Monday, 2018-09-24

« Sunday, 2018-09-23 Index Tuesday, 2018-09-25 »
*** rcernin has quit IRC01:40
*** rcernin has joined #openstack-lbaas01:40
*** hongbin has joined #openstack-lbaas01:49
*** kiennt26 has joined #openstack-lbaas02:01
*** yamamoto has quit IRC02:42
*** kiennt26 has quit IRC02:43
*** sapd1 has joined #openstack-lbaas03:05
sapd1Is OVH running Octavia?03:32
*** hongbin has quit IRC03:39
*** yamamoto has joined #openstack-lbaas04:08
*** jiteka has quit IRC04:52
*** ramishra has joined #openstack-lbaas04:59
*** abaindur has quit IRC05:29
*** yamamoto has quit IRC05:30
*** yamamoto has joined #openstack-lbaas05:39
*** dayou has quit IRC05:43
*** dayou has joined #openstack-lbaas05:44
*** pcaruana has joined #openstack-lbaas06:05
openstackgerritReedip proposed openstack/octavia-tempest-plugin master: Add configuration support for skipping tests  https://review.openstack.org/59939306:31
*** reedipb has joined #openstack-lbaas06:53
reedipbxgerman_, johnsom : https://storyboard.openstack.org/#!/story/2003833 is something to look at, probably. Seems to be a timing issue06:54
*** maciejjozefczyk has joined #openstack-lbaas07:00
*** rcernin has quit IRC07:06
*** velizarx has joined #openstack-lbaas07:21
*** velizarx has quit IRC07:31
*** velizarx has joined #openstack-lbaas07:36
maciejjozefczykrm_work: Hey! I commented the line that causes my issues here: https://review.openstack.org/#/c/435612/148 with my explination. Please take a look :) thanks!07:46
*** velizarx has quit IRC07:51
*** velizarx has joined #openstack-lbaas07:54
*** velizarx has quit IRC08:20
*** openstackgerrit has quit IRC08:22
*** yamamoto has quit IRC08:45
*** velizarx has joined #openstack-lbaas09:12
*** salmankhan has joined #openstack-lbaas09:15
*** rpittau has joined #openstack-lbaas09:20
*** ianychoi_ has joined #openstack-lbaas09:30
*** ianychoi has quit IRC09:34
*** yamamoto has joined #openstack-lbaas09:40
*** sapd1__ has quit IRC09:54
*** sapd1_ has joined #openstack-lbaas09:59
*** Emine has joined #openstack-lbaas10:01
*** sapd1 has quit IRC10:55
*** jitek4 has joined #openstack-lbaas11:12
jitek4I noticed that logs for this chat are not archived here http://eavesdrop.openstack.org/irclogs/11:15
jitek4is it by choice or it will in future ?11:15
*** velizarx has quit IRC11:48
*** velizarx has joined #openstack-lbaas11:52
*** velizarx has quit IRC12:35
*** velizarx has joined #openstack-lbaas12:37
*** openstackgerrit has joined #openstack-lbaas13:00
openstackgerritmelissaml proposed openstack/neutron-lbaas-dashboard master: Update the URL in HACKING.rst  https://review.openstack.org/60476913:00
*** yamamoto has quit IRC13:07
*** yamamoto has joined #openstack-lbaas13:08
*** reedipb has quit IRC13:32
*** yamamoto has quit IRC13:54
*** yamamoto has joined #openstack-lbaas13:55
*** yamamoto has quit IRC13:55
*** yamamoto has joined #openstack-lbaas13:55
*** yamamoto has quit IRC13:55
*** ramishra has quit IRC14:14
*** velizarx has quit IRC14:15
*** velizarx has joined #openstack-lbaas14:20
*** dulek has quit IRC14:38
*** dulek has joined #openstack-lbaas14:39
xgerman_jitek4: we want them archived - but not sure how to make that happen if it’s not currently working15:02
johnsomjitek4 I see them being logged here: http://eavesdrop.openstack.org/irclogs/%23openstack-lbaas/15:12
*** yamamoto has joined #openstack-lbaas15:13
*** velizarx has quit IRC15:20
*** sapd1 has joined #openstack-lbaas15:53
*** velizarx has joined #openstack-lbaas16:37
openstackgerritmelissaml proposed openstack/neutron-lbaas-dashboard master: Update the URL in HACKING.rst  https://review.openstack.org/60476916:38
*** velizarx has quit IRC16:41
*** salmankhan has quit IRC17:37
johnsomSo frustrated, just wrote up a big RFE for backend re-encryption and the "save changes" button was grayed out. Ended up losing all of my work....17:46
johnsomStoryboard.....  arg17:46
colin-omg17:47
colin-i hate that. something, somewhere should suffer for this injustice17:47
colin-i just don't know what, or where17:47
johnsomlol17:47
johnsomYeah, I had about 15 tasks defined for it too. Lesson here, do one-by-one as storyboard is unstable17:48
colin-vim would've at least saved your swap file :p17:48
johnsomHa, yes17:49
rm_workI usually at least copy/paste all of my TEXT out into something else when that happens18:04
rm_workannoying but not catastrophic18:04
openstackgerritGerman Eichberger proposed openstack/octavia master: Refactor the AAP driver to not depend on nova  https://review.openstack.org/60422618:04
johnsomYeah, I started to and hit a bad key, that closed the window18:04
johnsomrm_work and others, see if I am forgetting anything here: https://storyboard.openstack.org/#!/story/200385818:04
johnsomI will add another one for backend client certs18:04
johnsomI also think per-member CAs is a "future" think, not part of this RFE18:05
rm_workmaciejjozefczyk: so, we can't add params or anything that would be API-changing18:05
rm_workwe have to re-use the existing stuff18:05
xgerman_might want to set it up in Word first since they have a spell checker, etc.18:05
rm_workwhich is somewhat limiting18:05
rm_workmaciejjozefczyk: also -- neutron doesn't let you specify a subnet on a vip create, does it? ours doesn't i think18:06
rm_workor maybe we just don't18:06
johnsomYes, we allow subnet specification18:06
rm_workerr18:06
rm_worksorry18:07
rm_worki meant, neutron doesn't let you specify a subnet on FLIP create18:07
johnsomYes, it does18:07
rm_workhmmm18:07
johnsomhttps://developer.openstack.org/api-ref/network/v2/index.html#id13318:08
rm_workah yeah it does, ok18:08
rm_workhmmm18:08
rm_workdo people use that?18:08
johnsomWell, you can have many subnets on one network18:09
rm_workright18:09
johnsomI see IPv4 and IPv6 subnets on a given network pretty regularly18:09
rm_workand we do have like 20 subnets on our FLIP network18:09
rm_workbut how would a customer be expected to choose one?18:09
rm_workah are FLIPs normally on user-networks too?18:10
johnsomThey should known the subnet they want their address on18:10
johnsomOr just don't specify and roll the dice18:10
rm_workright, normally you let neutron pick the subnet18:10
rm_workbased on what still has IPs available18:11
rm_worki think i have been in this environment too long18:11
rm_worki think this is a GD thing18:11
johnsomAgain, probably the most common case is picking IPv4 or IPv618:11
rm_workwe have "public floating ips" in a network, and as the IPs dry up, we add new subnets18:11
rm_workso we have like 20 subnets18:12
rm_workand most of them have no IPs left18:12
rm_workand a user would have no insight into which one to use18:12
rm_workthey could guess and hopefully not just get an error back :P18:12
rm_workbut yeah ok, my FLIP experience is very unlike the normal upstream FLIP experience I think18:13
*** coreycb has joined #openstack-lbaas18:13
rm_workmaciejjozefczyk: i think we need to schedule some time when we're both online, to have a chat :P18:18
openstackgerritMichael Johnson proposed openstack/octavia master: Remove duplicate check gate jobs  https://review.openstack.org/60486918:24
*** fnaval has joined #openstack-lbaas18:33
*** ianychoi_ is now known as ianychoi18:54
*** aojea has joined #openstack-lbaas19:05
*** sapd1 has quit IRC19:15
*** pcaruana has quit IRC19:17
*** aojea has quit IRC19:23
*** aojea has joined #openstack-lbaas19:24
johnsomFYI: https://review.openstack.org/60489019:36
johnsomStarted the process to create octavia-lib19:36
*** pck has quit IRC19:45
*** pck has joined #openstack-lbaas19:45
colin-qq if anyone cares to weigh in, since upgrading octavia to rocky i've been encountering this error when attemtping to use magnum, which employs a heat template that creates octavia resources:19:56
colin-"Resource CREATE failed: OctaviaClientException: resources.api_lb.resources.monitor: http_method is not a valid option for health monitors of type TCP (HTTP 400)"19:56
colin-my impression up until now was that my call does not include an http_method19:57
colin-additionally the properties of that resource have a "type" (of monitor) defined as "TCP"20:01
colin-realize this channel is not the heat stack or magnum cluster channel just wondering if anyone has suggestions20:01
rm_workeither we have some sort of bug with our defaults around the HM args (which i doubt because i think we test it pretty well, but possible) or their templates have some way for the type it thinks it's editing to become mismatched20:05
colin-https://docs.openstack.org/heat/latest/template_guide/openstack.html#OS::Octavia::HealthMonitor20:06
colin-seems to be conforming to tha so, yeah not sure20:06
rm_workthat doesn20:08
rm_work*doesn't really mean anything20:08
rm_workneed to see the template20:09
rm_workhttp_method is a valid arg... in the case that the type is HTTP, and yeah those docs confirm they're aware of that (though i assume they're just copy/pasted)20:09
colin-i get that20:10
rm_workwhat template is it that you're using/20:10
rm_worksomething in magnum....20:10
rm_workahh are they still using N-LBaaS?20:11
rm_workhttps://github.com/openstack/magnum/blob/master/magnum/drivers/common/templates/environments/with_master_lb_octavia.yaml20:12
rm_worklooks like20:13
rm_workor... does that mean they are like ... aliasing in the octavia stuff over the n-lbaas norm?20:13
johnsomcolin- Yes, there was a bug that was fixed in Rocky where the API was filling in HTTP fields for TCP health monitors. I assume that is where this change originated.  There is now better validation that HTTP settings are being attempted to be used on TCP health monitors20:15
rm_worklooking through heat and magnum both and i can't actually find where the template would be defined20:16
rm_worksomething has to pass a template yaml into this engine right?20:16
colin-https://github.com/openstack/magnum/blob/stable/rocky/magnum/drivers/common/templates/lb.yaml#L40-L4720:17
colin-+20:17
colin-https://github.com/openstack/magnum/blob/stable/rocky/magnum/drivers/common/templates/environments/with_master_lb_octavia.yaml20:17
colin-i think?20:17
rm_worklooks right, ok20:17
johnsomcolin- that looks ok to me....20:17
colin-same20:17
rm_workand we can't CHANGE a type of a HM right?20:18
rm_workit's immutable?20:18
colin-"Can be updated without replacement."20:18
colin-under the optional properties http_method def https://docs.openstack.org/heat/latest/template_guide/openstack.html#OS::Octavia::HealthMonitor20:18
colin-unless i misunderstood your Q20:18
johnsomType cannot be changed on a health monitor per our API spec: https://developer.openstack.org/api-ref/load-balancer/v2/index.html#update-a-health-monitor20:19
colin-ah20:19
rm_workyeah in our API-Ref, health-monitor type cannot be updated20:19
rm_workjust verified20:19
rm_workis it in the CREATE or is it an update that fails?20:19
colin-create20:20
johnsomIt looks like heat did something different, they have "Updates cause replacement." for type, so...20:20
rm_workdoes heat dump out the call into logs?20:20
rm_workyeah i assume that's a heat thing that means "we delete and re-create elements that have immutable properties changed"20:20
colin-let me see what heat sees, good point. easy to forget there's an intermediary or two20:21
rm_workand does octavia api dump the whole call? i forget... might be up to the WSGI runner?20:21
rm_worki think mine does20:21
*** salmankhan has joined #openstack-lbaas20:22
rm_work(logs incoming requests by just dumping the whole url/body)20:22
rm_workbasically i want to know exactly what heat is sending, to know if it's them constructing the params wrong somehow, or us mucking with it20:22
colin-you guys are going to ban me20:24
rm_worklol20:24
colin-octavia isn't failing to create the LBs20:24
colin-you are just dealign wiith an operator who believed heat rather than checking with ocatvia first20:24
colin-...sorry!20:24
colin-mortified20:24
rm_worki mean, the LB create would be first20:25
rm_workfollowed by the listener/pool20:25
rm_workthen the HM20:25
colin-well what it's sending is apparently good enough for octavia cause the LB is coming up fine20:25
rm_workthe HM was where the error was20:25
colin-it's just not reporting that way20:25
rm_workso it's probably up, but with no HM?20:25
colin-good question let me see what its HM looks like20:25
rm_worki assume the error is not completely invalid20:26
rm_workwe'd be looking for the call Heat made for the HM create, which is like step 3-5 somewhere probably20:27
colin-you called it, no HM20:27
rm_worksince i assume they don't do single-creates20:27
colin-yeah i think you're right20:28
rm_workso see if you can get the logs for the HM create call (hopefully including the body) either on the heat side or the octavia side20:28
colin-https://docs.openstack.org/heat/queens/template_guide/openstack.html#OS::Octavia::HealthMonitor-prop-http_method20:36
colin-queens heat, heh20:36
colin-Defaults to “GET”.20:36
rm_workah yep20:38
colin-fiddlesticks20:39
rm_workcaw caw caw20:39
rm_worksorry, game reference20:39
johnsomYou can probably update your TCP heat template to set that value to None or null to work around the queens default20:43
colin-i suggested that and got laughed at here!20:43
colin-i'm doing it, ptl said so20:43
johnsomYeah, well, I'm not the PTL of heat so....20:43
colin-haha fiar20:43
colin-fair*20:43
colin-i'll try that, still finding the request body from heat->octa. appreciate the feedback thanks to you both20:44
colin-https://github.com/openstack/heat/blob/stable/queens/heat/engine/resources/openstack/octavia/health_monitor.py#L8020:58
colin-https://github.com/openstack/heat/blob/stable/queens/heat/engine/resources/openstack/octavia/health_monitor.py#L39-L4120:58
johnsomDoh20:59
colin-same in rocky21:00
openstackgerritMichael Johnson proposed openstack/octavia master: Add API versioning to the api-ref  https://review.openstack.org/60491121:04
rm_workdoes None/null not bypass that? :(21:09
rm_workwhat even is this syntax, i have never seen it21:10
rm_workhttps://github.com/openstack/heat/blob/stable/queens/heat/engine/resources/openstack/octavia/health_monitor.py#L39-L4521:10
johnsomYeah, new one on me too21:11
rm_worktesting it locally, python doesn't explode like i'd expect it to, but it also doesn't seem to do anything special like i'd expect it to :/21:12
rm_workresult seems to just be the second tuple21:12
rm_workOH21:12
rm_workI get it21:12
rm_workthat's actually kinda cool21:12
rm_workgives you a single var with the tuple, and also DEFINES the rest21:13
rm_workso it's a shorthand way of doing some of the shit we do in our constants21:13
rm_workwhere we do like, 10 lines to define a bunch of vars, then make a list of them21:13
rm_workwhether it's more readable or not, i question21:14
johnsomAh, my next task, working on the gates is going to be fun with the backlog21:23
colin-going to test a local change to heat adding a none/emptystr to that list and see how it goes21:26
*** salmankhan has quit IRC21:39
*** abaindur has joined #openstack-lbaas21:47
*** aojea has quit IRC21:47
*** abaindur has quit IRC21:48
*** abaindur has joined #openstack-lbaas21:48
openstackgerritMichael Johnson proposed openstack/octavia master: Remove deprecated API settings  https://review.openstack.org/60081921:58
*** Emine has quit IRC22:03
openstackgerritAdam Harwell proposed openstack/octavia master: HTTPS HMs need the same validation path as HTTP  https://review.openstack.org/60492422:14
rm_work^^22:18
openstackgerritMichael Johnson proposed openstack/octavia master: Add amphora statistics to the admin API  https://review.openstack.org/58503122:31
*** fnaval has quit IRC22:38
*** rcernin has joined #openstack-lbaas22:45
*** dims_ has joined #openstack-lbaas22:59
*** dims has quit IRC23:00
*** ArchiFleKs has quit IRC23:00
*** nmagnezi has quit IRC23:00
*** strigazi has quit IRC23:00
*** andreykurilin has quit IRC23:03
*** andreykurilin has joined #openstack-lbaas23:05
*** ArchiFleKs has joined #openstack-lbaas23:06
openstackgerritMichael Johnson proposed openstack/octavia master: Pass through DIB_LOCAL_ELEMENTS from localrc  https://review.openstack.org/60493323:35
openstackgerritMichael Johnson proposed openstack/octavia-tempest-plugin master: Use the infra mirrors for DIB  https://review.openstack.org/60133223:36
johnsomcgoncalves nmagenzi You may want to jump into this thread about storyboard priorities: http://lists.openstack.org/pipermail/openstack-dev/2018-September/135030.html23:57
« Sunday, 2018-09-23 Index Tuesday, 2018-09-25 »

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