Thursday, 2018-01-18

« Wednesday, 2018-01-17 Index Friday, 2018-01-19 »
*** yamamoto has joined #openstack-lbaas00:06
*** jniesz has quit IRC00:25
*** threestrands_ has quit IRC00:48
*** threestrands has joined #openstack-lbaas00:48
openstackgerritMichael Johnson proposed openstack/octavia master: ACTIVE-ACTIVE: OVS diskimage-builder element  https://review.openstack.org/53384901:14
*** armax has quit IRC01:18
dayoujohnsom, I'll look into it and report status back later today01:21
johnsomOk, thanks, there is some dependency missing after the zuulv3 change that I'm not sure where it goes01:21
dayouyes, seems so01:22
*** openstackstatus has quit IRC01:55
*** harlowja has quit IRC01:55
*** johnsom has quit IRC01:55
*** robcresswell has quit IRC01:55
*** jmccrory has quit IRC01:55
*** mnaser has quit IRC01:55
*** reedip has quit IRC01:55
*** jmccrory has joined #openstack-lbaas01:55
*** johnsom has joined #openstack-lbaas01:56
*** mnaser has joined #openstack-lbaas01:56
*** robcresswell has joined #openstack-lbaas01:56
*** openstackstatus has joined #openstack-lbaas01:56
*** ChanServ sets mode: +v openstackstatus01:56
*** reedip has joined #openstack-lbaas02:06
openstackgerritHengqing Hu proposed openstack/octavia master: Fix a few E123, E126 pep8 errors  https://review.openstack.org/53499702:06
*** annp has joined #openstack-lbaas02:19
*** AlexeyAbashkin has joined #openstack-lbaas02:46
*** AlexeyAbashkin has quit IRC02:50
*** armax has joined #openstack-lbaas03:07
*** sanfern has quit IRC03:21
*** rm_work has quit IRC03:30
*** coreycb has quit IRC03:30
*** Aju has quit IRC03:30
*** afranc has joined #openstack-lbaas03:30
*** rm_work has joined #openstack-lbaas03:30
*** rm_work has quit IRC03:30
*** rm_work has joined #openstack-lbaas03:30
*** coreycb has joined #openstack-lbaas03:31
*** armax has quit IRC03:36
*** armax has joined #openstack-lbaas03:37
*** AlexeyAbashkin has joined #openstack-lbaas03:46
*** AlexeyAbashkin has quit IRC03:51
*** armax has quit IRC04:10
*** fnaval has joined #openstack-lbaas04:23
*** sanfern has joined #openstack-lbaas04:25
*** gans has joined #openstack-lbaas04:39
openstackgerritGanpat Agarwal proposed openstack/octavia master: ACTIVE-ACTIVE ExaBGP rest api driver  https://review.openstack.org/52700904:47
openstackgerritGanpat Agarwal proposed openstack/octavia master: ACTIVE-ACTIVE ExaBGP rest api driver  https://review.openstack.org/52700904:59
*** gans has quit IRC04:59
*** gans has joined #openstack-lbaas05:11
*** gans has quit IRC05:28
*** links has joined #openstack-lbaas05:31
*** fnaval has quit IRC05:36
*** sticker has quit IRC05:39
*** gcheresh has joined #openstack-lbaas06:07
*** gans has joined #openstack-lbaas06:46
*** dims has quit IRC07:12
*** dims has joined #openstack-lbaas07:14
*** pcaruana has joined #openstack-lbaas07:15
*** dims has quit IRC07:19
*** dims has joined #openstack-lbaas07:20
*** aojea has joined #openstack-lbaas07:38
dayoujohnsom, a few changes introduced in pike were not back ported to ocata, which was causing that fail, I cherry picked them and rebased that patch, now it works.07:42
*** gans has quit IRC07:51
*** annp has quit IRC07:56
*** Alex_Staf has joined #openstack-lbaas07:58
*** threestrands has quit IRC08:03
*** tesseract has joined #openstack-lbaas08:08
*** b_bezak has joined #openstack-lbaas08:12
*** annp has joined #openstack-lbaas08:13
*** AlexeyAbashkin has joined #openstack-lbaas08:26
*** rcernin has quit IRC08:28
openstackgerritMerged openstack/octavia-dashboard master: Update the installation and contributors documents  https://review.openstack.org/53376008:32
*** pcaruana has quit IRC08:43
*** pcaruana has joined #openstack-lbaas08:43
openstackgerritHengqing Hu proposed openstack/octavia master: Make new-octavia-devstack.sh os agnostic  https://review.openstack.org/53523108:55
*** dulek has quit IRC09:21
*** dulek has joined #openstack-lbaas09:24
openstackgerritGanpat Agarwal proposed openstack/octavia master: Updated the function name with "get_remove_vip_subflow" to fix the typo  https://review.openstack.org/53526509:59
*** gans has joined #openstack-lbaas09:59
*** salmankhan has joined #openstack-lbaas10:06
*** tesseract has quit IRC10:27
*** tesseract has joined #openstack-lbaas10:29
*** salmankhan has quit IRC10:46
*** sanfern has quit IRC10:50
*** salmankhan has joined #openstack-lbaas10:57
*** gans has quit IRC11:18
*** fnaval has joined #openstack-lbaas11:18
*** fnaval has quit IRC11:22
*** salmankhan has quit IRC11:40
*** salmankhan has joined #openstack-lbaas11:43
*** yamamoto has quit IRC11:51
*** yamamoto has joined #openstack-lbaas11:55
*** annp has quit IRC12:00
*** salmankhan has quit IRC12:11
*** fnaval has joined #openstack-lbaas12:17
*** fnaval has quit IRC12:22
*** salmankhan has joined #openstack-lbaas12:23
*** Alex_Staf has quit IRC12:24
*** aojea has quit IRC12:26
*** sanfern has joined #openstack-lbaas12:37
*** Alex_Staf has joined #openstack-lbaas12:40
*** salmankhan has quit IRC12:50
*** salmankhan has joined #openstack-lbaas12:52
*** yamamoto has quit IRC12:58
*** yamamoto has joined #openstack-lbaas13:01
*** salmankhan has quit IRC13:05
*** ianychoi has quit IRC13:12
*** yamamoto has quit IRC13:16
*** aojea has joined #openstack-lbaas13:27
*** aojea has quit IRC13:32
*** yamamoto has joined #openstack-lbaas13:35
*** Alex_Staf has quit IRC13:51
*** Alex_Staf has joined #openstack-lbaas13:52
*** ianychoi has joined #openstack-lbaas13:55
*** gans has joined #openstack-lbaas13:55
*** salmankhan has joined #openstack-lbaas14:04
*** atoth has joined #openstack-lbaas14:13
*** gans has quit IRC14:22
*** HW_Peter has quit IRC14:29
*** HW-Peter has joined #openstack-lbaas14:29
*** aojea has joined #openstack-lbaas14:29
*** dayou has quit IRC14:30
*** fnaval has joined #openstack-lbaas14:32
*** aojea has quit IRC14:35
*** dayou has joined #openstack-lbaas14:44
*** Alex_Staf has quit IRC14:51
*** Alex_Staf has joined #openstack-lbaas14:54
*** salmankhan has quit IRC15:10
*** ndanl4 has joined #openstack-lbaas15:10
*** gcheresh has quit IRC15:12
*** aojea has joined #openstack-lbaas15:15
*** salmankhan has joined #openstack-lbaas15:17
*** armax has joined #openstack-lbaas15:22
*** ndanl4 is now known as ndanl15:22
*** b_bezak has quit IRC15:27
*** b_bezak has joined #openstack-lbaas15:27
*** b_bezak has quit IRC15:32
cgoncalvesjohnsom: hey. neutron-lbaas bugs should be filled against neutron@launchpad or octavia@storyboard?15:34
johnsomThere is a neutron-lbaas project in storyboard15:35
cgoncalvesoh! :)15:36
cgoncalvesAlex_Staf: ^ :D15:36
*** links has quit IRC15:38
Alex_Stafcgoncalves, tnx man15:40
*** bbzhao has quit IRC15:47
*** bbzhao has joined #openstack-lbaas15:48
*** jerrygb has joined #openstack-lbaas15:50
*** jerrygb has left #openstack-lbaas15:51
*** yamamoto has quit IRC16:09
*** AlexeyAbashkin has quit IRC16:13
*** yamamoto has joined #openstack-lbaas16:14
*** bbzhao has quit IRC16:26
*** yamamoto has quit IRC16:27
*** bbzhao has joined #openstack-lbaas16:28
*** pcaruana has quit IRC16:29
*** atoth has quit IRC16:31
*** yamamoto has joined #openstack-lbaas16:39
*** yamamoto has quit IRC16:44
*** yamamoto has joined #openstack-lbaas16:49
*** yamamoto has quit IRC16:49
*** salmankhan has quit IRC16:54
johnsomAlex_Staf If you are still online, I commented on that LBaaS story and marked invalid. Happy to chat about it here if you want.16:56
Alex_Stafjohnsom, the members stuff ?16:57
johnsomYes16:57
Alex_Stafjohnsom, yep glad to hear :)16:57
Alex_Stafjohnsom, yeah I agree than neutron lbaas is less interesting . BUT, I think we should see if it is repeated in openstack loadbalancer member show/list16:59
johnsomThe status is reported on members in the OSC client.17:00
Alex_Stafjohnsom, awesome17:03
Alex_Stafcgoncalves, ^^17:03
johnsomYeah, it was a peeve of mine on LBaaSv2, so I made sure we fixed it with the new API.17:04
*** salmankhan has joined #openstack-lbaas17:04
johnsomIt's just crazy to not have the status on the objects IMO17:04
Alex_Stafjohnsom, totally agreed .17:04
*** bbzhao has quit IRC17:16
*** bbzhao has joined #openstack-lbaas17:16
*** Alex_Staf has quit IRC17:18
openstackgerritMichael Johnson proposed openstack/python-octaviaclient master: Update and add listener stats release notes  https://review.openstack.org/53541417:20
*** ivve has quit IRC17:25
openstackgerritMerged openstack/python-octaviaclient master: Updated from global requirements  https://review.openstack.org/53512217:27
*** aojea has quit IRC17:29
*** atoth has joined #openstack-lbaas17:29
*** aojea has joined #openstack-lbaas17:35
*** ivve has joined #openstack-lbaas17:37
ivveits not possible to assign a security group to a loadbalancer (lbaas v2) ?17:45
*** Swami has joined #openstack-lbaas17:46
*** SumitNaiksatam has joined #openstack-lbaas17:47
*** yamamoto has joined #openstack-lbaas17:50
*** KeithMnemonic has joined #openstack-lbaas17:55
openstackgerritMerged openstack/python-octaviaclient master: Update and add listener stats release notes  https://review.openstack.org/53541417:57
*** AlexeyAbashkin has joined #openstack-lbaas17:58
*** tesseract has quit IRC18:00
*** yamamoto has quit IRC18:00
*** bbzhao has quit IRC18:01
*** bbzhao has joined #openstack-lbaas18:02
*** AlexeyAbashkin has quit IRC18:03
*** AlexeyAbashkin has joined #openstack-lbaas18:14
*** bbzhao has quit IRC18:14
*** bbzhao has joined #openstack-lbaas18:15
*** AlexeyAbashkin has quit IRC18:19
*** AlexeyAbashkin has joined #openstack-lbaas18:22
sanfernhi johnsom18:24
johnsomsanfern Hi18:25
sanferni feel we need to normalized the tables of L3 AA18:25
johnsomivve You can, but it is not recommended as the LB manages that SG applied to the ports18:25
sanfernnow i am assuming one peer only support18:26
johnsomI thought we did that here: https://review.openstack.org/#/c/528850/18:26
*** AlexeyAbashkin has quit IRC18:26
ivveok, just getting no route to host now.. no clue why. all routes look fine to me18:26
ivvevia the FIP18:26
ivveotherwise stuff works18:26
johnsomOh, are you discussing the concern of multiple vips per amphora?18:27
sanfernyes18:27
johnsomivve If you are using DVR it might be a DVR related bug18:27
sanfernmultiple tor switch18:27
ivveoh a bug?18:27
sanfernone for v6 and one for v418:27
ivvei am using dvr18:27
ivveso its known18:27
johnsomsanfern, if we can leave the design open to support it in the future cool.  If it's really causing complications we can consider changing it.18:27
ivvei guess i could turn off ha on the switch to test18:28
johnsomivve DVR has had a number of issues with FLIPs.18:28
ivvejohnsom: checkbox!18:28
ivvebut regarding sec groups18:28
ivvethe port is getting the default sec group18:28
ivve(i am using heat)18:28
*** aojea has quit IRC18:28
johnsomivve You may need to rebuild the flip to get it resolved after disabling DVR18:29
sanfernjohnsom, i am considering one peer now but future we can drop L3_BGP_peer_reg table18:29
ivveshould i not try to assign a secgroup to the vip?18:29
johnsomivve each VIP should get a unique security group....  If it's an AAP port, check the base port SG18:30
ivveaye ill rebuild the hole stack18:30
ivvejohnsom: cool!18:30
johnsomsanfern Ah, I see.  So, you are proposing having a speaker only talk to one peer (tor)?18:31
sanfernyes for this release18:32
johnsomsanfern I thought Janson wanted to have a centralized speaker that did all of the advertisements in the future.18:32
johnsomI am all for simplifying and incremental improvement, we just need to make sure we have an upgrade strategy/plan of how we can get to where Jason want's to go with it.18:33
sanfernjohnsom, we will have one peer i.e. TOR switch, speaker will be running in all the amps18:33
*** gcheresh has joined #openstack-lbaas18:34
sanfernDATA_MODEL needs to be normalized we will be good, jason requirements changes every meeting ;)18:34
johnsomHa, ok18:35
johnsomWell, no harm in putting up a proposed patch18:35
sanfernok18:35
*** salmankhan has quit IRC18:35
johnsomMaybe add a comment that describes "why" to help reviewers18:35
*** aojea has joined #openstack-lbaas18:35
*** rstarmer has joined #openstack-lbaas18:36
sanfernjohnsom, I agree, Will add that.18:40
*** aojea has quit IRC18:41
ivvejohnsom: didn't do it, router was created with "ha: false" but no cigar. no route to host18:53
johnsomHmm, ok, that is odd then18:54
sanfernjohnsom, in delete_distributor task i am just maring distributor prov_status to deleted. I am not removing speaker or peer entries18:54
ivvethere is "distributed: false" as well, could try that18:54
johnsomsanfern that is good.  We should clean those up with housekeeping per the deleted retention policy in the config18:55
sanfernok we need to write housekeeping flow to delete tables18:55
johnsomivve Ah, yes, they are different. DVR is distributed, HA is just HA routers with keepalived18:55
johnsomsanfern there is an existing flow, it probably just needs extended18:56
sanfernok18:56
ivvetesting...18:56
ivvethats wierd18:56
ivveCreate Failed18:57
ivveResource Create Failed: Badrequest: Resources.Router: Unrecognized Attribute(S)18:57
ivve'Distributed' Neutron Server Returns Reques18:57
ivve:)18:57
*** jniesz has joined #openstack-lbaas18:57
ivvesomeone really doesn't want me doing this lbaas thing18:57
*** atoth has quit IRC18:58
johnsomWell, that doesn't have anything to do with lbaas....18:58
johnsomWe are just victims of other project bugs.... grin18:58
ivvetrue dat19:01
*** atoth has joined #openstack-lbaas19:04
*** gcheresh has quit IRC19:11
ivvejohnsom: well i was under the impression i was using dvr... but neutron.conf in server says false on both settings... l3_ha says true though19:13
ivvehaha. any thoughts? :D19:13
johnsomUmm, I don't know why it would say L3 HA but not be enabled in the config.  That is a mystery. But, if distributed is disabled when you do show router for the VIP router, it's not DVR related....19:14
johnsomOn devstack, it doesn't setup route to the private network from the host anymore, but I assume that is not your setup....19:15
sanfernjohnsom, http://paste.openstack.org/show/P7smyrv1T9FV3IH4IMv1/ can we have dynamic j2 template, like neig19:16
sanfernmore neighbors entry means more peer_ip's19:16
jnieszthat would require another python process running19:17
jnieszwould prefer not to use sub process19:17
*** atoth has quit IRC19:17
openstackgerritOpenStack Release Bot proposed openstack/python-octaviaclient master: Update reno for stable/queens  https://review.openstack.org/53545619:18
johnsomjniesz Not sure I follow.19:19
ivvejohnsom: nah its a 7 node setup with 3 controllers where all 7 are compute as well as ceph nodes19:20
jnieszprocess flask-api {19:20
jniesz                run /path/to/python /path/to/app.py;19:20
jniesz        }19:20
jnieszthat is basically running a separate python process to import stuff into exabgp19:20
johnsomsanfern Sure, not a problem. This is how we configure keepalived, the neighbor blocks are just in a for loop in jina19:20
sanfernjniesz, ignore run19:20
johnsomjniesz Why is that needed?19:20
jnieszi am saying it is not19:20
johnsomAh, ok19:21
johnsomYeah, I can't think why it would be needed either19:21
jnieszi think sanfern was just showing two neighbors19:21
sanfernI wanted to make multiple entry of neighbor in j219:21
jnieszoption19:21
jnieszthe case we want to handle is multiple neighbors19:21
jnieszsome people might need to peer to more than one19:22
ivvejohnsom: so the lbaas namespace has a default route (to private subnet)19:22
johnsomsanfern https://github.com/openstack/octavia/blob/master/octavia/common/jinja/haproxy/templates/macros.j2#L24819:22
ivvefrom that namespace i can reach both the provider router as well as the private subnet router19:22
ivvesnat is enabled on all routers19:23
jnieszusing for loop seems like right way to handle multiple neighbors19:23
johnsomivve Default route should be to the VIP network.  There also might be policy base routing rules. Not sure on that old driver19:23
ivvenever thought about that... :P19:23
sanfernjohnsom, this is how i did it for multiple vip to dummy0.cfg file19:23
johnsomsanfern +119:24
ivveits the rootwrap that sets it19:24
ivve(the gw)19:24
jnieszjohnson: one thing I was discussing with sanfern is removing the speaker and bgp peer tables19:25
jnieszsince that will now be coming from config file19:25
jnieszi think it only makes sense to keep the tables, if we also build the cli / api's for the administrator to define those19:26
johnsomjniesz I think they should stay as you want failover rebuilds to use the stored values and not the "new" values that may be in the config.19:26
sanfernhttps://review.openstack.org/#/c/524722/33/octavia/db/models.py - DistributorL3BGPPeerRegistration19:26
ivvejohnsom: ah it is.. i read flip19:26
sanfernDo we need this table - DistributorL3BGPPeerRegistration - one to many relationship19:27
ivveso default route on lbaas netns is correct, everything is reachable there all routers. all the instances can reach those routers as well19:27
jnieszthe reason I created that table was to associate peers to one or more bgp speakers19:28
jnieszthat will be more relevant when we have dedicated distributors19:28
johnsomRight19:28
jnieszfrom a modeling perspective you register bgp peers to a bgp speaker19:29
jnieszand bgp speaker has some config info19:29
jnieszand bgp peers = neighbor19:29
johnsomIf we have a upgrade/migration path for that in the future when we do a "dedicated distributor for L3" I am fine with dropping that now. But we should balance if it's just easier to leave it with single records for now vs. dropping it19:29
sanfernwe will have speaker table, peer table19:30
jnieszit makes sense to keep, right now it is 1:119:31
johnsomFYI, I have released python-octaviaclient 1.4.0 which will be our Queens OSC plugin release.19:31
jnieszactually if there are two neighbors it would be 1:2119:31
jniesz1:219:31
*** rstarmer has quit IRC19:32
*** harlowja has joined #openstack-lbaas19:52
*** slaweq has joined #openstack-lbaas20:03
*** salmankhan has joined #openstack-lbaas20:08
*** salmankhan has quit IRC20:12
*** SumitNaiksatam has quit IRC20:17
openstackgerritSanthosh Fernandes proposed openstack/octavia master: [WIP] ACTIVE-ACTIVE create distributor flow  https://review.openstack.org/52778420:19
openstackgerritSanthosh Fernandes proposed openstack/octavia master: [WIP] ACTIVE-ACTIVE create distributor flow  https://review.openstack.org/52778420:20
*** sanfern has quit IRC20:29
*** sanfern has joined #openstack-lbaas20:30
*** sanfern has quit IRC20:30
*** sanfern has joined #openstack-lbaas20:30
*** gcheresh has joined #openstack-lbaas20:31
*** sanfern has quit IRC20:31
*** sanfern has joined #openstack-lbaas20:31
*** sanfern has quit IRC20:32
*** sanfern has joined #openstack-lbaas20:32
*** sanfern has quit IRC20:33
*** gcheresh has quit IRC20:37
*** AlexeyAbashkin has joined #openstack-lbaas20:45
*** Alex_Staf has joined #openstack-lbaas20:45
*** AlexeyAbashkin has quit IRC20:49
*** gcheresh has joined #openstack-lbaas21:19
*** ivveh has quit IRC21:27
*** ivveh has joined #openstack-lbaas21:27
*** slaweq has quit IRC21:27
*** aojea has joined #openstack-lbaas21:30
*** slaweq has joined #openstack-lbaas21:32
*** openstackgerrit has quit IRC21:33
*** slaweq has quit IRC21:34
*** aojea_ has joined #openstack-lbaas21:35
*** rstarmer has joined #openstack-lbaas21:37
*** aojea has quit IRC21:38
*** aojea has joined #openstack-lbaas21:40
*** aojea_ has quit IRC21:43
*** AlexeyAbashkin has joined #openstack-lbaas21:45
*** aojea_ has joined #openstack-lbaas21:45
*** gcheresh has quit IRC21:47
*** aojea has quit IRC21:48
*** AlexeyAbashkin has quit IRC21:50
*** aojea has joined #openstack-lbaas21:51
*** aojea_ has quit IRC21:54
*** aojea_ has joined #openstack-lbaas21:55
*** threestrands has joined #openstack-lbaas21:56
*** threestrands has quit IRC21:56
*** threestrands has joined #openstack-lbaas21:56
*** aojea has quit IRC21:58
*** aojea has joined #openstack-lbaas22:00
*** aojea_ has quit IRC22:03
*** aojea_ has joined #openstack-lbaas22:05
*** rcernin has joined #openstack-lbaas22:06
*** aojea has quit IRC22:08
*** aojea has joined #openstack-lbaas22:11
*** aojea_ has quit IRC22:14
*** aojea_ has joined #openstack-lbaas22:16
*** aojea has quit IRC22:19
*** aojea_ has quit IRC22:24
*** openstackgerrit has joined #openstack-lbaas22:37
openstackgerritMichael Johnson proposed openstack/octavia master: ACTIVE-ACTIVE: OVS diskimage-builder element  https://review.openstack.org/53384922:37
openstackgerritMichael Johnson proposed openstack/octavia master: ACTIVE-ACTIVE: OVS diskimage-builder element  https://review.openstack.org/53384923:33
« Wednesday, 2018-01-17 Index Friday, 2018-01-19 »

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