Thursday, 2018-05-31

« Wednesday, 2018-05-30 Index Friday, 2018-06-01 »
*** longkb has joined #openstack-lbaas00:26
*** JudeC_ has quit IRC01:05
*** yamamoto has joined #openstack-lbaas01:14
*** hongbin has joined #openstack-lbaas01:16
*** harlowja has quit IRC01:19
openstackgerritLingxian Kong proposed openstack/octavia master: Use HMAC.hexdigest to avoid non-ascii characters for package data  https://review.openstack.org/57133301:31
*** thomasem has quit IRC01:59
*** thomasem has joined #openstack-lbaas01:59
*** annp has joined #openstack-lbaas02:13
openstackgerritMichael Johnson proposed openstack/octavia master: Implement provider drivers - Driver Library  https://review.openstack.org/57135802:16
openstackgerritMichael Johnson proposed openstack/octavia master: Implement provider drivers - Driver Library  https://review.openstack.org/57135802:18
johnsomStill WIP, but a decent start02:20
*** annp has quit IRC03:25
*** annp has joined #openstack-lbaas03:31
*** rcernin has quit IRC03:35
*** ianychoi has quit IRC03:51
*** hongbin has quit IRC03:52
*** rcernin has joined #openstack-lbaas03:54
*** JudeC_ has joined #openstack-lbaas04:08
*** rcernin has quit IRC04:28
*** rcernin has joined #openstack-lbaas04:29
openstackgerritLingxian Kong proposed openstack/octavia master: Use HMAC.hexdigest to avoid non-ascii characters for package data  https://review.openstack.org/57133304:29
*** links has joined #openstack-lbaas04:38
*** AlexeyAbashkin has joined #openstack-lbaas05:04
*** JudeC_ has quit IRC05:19
*** AlexeyAbashkin has quit IRC05:23
*** AlexeyAbashkin has joined #openstack-lbaas05:28
*** AlexeyAbashkin has quit IRC05:35
*** imacdonn has quit IRC06:16
*** imacdonn has joined #openstack-lbaas06:16
*** gans has joined #openstack-lbaas06:26
*** kobis has joined #openstack-lbaas06:28
*** pcaruana has joined #openstack-lbaas06:43
*** nmanos has joined #openstack-lbaas07:01
*** apple01 has joined #openstack-lbaas07:05
*** rcernin has quit IRC07:11
*** JudeC_ has joined #openstack-lbaas07:15
*** apple01 has quit IRC07:17
*** tesseract has joined #openstack-lbaas07:22
*** apple01 has joined #openstack-lbaas07:26
*** AlexeyAbashkin has joined #openstack-lbaas07:30
*** ianychoi has joined #openstack-lbaas07:32
*** apple01 has quit IRC07:38
*** gans has quit IRC07:50
*** JudeC_ has quit IRC07:56
*** JudeC_ has joined #openstack-lbaas07:57
*** velizarx has joined #openstack-lbaas08:11
*** rcernin has joined #openstack-lbaas08:12
*** rcernin has quit IRC08:12
sapd_rm_work:  how can I get loadbalancer id in amphora instance?08:13
*** apple01 has joined #openstack-lbaas08:13
openstackgerritLingxian Kong proposed openstack/octavia master: Use HMAC.hexdigest to avoid non-ascii characters for package data  https://review.openstack.org/57133308:43
*** JudeC_ has quit IRC08:44
*** yboaron has joined #openstack-lbaas08:59
openstackgerritzhulingjie proposed openstack/octavia master: Update overriden to overridden  https://review.openstack.org/57141909:03
*** salmankhan has joined #openstack-lbaas09:08
*** yboaron has quit IRC09:24
*** salmankhan1 has joined #openstack-lbaas09:32
*** salmankhan has quit IRC09:34
*** salmankhan1 is now known as salmankhan09:34
*** apple01 has quit IRC09:37
*** apple01 has joined #openstack-lbaas09:37
*** apple01 has quit IRC09:50
*** apple01 has joined #openstack-lbaas09:57
*** kobis has quit IRC10:01
numanscgoncalves, ping10:01
*** apple01 has quit IRC10:13
*** apple01 has joined #openstack-lbaas10:15
*** apple01 has quit IRC10:16
*** salmankhan has quit IRC10:33
*** kobis has joined #openstack-lbaas10:37
*** salmankhan has joined #openstack-lbaas10:49
*** longkb has quit IRC10:54
*** salmankhan has quit IRC10:54
*** salmankhan has joined #openstack-lbaas10:55
*** apple01 has joined #openstack-lbaas10:57
*** annp has quit IRC10:57
*** salmankhan has quit IRC11:01
*** tesseract has quit IRC11:07
*** tesseract has joined #openstack-lbaas11:08
*** salmankhan has joined #openstack-lbaas11:11
*** jiteka has quit IRC11:21
*** jcarpentier has joined #openstack-lbaas11:23
openstackgerritZhaoBo proposed openstack/octavia master: UDP for [3][5][6]  https://review.openstack.org/53939111:26
*** yamamoto has quit IRC11:41
*** yboaron has joined #openstack-lbaas11:42
*** atoth has joined #openstack-lbaas11:59
openstackgerritKobi Samoray proposed openstack/octavia master: Providers: propose changes to API  https://review.openstack.org/57145312:02
*** velizarx has quit IRC12:08
*** velizarx has joined #openstack-lbaas12:11
*** amuller has joined #openstack-lbaas12:11
*** kobis has quit IRC12:22
*** kobis has joined #openstack-lbaas12:28
*** yamamoto has joined #openstack-lbaas12:41
*** yamamoto has quit IRC12:46
*** yamamoto has joined #openstack-lbaas12:56
*** apple01 has quit IRC13:19
*** apple01 has joined #openstack-lbaas13:28
*** links has quit IRC13:29
*** fnaval has joined #openstack-lbaas14:15
*** kobis has quit IRC14:50
*** yboaron has quit IRC15:14
*** Swami_ has joined #openstack-lbaas15:15
*** pcaruana has quit IRC15:28
cgoncalvesnumans, hi. I'm off today (bank holiday)15:36
*** AlexeyAbashkin has quit IRC15:37
*** apple01 has quit IRC15:47
*** kobis has joined #openstack-lbaas15:48
johnsomEveryday is a holiday for Carlos....15:49
*** ispp has joined #openstack-lbaas16:01
*** kobis has quit IRC16:03
*** JudeC_ has joined #openstack-lbaas16:05
*** kobis has joined #openstack-lbaas16:06
*** kobis has quit IRC16:07
openstackgerritMichael Johnson proposed openstack/octavia master: Allow DB retries on controller_worker creates  https://review.openstack.org/57110716:07
*** ispp has quit IRC16:14
*** amotoki has quit IRC16:35
*** rpittau has quit IRC16:36
*** amotoki has joined #openstack-lbaas16:38
*** SumitNaiksatam has joined #openstack-lbaas16:56
johnsomSo evidently since we added Castellan support Octavia can't be labeled "inter-operable" since there would be multiple workflows to upload certs and keys.  Ugh, I think that is a bit strange. It sounds like they require one and only one option for everything.17:05
*** JudeC_ has quit IRC17:05
*** JudeC_ has joined #openstack-lbaas17:06
openstackgerritBen Nemec proposed openstack/python-octaviaclient master: update cmd2 dependency to handle py3 only versions  https://review.openstack.org/57152717:09
*** salmankhan has quit IRC17:20
*** atoth has quit IRC17:22
rm_workjohnsom: there's only one option *per deployment*...17:22
*** atoth has joined #openstack-lbaas17:22
johnsomYeah, but the fact that there is an option at all seems to be their issue.17:23
johnsomI'm not sure I care at the moment.17:23
rm_workit seems like the response to that would be "ok, we'll remove that driver", which would make it LESS inter-operable17:23
johnsomIt was part of a larger TC discussion about barbican/castellan,  etc17:23
rm_workso i'm not sure what their definition of inter-operable is17:23
johnsomRight, that is what I asked and basically what I got back17:24
johnsomI guess we should require everything to go through our API so that our flexibility is hidden, but...  eh17:25
xgerman_The time we spend on BBQ…17:25
johnsomThen it devolved into confusion about doing signing/decryption inside HSMs, etc...17:27
*** AlexeyAbashkin has joined #openstack-lbaas17:32
*** JudeC_ has quit IRC17:35
*** JudeC_ has joined #openstack-lbaas17:39
*** AlexeyAbashkin has quit IRC17:45
*** AlexeyAbashkin has joined #openstack-lbaas17:45
*** AlexeyAbashkin has quit IRC17:45
*** AlexeyAbashkin has joined #openstack-lbaas17:46
rm_workjohnsom: did you realize haproxy can take a config dir with `-f`?17:46
rm_worksince 1.717:46
rm_workhttps://cbonte.github.io/haproxy-dconv/1.7/management.html#3-where17:47
johnsomUmm, maybe.  Why do you ask?  We are using two configs now, but not by directory17:47
rm_workyeah17:47
rm_workso17:47
johnsomPlus, well, CentOS17:47
rm_worklol17:47
rm_workwell that 1.8 patch can merge now :)17:47
rm_workhttps://review.openstack.org/#/c/548420/17:48
rm_workfeel free to +A17:48
rm_work"It should be safe to use that repo. In fact, in Red Hat OSP 14 (Rocky-based) we're trying to cross-tag it from same (or almost) source as that repo. We'd still RDO to cross-tag & distribute such package. For now, LGTM."17:49
johnsomSo, again, do we have a need for that now or  ?17:50
rm_workit's just a comment17:57
rm_workit's interesting17:57
rm_workif we look at consolidating processes...17:57
rm_workwe can just do a single config dir, and switch to one file per listener, like .... ${LISTENER_ID}.cfg17:58
rm_workand not have to worry about chaining -f17:58
johnsomYeah, would be interesting to see how it merges the files17:59
rm_workit says in the doc, it's by newest17:59
rm_workso it applies from oldest to newest, overwriting anything that's duplicated17:59
rm_workTECHNICALLY that could be used for versioning :P18:00
rm_worknot that we'd want to, it'd be gross18:00
rm_workbut again, interesting18:00
rm_work"all the files (and only files) it contains are added in lexical order (using LC_COLLATE=C) to the list of configuration files to be loaded ; only files with ".cfg" extension are added, only non hidden files (not prefixed with ".") are added."18:01
*** Alexey_Abashkin has joined #openstack-lbaas18:04
*** AlexeyAbashkin has quit IRC18:07
*** Alexey_Abashkin is now known as AlexeyAbashkin18:07
*** AlexeyAbashkin has quit IRC18:13
rm_workjohnsom: i think the error in the requirements check is NOT mine: http://logs.openstack.org/07/571107/6/check/requirements-check/8b180a1/job-output.txt.gz#_2018-05-31_16_10_55_14973418:14
rm_workunless it's a weird side-effect thing18:14
rm_worki think if we ran any requirements check on master it might fail18:15
johnsomCould be18:15
rm_worki'll see about fixing it18:15
johnsomFallout of the changes in GR management I guess18:15
rm_workyeah it's in our doc reqs18:16
*** kobis has joined #openstack-lbaas18:21
rm_workyep fixed18:25
openstackgerritAdam Harwell proposed openstack/octavia master: Allow DB retries on controller_worker creates  https://review.openstack.org/57110718:26
openstackgerritAdam Harwell proposed openstack/octavia master: Fix requirements gate  https://review.openstack.org/57154718:26
rm_workjohnsom: ^^18:26
johnsomThanks!18:26
rm_worki wonder if our docs requirements are not getting auto-updated, but still checked :/18:26
rm_work(I actually ran the reqs check locally BTW, I'm not just guessing)18:27
rm_worki kinda want to add a tox env for it...18:27
rm_workerg i'm gonna do that i think18:27
numanscgoncalves, thanks for the reply. I figured out what i was looking for :)18:29
*** tesseract has quit IRC18:34
*** JudeC_ has quit IRC18:54
*** JudeC_ has joined #openstack-lbaas18:55
openstackgerritAdam Harwell proposed openstack/octavia master: Fix requirements gate + add local tox profile  https://review.openstack.org/57154719:00
rm_workjohnsom: ^^ try that out? when you have a sec19:00
rm_workmake sure "tox -e requirements" works for you19:00
*** amuller has quit IRC19:11
*** kobis has quit IRC19:19
openstackgerritAdam Harwell proposed openstack/octavia master: Fix requirements gate + add local tox profile  https://review.openstack.org/57154719:23
rm_workadjusted for comments from dhellman19:43
*** nmanos has quit IRC19:52
johnsomSpeaking of this fun, can someone take a look at this: https://review.openstack.org/#/c/571527/19:54
*** salmankhan has joined #openstack-lbaas19:58
*** harlowja has joined #openstack-lbaas20:01
xgerman_yeah, spend yesterday updating my ansible project… but this fix is nicer…20:12
*** kobis has joined #openstack-lbaas20:18
openstackgerritMerged openstack/python-octaviaclient master: update cmd2 dependency to handle py3 only versions  https://review.openstack.org/57152720:20
openstackgerritKobi Samoray proposed openstack/octavia master: Providers: propose changes to API  https://review.openstack.org/57145320:22
*** kobis has quit IRC20:24
*** AlexeyAbashkin has joined #openstack-lbaas20:33
openstackgerritLingxian Kong proposed openstack/octavia master: Use HMAC.hexdigest to avoid non-ascii characters for package data  https://review.openstack.org/57133320:36
johnsomOk, so we might have an issue with drivers on Octavia and stats.20:44
xgerman_mmh20:44
johnsomLBaaS had an LB stats table. Octavia has a listener stats table and rolls them up.20:44
johnsomWhat if we have a driver that only supports providing load balancer stats20:45
johnsom?20:45
rm_workhmmm20:45
xgerman_divide by # of listeners?20:45
rm_worklolol20:45
johnsomDo we force them to only send listener stats (our current model) and the just divide the counts by the number of listeners?20:46
johnsomI lean towards just saying "we accept stats per listener only" and leave it at that.20:46
xgerman_+120:46
xgerman_they can divide it needed20:47
xgerman_if needed20:47
johnsomYeah, that is the way I lean too20:47
*** AlexeyAbashkin has quit IRC20:50
*** salmankhan has quit IRC20:53
*** harlowja has quit IRC21:00
*** samccann has quit IRC21:19
rm_workwe need https://review.openstack.org/#/c/571547/321:20
rm_workthis is a gatefix21:20
rm_workand yes, I agree, let them figure it out21:20
rm_worki honestly don't know if there will be a driver that can't do it that way21:21
rm_workjohnsom / xgerman_ ^^21:22
johnsomAlready done21:22
rm_workthen we can merge the NEXT fix21:22
johnsomWhat is up with the noop API tests?21:22
rm_workthat's what these fixes are all fow21:23
rm_work*for21:23
rm_workthis chain will solve that21:23
rm_workit's the race thing21:23
rm_workso you could go ahead with https://review.openstack.org/#/c/571107/ if you have time :P21:23
rm_workxgerman_: responded21:26
xgerman_better description could have prevented that21:30
xgerman_(my comment)21:31
*** fnaval has quit IRC21:31
rm_workyes, true21:33
rm_workthen we just need https://review.openstack.org/#/c/571107/21:35
rm_workwere you ok with that fix johnsom?21:35
rm_workI had tweaked it a bit for you21:35
rm_worki still think it's gross21:36
rm_workbut basically we have no other option until we do a full driver rewrite, i guess21:36
*** links has joined #openstack-lbaas21:42
*** Swami_ has quit IRC21:57
rm_workjohnsom: what do you think, can we get the rest of the tests (at least through l7rules) merged this week?22:02
rm_workit means merging that db fix (or something to fix that issue) and three more test patches22:02
rm_workabout to post the scenario for l7policy22:02
rm_workwaiting on final runs22:03
johnsomI think it is possible. Are you happy with the driver patches? HM, L7s, and cleanup are still open for reviews22:04
rm_workneed to land these fixes and see tests passing, otherwise i just need to quickly review them after this22:11
rm_workit's all basically stamping a template isn't it?22:11
rm_workkinda like what i'm doing22:12
rm_workthe real review was in the first couple22:12
rm_workand then just tedium...22:12
*** rcernin has joined #openstack-lbaas22:13
rm_worki have noticed one kinda annoying thing, which is that we update statuses of objects in different sessions22:15
rm_workso the flows will make like... a pool active, THEN it'll make the listener and LB active22:15
rm_workso if you have waiters on the subobject, you can race the real object and get 409s22:15
rm_worki realize we can say "you should be checking the LB status" but it's admittedly a little bit lame22:16
*** links has quit IRC22:16
*** harlowja has joined #openstack-lbaas22:21
johnsomhmm, yeah, it's true that we do those separately. They probably should be in one transaction. It would be just changing the tasks to do both in the object unlock and then change the flows to not have separate tasks for that.22:34
johnsomI would guess 99+% of the flows do those two right after each other22:35
rm_workyes]22:35
rm_workthat is what happens now22:35
rm_workbut i have ACTUALLY run into races22:36
rm_worklulz22:36
rm_work(I was doing it wrong in the tests, assuming I could just wait_for_status on the actual object I was checking)22:36
rm_workanywho... working on the last of the l7rule test now (adding an actual traffic test for everything l7)22:36
rm_workthis one is complex :P22:36
johnsomYeah, I'm noodling a bulk upsert for stats22:37
johnsomSo clean to do it direct to the DB, but sqlalchemy...22:37
openstackgerritLingxian Kong proposed openstack/octavia master: Use HMAC.hexdigest to avoid non-ascii characters for package data  https://review.openstack.org/57133322:47
openstackgerritMerged openstack/octavia master: Fix requirements gate + add local tox profile  https://review.openstack.org/57154723:03
*** hyang has joined #openstack-lbaas23:23
*** SumitNaiksatam has quit IRC23:26
hyanghi there, I have a question regarding the network config for Octavia. I'm using queens version and the official doc says create a 'lb-mgmt-net' besides tenant networks. In my cluster deployment, there is only one provider network that is shared by all tenants. I'm wondering can I just let Octavia also use that provider network? If I do so, does it mean I also don't need create an interface for Octavia? (like o-hm0 in devstack)23:28
rm_workhyang: yep, that works23:32
rm_workI call it "management on vip net" and I do it, as well as some others -- it will continue to work23:32
rm_workI actually don't really think it's a problem, as long as that provider net isn't *internet accessible*23:33
hyangok cool, I'll give a try23:34
hyangthank you @rm_work23:34
rm_worknp23:36
johnsomugh, amphora_id is NOT NULL and a foreign key on listener stats.23:36
rm_workorly23:36
johnsomI kind of wish I could just start stats over.23:37
rm_worki guess stats wasn't set up to be split out the way health was >_>23:37
johnsomIt was slapped together (by me) as a place holder and got away from me23:37
johnsomWell, it was, listener ID is not a foreign key....23:37
johnsomWhat the....23:40
johnsom4a6ec0ab7284_remove_fk_constraints_on_listener_.py23:40
johnsom9bf4d21caaea_adding_amphora_id_to_listener_.py23:40
johnsomAh, ok, it's not a true foreign ke23:43
johnsomkey23:43
openstackgerritAdam Harwell proposed openstack/octavia master: Add usage admin resource  https://review.openstack.org/55754823:52
openstackgerritAdam Harwell proposed openstack/octavia master: Fix amp failover where failover already failed  https://review.openstack.org/54898923:52
openstackgerritAdam Harwell proposed openstack/octavia master: Experimental multi-az support  https://review.openstack.org/55896223:52
openstackgerritAdam Harwell proposed openstack/octavia master: WIP: AZ Evacuation resource  https://review.openstack.org/55987323:54
openstackgerritAdam Harwell proposed openstack/octavia master: WIP: Floating IP Network Driver (spans L3s)  https://review.openstack.org/43561223:54
rm_workweekly rebasing >_>23:54
« Wednesday, 2018-05-30 Index Friday, 2018-06-01 »

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!