoi mean, i can understand the 404 argument, but i like the way cdent stated it in his comments on the review16:52
dtantsur"there is no specific error" is a typical problem with HTTP API16:52
elmikosadly, yes16:52
dtantsurI can tell you horror stories about 409 Conflict in Ironic :)16:53
edleafeI'll respond to cdent on the review16:53
cdentdtantsur: we've got those in placement too, but there's a plan to start using the errors guideline to address it16:53
elmikoi was just stating the obvious that since its a client side error and there is nothing more specific that 400 was probably the best16:53
*** VW has quit IRC16:53
*** Patifa has joined #openstack-meeting-316:53
*** VW has joined #openstack-meeting-316:54
*** armstrong has quit IRC16:54
dtantsur"410 Gone16:54
dtantsur    Indicates that the resource requested is no longer available and will not be available again. "16:54
*** rmart04 has joined #openstack-meeting-316:54
dtantsur(yes, I know that you want to kill me now :)16:54
cdentdefinitely not 410, that means something very specific: never ever try this URI again16:54
elmikobut to cdent's point, the resource (the collection) *is* still there16:54
cdentfive minute warning16:55
dtantsurmy only problem with 400 is that it's hard to distinguish from other bad requests16:55
*** VW has quit IRC16:55
elmikothat's why we have the error guideline!16:55
dtantsurso e.g. on receiving a 404, a client could try to re-get the marker16:55
dtantsuror just start from the beginning of the collection16:55
dtantsurwith 400 it's probably doomed to just fail16:55
cdentI think in this case, no matter what the response code, there's going to need to be special handling16:55
*** VW has joined #openstack-meeting-316:55
elmikoah, i see, you're talking about programmatically resolving the error16:55
dtantsurelmiko, yep16:56
edleafedtantsur: or re-request the previous page, which should now have a new marker16:56
elmikocdent: i agree16:56
cdentin which case 400, because 404 already has clear meaning: this uri is not found16:56
cdentand that will be misleading in many ways16:56
dtantsurwith just 400 you never know if the request is broken or what16:56
edleafedtantsur: that's why the error message is super-critical16:57
cdentyes, but that's the true case in this case: something went wrong, but it's non obvious16:57
elmikoyeah, i can see how this is going to be tough to resolve16:57
dtantsuredleafe, assuming you want clients to parse them16:57
dtantsurcdent, my definition of "this uri is not found" allows this actually16:57
edleafewe are discussing an edge case of an edge case, just for perspective16:57
elmikodtantsur: to play devil's advocate (which you seem to enjoy ;) ), how will your client know that the pagination request had a bad marker and the whole URI was bad though?16:57
cdentdtantsur: strictly speaking, yes, because the query param is part of the uri, but from a human standpoint, people don't think like that16:57
cdent2 minute warning16:58
cdentgiven that, I think we should punt this fun discussion to the review16:58
edleafeHumans?? Where did they come into the discussion??16:58
dtantsurcdent, many do, but I see your point16:58
elmikoedleafe: LOL16:58
dtantsuryeah, I'll think about it til tomorrow morning, I'm not the brightest thinker at 7pm16:58
cdentedleafe: I don't know, for that, you might have to do the newsletter16:58
* edleafe proposes changing our name to the Human Programming Interface Working Group16:59
elmikoi could do the newsletter, but it won't go out for a few hours, i have solid meetings till about 416:59
cdent(but if you haven't got the cycles I will, but need to go have a cup of tea and a sit down in the sun outside)16:59
elmikoedleafe: LOL, love it!16:59
edleafeI'm kinda booked for the next 2 hours17:00
cdentokay I'll do it17:00
cdentthanks everyone!17:00
*** openstack changes topic to "OpenStack Meetings ||"17:00
elmikoso, cdent + tea + sunshine + newsletter?17:00
openstackMeeting ended Thu Apr  6 17:00:25 2017 UTC.  Information about MeetBot at . (v 0.1.4)17:00
*** baoli has quit IRC17:00
openstackMinutes (text):
elmikocdent: thanks!17:00
*** dtantsur has left #openstack-meeting-317:00
*** tosky has left #openstack-meeting-317:01
*** rmart04 has quit IRC17:01
*** donghao has quit IRC17:03
*** stevelle has left #openstack-meeting-317:03
*** rbak has joined #openstack-meeting-317:03
*** rossella_s has quit IRC17:04
*** mickeys has quit IRC17:05
*** mickeys has joined #openstack-meeting-317:06
*** lhx_ has quit IRC17:06
*** ltomasbo is now known as ltomasbo|away17:08
*** baoli has joined #openstack-meeting-317:11
*** jtomasek has quit IRC17:16
*** srobert has left #openstack-meeting-317:16
*** donghao has joined #openstack-meeting-317:17
*** srobert has joined #openstack-meeting-317:17
*** rbak has quit IRC17:19
*** harlowja has joined #openstack-meeting-317:23
*** donghao has quit IRC17:27
*** alexismonville has joined #openstack-meeting-317:28
*** seanatcisco has quit IRC17:36
*** jtomasek has joined #openstack-meeting-317:38
*** Swami has joined #openstack-meeting-317:50
*** mlavalle has quit IRC17:53
*** seanatcisco has joined #openstack-meeting-317:54
*** Patifa has quit IRC17:57
*** soliosg has left #openstack-meeting-317:57
*** Patifa has joined #openstack-meeting-317:58
*** Patifa has quit IRC17:58
*** Patifa has joined #openstack-meeting-317:59
*** Patifa has quit IRC18:00
*** Patifa has joined #openstack-meeting-318:01
*** Patifa has quit IRC18:04
*** s3wong has joined #openstack-meeting-318:08
*** Sukhdev has joined #openstack-meeting-318:18
*** lucasxu has joined #openstack-meeting-318:22
*** donghao has joined #openstack-meeting-318:25
*** alexismonville has quit IRC18:26
*** donghao has quit IRC18:31
*** esikache1 has quit IRC18:33
*** sdague has joined #openstack-meeting-318:34
*** esikache1 has joined #openstack-meeting-318:35
*** esikache1 has quit IRC18:39
*** stevejims has quit IRC18:40
*** ibmko has joined #openstack-meeting-318:44
*** Rockyg has joined #openstack-meeting-318:47
*** esikache1 has joined #openstack-meeting-318:48
*** dtroyer has joined #openstack-meeting-318:54
*** seanatcisco has quit IRC18:56
*** esikache1 has quit IRC18:58
*** rbak has joined #openstack-meeting-319:00
*** baoli has quit IRC19:03
*** tobberydberg has joined #openstack-meeting-319:03
*** esikache1 has joined #openstack-meeting-319:03
*** galstrom is now known as galstrom_zzz19:03
*** cdent has quit IRC19:06
*** choyj has quit IRC19:07
*** esikache1 has quit IRC19:12
*** Rockyg has quit IRC19:19
*** esikache1 has joined #openstack-meeting-319:19
*** esikache1 has quit IRC19:24
*** esikache1 has joined #openstack-meeting-319:25
*** baoli has joined #openstack-meeting-319:26
*** donghao has joined #openstack-meeting-319:28
*** galstrom_zzz is now known as galstrom19:33
*** donghao has quit IRC19:33
*** esikache1 has quit IRC19:38
*** openstack has joined #openstack-meeting-319:47
*** ChanServ sets mode: +o openstack19:47
*** esikache1 has quit IRC19:51
*** annegentle has joined #openstack-meeting-319:59
*** rbak has quit IRC20:01
*** esikache1 has joined #openstack-meeting-320:01
*** harlowja has quit IRC20:03
*** scottda has quit IRC20:05
*** rockyg has joined #openstack-meeting-320:05
*** alaski has quit IRC20:11
*** alaski has joined #openstack-meeting-320:11
*** mariannelm has quit IRC20:12
*** zerick has quit IRC20:21
*** zerick has joined #openstack-meeting-320:21
*** markmcclain has quit IRC20:26
*** markmcclain has joined #openstack-meeting-320:27
*** Sukhdev has quit IRC20:31
*** iyamahat has joined #openstack-meeting-320:32
*** abalutoiu has joined #openstack-meeting-320:32
*** rbak has joined #openstack-meeting-320:33
*** tonytan4ever has quit IRC20:37
*** iyamahat has quit IRC20:37
*** esikache1 has quit IRC20:40
*** david-lyle has quit IRC20:44
*** esikache1 has joined #openstack-meeting-320:46
*** kzaitsev_mb has joined #openstack-meeting-320:53
*** cleong has quit IRC20:54
*** tobberydberg has quit IRC20:55
*** annegentle has quit IRC20:59
*** iyamahat has joined #openstack-meeting-320:59
*** esikache1 has quit IRC21:00
*** dtroyer has left #openstack-meeting-321:02
*** harlowja has joined #openstack-meeting-321:03
*** kzaitsev_mb has quit IRC21:03
*** rockyg has quit IRC21:05
*** lamt has joined #openstack-meeting-321:11
*** gouthamr has quit IRC21:11
*** dimtruck is now known as zz_dimtruck21:12
*** marst_ has quit IRC21:13
*** seanatcisco has joined #openstack-meeting-321:16
*** alexpilotti has joined #openstack-meeting-321:18
*** donghao has joined #openstack-meeting-321:23
*** spzala has quit IRC21:26
*** spzala has joined #openstack-meeting-321:26
*** annegentle has joined #openstack-meeting-321:26
*** donghao has quit IRC21:29
*** Sukhdev has joined #openstack-meeting-321:29
*** spzala has quit IRC21:31
*** haleyb has quit IRC21:35
*** gouthamr has joined #openstack-meeting-321:36
*** lucasxu has quit IRC21:36
*** rbak has quit IRC21:42
*** bzhao_ has joined #openstack-meeting-321:43
*** zz_dimtruck is now known as dimtruck21:43
*** srobert_ has joined #openstack-meeting-321:43
*** srobert has quit IRC21:46
*** srobert_ has quit IRC21:48
*** baoli has quit IRC21:55
*** galstrom is now known as galstrom_zzz21:58
*** openstack has joined #openstack-meeting-322:03
*** ChanServ sets mode: +o openstack22:03
*** iyamahat has quit IRC22:03
*** haleyb has joined #openstack-meeting-322:04
*** lyan has quit IRC22:10
*** bobh has quit IRC22:15
*** vishwana_ has quit IRC22:17
*** esikache1 has joined #openstack-meeting-322:18
*** iyamahat has joined #openstack-meeting-322:23
*** esikache1 has quit IRC22:24
*** alexismonville has joined #openstack-meeting-322:33
*** esikache1 has joined #openstack-meeting-322:34
*** alexismonville has quit IRC22:38
*** gugl_ has joined #openstack-meeting-322:40
*** esikache1 has quit IRC22:41
*** iyamahat has quit IRC22:51
*** tobberydberg has joined #openstack-meeting-322:56
*** tobberydberg has quit IRC23:00
*** alexpilotti has quit IRC23:02
*** alexpilotti has joined #openstack-meeting-323:05
*** seanatcisco has quit IRC23:06
*** seanatcisco has joined #openstack-meeting-323:07
*** alexpilotti has quit IRC23:08
*** alexpilotti has joined #openstack-meeting-323:09
*** alexpilotti has quit IRC23:12
*** alexpilotti has joined #openstack-meeting-323:13
*** alexpilotti has quit IRC23:15
*** alexpilotti has joined #openstack-meeting-323:15
*** spzala has joined #openstack-meeting-323:22
*** bobh has joined #openstack-meeting-323:24
*** gugl has quit IRC23:26
*** gugl_ has quit IRC23:27
*** Swami has quit IRC23:28
*** donghao has joined #openstack-meeting-323:28
*** esikache1 has joined #openstack-meeting-323:29
*** donghao has quit IRC23:34
*** esikache1 has quit IRC23:34
*** markvoelker has quit IRC23:37
*** neiljerram has quit IRC23:43
*** esikache1 has joined #openstack-meeting-323:43
*** esikache1 has quit IRC23:48
*** david-lyle has joined #openstack-meeting-323:48
*** claudiub has quit IRC23:48
*** claudiub has joined #openstack-meeting-323:48
*** lyan has joined #openstack-meeting-323:59

Generated by 2.14.0 by Marius Gedminas - find it at!