Wednesday, 2014-08-13

*** xgerman has quit IRC00:07
*** fnaval has quit IRC00:24
*** fnaval has joined #openstack-lbaas00:25
*** fnaval has quit IRC00:29
*** sbfox has joined #openstack-lbaas00:29
*** markmcclain has quit IRC00:50
*** VijayB has quit IRC01:00
*** rm_you has quit IRC01:41
*** rm_you has joined #openstack-lbaas01:48
*** crc32 has quit IRC01:55
*** Krast has joined #openstack-lbaas02:00
*** woodster_ has quit IRC02:05
*** magesh_ has quit IRC02:06
*** Krast has quit IRC02:23
*** sbfox has quit IRC02:28
*** Krast has joined #openstack-lbaas02:42
*** fnaval has joined #openstack-lbaas02:47
*** vjay has joined #openstack-lbaas03:00
*** sbfox has joined #openstack-lbaas03:01
*** fnaval has quit IRC03:06
*** magesh_ has joined #openstack-lbaas03:10
*** magesh_ has quit IRC03:37
*** fnaval has joined #openstack-lbaas03:50
*** orion__ has quit IRC03:54
*** orion_ has joined #openstack-lbaas03:55
*** orion_ has quit IRC03:59
*** vjay has quit IRC04:02
*** magesh_ has joined #openstack-lbaas04:09
*** vjay has joined #openstack-lbaas04:24
*** fnaval has quit IRC04:49
*** rm_you| has joined #openstack-lbaas05:02
*** rm_you has quit IRC05:04
*** rm_you|wtf has joined #openstack-lbaas05:10
*** yfried__ has joined #openstack-lbaas05:13
*** rm_you| has quit IRC05:13
*** sbalukoff has quit IRC05:26
*** magesh_ has quit IRC05:52
*** HenryG_ has joined #openstack-lbaas06:09
*** HenryG has quit IRC06:12
*** vjay has quit IRC06:21
*** vjay has joined #openstack-lbaas06:43
*** evgenyf has joined #openstack-lbaas06:54
*** sbalukoff has joined #openstack-lbaas06:55
*** jschwarz has joined #openstack-lbaas07:09
*** vjay has quit IRC07:21
*** sballe has joined #openstack-lbaas07:29
*** vjay has joined #openstack-lbaas07:38
*** Youcef has joined #openstack-lbaas07:48
*** sballe has quit IRC07:51
*** Youcef has quit IRC08:00
*** sballe has joined #openstack-lbaas08:11
*** sbfox has quit IRC08:23
*** jschwarz has quit IRC08:29
*** sballe has quit IRC08:32
*** jschwarz has joined #openstack-lbaas08:35
*** jschwarz has quit IRC08:36
*** jschwarz has joined #openstack-lbaas08:36
*** enikanorov has quit IRC08:48
*** enikanorov has joined #openstack-lbaas08:48
*** vjay has quit IRC09:04
*** vjay has joined #openstack-lbaas09:08
*** sballe has joined #openstack-lbaas09:10
*** sballe has quit IRC09:20
*** sballe has joined #openstack-lbaas09:22
*** sballe_ has joined #openstack-lbaas09:24
*** sballe has quit IRC09:27
*** vjay has quit IRC09:35
*** vjay has joined #openstack-lbaas09:46
*** sballe__ has joined #openstack-lbaas09:46
*** sballe_ has quit IRC09:50
*** yfried has joined #openstack-lbaas09:53
*** yfried__ has quit IRC09:54
*** vjay has quit IRC09:57
*** vjay has joined #openstack-lbaas10:00
*** sballe has joined #openstack-lbaas10:06
*** vjay has quit IRC10:06
*** sballe__ has quit IRC10:07
*** sballe_ has joined #openstack-lbaas10:21
*** sballe has quit IRC10:21
*** sballe_ has quit IRC10:25
*** sballe has joined #openstack-lbaas10:26
*** sballe has quit IRC10:36
*** sballe has joined #openstack-lbaas10:37
*** yfried has quit IRC10:38
*** sballe_ has joined #openstack-lbaas10:44
*** sballe has quit IRC10:44
*** yfried has joined #openstack-lbaas11:05
*** [1]evgenyf has joined #openstack-lbaas11:14
*** yfried_ has joined #openstack-lbaas11:16
*** evgenyf has quit IRC11:16
*** [1]evgenyf is now known as evgenyf11:16
*** yfried has quit IRC11:16
*** yfried__ has joined #openstack-lbaas11:18
*** yfried_ has quit IRC11:21
*** sballe_ has quit IRC11:24
*** yfried has joined #openstack-lbaas11:52
*** yfried__ has quit IRC11:56
*** sballe has joined #openstack-lbaas12:34
*** sballe_ has joined #openstack-lbaas12:41
*** sballe has quit IRC12:44
jschwarzdougwig, ping12:49
*** yfried has quit IRC12:50
*** fnaval has joined #openstack-lbaas13:10
*** woodster_ has joined #openstack-lbaas13:12
*** sballe_ has quit IRC13:24
*** sballe has joined #openstack-lbaas13:51
*** orion__ has joined #openstack-lbaas13:53
*** sballe has quit IRC13:53
*** sballe has joined #openstack-lbaas13:56
*** sballe has quit IRC14:02
*** fnaval has quit IRC14:03
*** Krast has quit IRC14:13
*** Krast has joined #openstack-lbaas14:14
*** openstackgerrit has quit IRC14:16
*** openstackgerrit has joined #openstack-lbaas14:18
*** Krast has quit IRC14:28
*** Krast has joined #openstack-lbaas14:28
*** sbfox has joined #openstack-lbaas14:30
*** Krast has quit IRC14:31
*** sbfox has quit IRC14:31
*** Krast has joined #openstack-lbaas14:31
*** fnaval has joined #openstack-lbaas14:31
*** markmcclain has joined #openstack-lbaas14:48
*** sballe has joined #openstack-lbaas14:49
*** markmcclain1 has joined #openstack-lbaas14:51
*** markmcclain has quit IRC14:55
dougwigjschwarz: ack15:02
*** evgenyf has quit IRC15:04
jschwarzdougwig, good morning friend15:04
dougwigMorning. :)15:05
jschwarzIt would seem Kevin Benton is hard to reach, so I'm having trouble getting in touch with him visavi
jschwarzAlso it doesn't help that i'm in a different time zone than him (18:05 right now in Israel)15:05
jschwarzMaybe you can try pinging him during the day in #openstack-neutron?15:05
dougwigSure, I'll try.15:08
*** xgerman has joined #openstack-lbaas15:09
jschwarzGreat. Nothing fancy except letting him know of the issue and post a comment on the launchpad, unless he insists on patching it up himself ;-)15:09
*** yfried has joined #openstack-lbaas15:12
dougwigmorning.  you got double-rebased!15:24
blogani noticed15:24
bloganglad i didnt fix it last night15:24
ptoohillcrock of15:28
*** sbfox has joined #openstack-lbaas15:35
*** jschwarz has quit IRC15:36
*** magesh_ has joined #openstack-lbaas15:41
*** yfried has quit IRC15:53
*** magesh_ has quit IRC15:56
*** sbfox1 has joined #openstack-lbaas16:00
*** sbfox1 has quit IRC16:00
*** sbfox1 has joined #openstack-lbaas16:00
*** sbfox has quit IRC16:00
*** yfried has joined #openstack-lbaas16:01
*** yfried has quit IRC16:09
*** magesh has joined #openstack-lbaas16:25
*** sbalukoff has quit IRC16:28
*** rm_work has quit IRC16:29
*** rm_work has joined #openstack-lbaas16:30
*** rm_work has quit IRC16:34
*** rm_work has joined #openstack-lbaas16:35
*** rm_work has quit IRC16:35
*** rm_work has joined #openstack-lbaas16:35
*** rm_work has quit IRC16:41
*** rm_work has joined #openstack-lbaas16:42
*** rm_work has quit IRC16:42
*** rm_work has joined #openstack-lbaas16:42
*** rm_work has quit IRC16:47
*** rm_work has joined #openstack-lbaas16:47
*** mlavalle has joined #openstack-lbaas16:49
*** sballe has quit IRC16:50
*** rm_work has quit IRC16:52
*** VijayB has joined #openstack-lbaas16:53
*** rm_work has joined #openstack-lbaas16:53
*** rm_work has quit IRC16:53
*** rm_work has joined #openstack-lbaas16:53
*** rm_work has quit IRC16:58
*** rm_work has joined #openstack-lbaas16:58
*** rm_work has quit IRC16:58
*** rm_work has joined #openstack-lbaas16:58
*** VijayB has quit IRC17:03
*** rm_work has quit IRC17:03
*** rm_work has joined #openstack-lbaas17:04
*** rm_work has quit IRC17:04
*** rm_work has joined #openstack-lbaas17:04
*** rm_work has quit IRC17:08
*** sbfox1 has quit IRC17:08
*** sbfox has joined #openstack-lbaas17:09
*** sbfox has quit IRC17:09
*** sbfox has joined #openstack-lbaas17:09
*** rm_work has joined #openstack-lbaas17:10
*** VijayB has joined #openstack-lbaas17:11
*** rm_work has quit IRC17:15
*** VijayB has quit IRC17:16
*** rm_work has joined #openstack-lbaas17:19
*** rm_work has quit IRC17:19
*** rm_work has joined #openstack-lbaas17:19
*** sbfox has quit IRC17:24
*** sbfox1 has joined #openstack-lbaas17:24
*** rm_work has quit IRC17:26
*** rm_work has joined #openstack-lbaas17:27
*** sbalukoff has joined #openstack-lbaas17:31
*** rm_work has quit IRC17:32
*** rm_work has joined #openstack-lbaas17:32
*** rm_work has quit IRC17:37
*** rm_work has joined #openstack-lbaas17:38
*** rm_work has quit IRC17:38
*** rm_work has joined #openstack-lbaas17:38
*** fnaval has quit IRC17:39
*** rm_work has quit IRC17:43
*** rm_work has joined #openstack-lbaas17:44
*** rm_work has quit IRC17:44
*** rm_work has joined #openstack-lbaas17:44
*** rm_work has quit IRC17:49
*** rm_work has joined #openstack-lbaas17:49
*** rm_work has quit IRC17:50
*** rm_work has joined #openstack-lbaas17:50
*** rm_work has quit IRC17:54
*** VijayB has joined #openstack-lbaas18:00
*** rm_work has joined #openstack-lbaas18:05
*** rm_work has quit IRC18:05
*** rm_work has joined #openstack-lbaas18:05
*** yfried has joined #openstack-lbaas18:08
*** rm_work has quit IRC18:11
*** rm_work has joined #openstack-lbaas18:12
*** rm_work has quit IRC18:16
*** rm_work has joined #openstack-lbaas18:16
*** sbfox1 has quit IRC18:23
*** rm_work has quit IRC18:29
*** rm_work has joined #openstack-lbaas18:29
*** sbalukoff1 has joined #openstack-lbaas18:33
*** rm_work has quit IRC18:33
*** rm_work has joined #openstack-lbaas18:33
*** TrevorV_ has joined #openstack-lbaas18:35
*** blogan_ has joined #openstack-lbaas18:35
*** sbalukoff has quit IRC18:41
*** TrevorV has quit IRC18:41
*** blogan has quit IRC18:41
*** openstackgerrit has quit IRC18:41
*** TrevorV_ is now known as TrevorV18:41
*** sbfox has joined #openstack-lbaas18:44
*** openstackgerrit has joined #openstack-lbaas18:50
*** sbfox has quit IRC18:51
*** sbfox has joined #openstack-lbaas18:52
*** sbfox1 has joined #openstack-lbaas18:58
*** sbfox has quit IRC18:58
*** fnaval has joined #openstack-lbaas19:00
*** sbfox1 has quit IRC19:05
*** sbfox has joined #openstack-lbaas19:05
*** LouisF has joined #openstack-lbaas19:12
TrevorVdougwig, does the meeting number change at all?19:30
dougwigAlways the same19:34
TrevorVBrandon is trying to see if he can get it working on Linux, could you host it real quick dougwig ?19:39
TrevorVI can't do a login to be host :D19:39
dougwigAt lunch. I can in 10.19:40
TrevorVAlright thanks19:46
TrevorV(He can reboot into his OSX installation, he's just bein lazy)19:47
orion__Does anyone have any information regarding the "Load Balancing Methods" for Icehouse?  Specifically, what is "SOURCE_IP" balancing?  Is this actually referring to persistence?19:48
xgermanblogan ever heard about the cloud? Run your Linux in the cloud...19:49
xgermanI can hook you up with a box if that helps :-)19:49
blogan_orion__: we've had that same question and there is a minor different but it escapes me19:49
orion__blogan_: hah, ok :)19:50
blogan_orion__: i'm very helpful i know, sorry19:50
blogan_xgerman: can't I just toss my computer up in the air that is running linux and say its in the cloud?19:51
orion__blogan_: no worries.  i'll do a more exhaustive search later... moving on for now :)19:51
ptoohillorion__ SOURCE_IP balancing would be just that. There should also be SOURCE_IP session persistence which is a bit different from what i understand19:51
ptoohillFor lb_algo, it 'should' send connections to specified backends based off of the 'source_ip' in session persistence it should maintain the 'session' for that source_ip19:52
blogan_just a quick search on haproxy's stuff19:52
orion__blogan_: ...reading...19:53
*** fnaval has quit IRC19:54
ptoohillJava is not working.19:57
*** TrevorV_ has joined #openstack-lbaas19:58
sbalukoff1blogan_: For today's webex, can you buddy-up with someone else there?19:58
*** TrevorV_ has quit IRC19:59
blogan_sbalukoff1: whys that?19:59
TrevorVsbalukoff1, might be under the assumption that you can't join otherwise blogan_19:59
TrevorVwe're talking already and can't hear you if you are blogan19:59
TrevorVdougwig, will you be able to send me the recording links again so I can write up the minutes?20:00
sbalukoff1Yes, sorry, I was.20:00
*** sbalukoff1 is now known as sbalukoff20:00
*** blogan_ is now known as blogan20:00
*** fnaval has joined #openstack-lbaas20:01
*** VijayB has quit IRC20:05
dougwigno, i think only the host can record, which isn't me.20:33
dougwig(if you clicked the record button, yes, i can.)20:34
*** crc32 has joined #openstack-lbaas20:37
*** s3wong has joined #openstack-lbaas20:44
TrevorVDamnit... I forgot to click record.20:44
TrevorVI feel retarded20:44
TrevorVdougwig, I can't record this time... I'm host but it doesn't let me20:49
TrevorVIs that a feature that requires enabling or something?20:49
dougwigsbalukoff is the host20:49
*** VijayB has joined #openstack-lbaas20:50
TrevorVOh, he probably authenticated just before I did.  It had said I was going to Host, but oh well20:50
TrevorVI wonder if he clicked record... :(20:50
rm_workthe host is Stephen, so20:50
*** rm_work has quit IRC20:57
*** rm_work has joined #openstack-lbaas20:58
*** rm_work has quit IRC20:58
*** rm_work has joined #openstack-lbaas20:58
s3wongdougwig: hello21:00
dougwigs3wong: hiya.21:00
dougwigmost of us just got out of a webex, so we should get some good participation.21:01
s3wongdougwig: cool21:01
dougwigcan you link the spec here?21:01
*** rm_work has quit IRC21:01
*** hemanth_ has joined #openstack-lbaas21:01
dougwigto summarize, the port objects currently used by lbaas would be going away, and we'd have have many cycles to switch?21:03
s3wongdougwig (and other lbaas community friends): the idea is that each advanced service has its own method to insert into Neutron; and we want a more unified method'21:03
s3wongdougwig: the original task list have someone from the service insertion / advanced service side (me, actually) to do the work21:03
*** rm_work has joined #openstack-lbaas21:04
*** rm_work has quit IRC21:04
*** rm_work has joined #openstack-lbaas21:04
dougwigthat won't work with lbaas v2, which will might hit upstream at the same time as yours.21:04
dougwig(though anything making it is seriously in limbo right now.)21:04
s3wongdougwig: in the mean time, it is not disruptive, as we aren't immediately removing the port object unless it is approved by the lbaas team21:05
dougwigs3wong: see here:
dougwig(and follow the chain downhill)21:05
s3wongdougwig: as you saw today in the adv-service meeting: this framework change is even more in limbo :-)21:05
dougwigok, so it won't break anything.  can you give us a quick summary on the upside benefits?21:05
s3wongdougwig: instead of having lbaas doing all insertion work, it will be via a common framework from Neutron21:06
s3wongwe aim to extend ServicePluginBase to include methods to grab serviceInterface21:07
*** rm_work has quit IRC21:07
s3wongand since ALL service dbmixin inherits from ServicePluginBase, you get that for free21:07
*** rm_work has joined #openstack-lbaas21:07
s3wong(of course, still need to invoke the methods)21:07
*** rm_work has quit IRC21:07
*** rm_work has joined #openstack-lbaas21:07
dougwigso where are the dragons here that caused such a dustup at your meeting?21:08
s3wongdougwig: you know, I have no idea21:08
*** VijayB_ has joined #openstack-lbaas21:08
s3wongdougwig: for markmcclain1 - it is because it is only 9 days left before code needs to land. That is a legitimate concern21:08
s3wongfor marun, I actually don't know what his primary reason is21:09
s3wongbut as I told you guys in San Antonio, this is a very simple framework that we would like to introduce and unify21:09
dougwiglet me check if any of the current drivers are reaching around the plugin for this.21:09
s3wongand we have FWaaS and VPNaaS team backing it21:09
s3wongdougwig: OK. Thanks21:09
dougwigonly one driver directly references Port, so i'm not seeing any major heartache on our side.21:10
s3wongdougwig: good to know21:11
s3wongdougwig: who is that driver, perhaps I should talk to that vendor's rep also21:11
*** a2hill has joined #openstack-lbaas21:12
dougwigare there any weirdo network schemes you can wire up with the current interface that you can't with the new one?  from a brief scan of your doc (sorry, have had no free time since this came up at your meeting), it didn't look like it.  just an extra layer of abstraction, which could be gotten around.21:12
*** VijayB has quit IRC21:12
*** sbfox has quit IRC21:12
*** ptoohill has quit IRC21:12
bloganaren't all the drivers creating ports for the memebrs?21:12
blogansorry for the pool21:12
*** sbfox has joined #openstack-lbaas21:13
*** magesh has quit IRC21:13
bloganbecause the pool port creation is not done at hte plugin level, so its up to the drivers to do that21:13
s3wongdougwig: not that I know of - in fact, almost all drivers should insert via Neutron port today, right?21:13
dougwigdid we server split, or is it just me?21:13
blogancan you see me?21:14
bloganis anyone out there?21:14
dougwigblogan: for hardware, no, we assume hardwired/routed.21:14
dougwigwell, we can.21:14
s3wongdougwig: blogan: so the new service insertion framework actually introduces more ways to insert: via external port (a controversial point, I know) or directly to another service (like a router)21:14
dougwigi can see you.. my writes are taking a long time.21:14
sbalukoffI'm here.21:14
sbalukoffI think.21:14
*** magesh has joined #openstack-lbaas21:14
bloganokay so I guess i don't understand the problem here then21:14
*** rm_work has quit IRC21:14
s3wongblogan: pool port? so there is a Neutron port for a pool?21:14
dougwigblogan: i don't either.  seems fine.21:14
dougwigs3wong: he means vip.  lbaas v1 put it on the pool 1:1.21:15
blogans3won: no there isn't, but i have seen drivers go and create them21:15
s3wongdougwig: blogan: hmm... OK, pool isn't really a service, so it doesn't inherit from ServiceBasePlugin...21:15
dougwigblogan: only one driver doesn't do it via the plugin methods, which means the implementation is easy to swap.21:15
blogani couda sworn radware did that21:16
bloganhold on21:16
dougwigi saw them calling a plugin method to do it.21:16
dougwigbut yeah, if my grep missed something, please check.21:16
s3wongblogan: but if the pool port isn't created via plugin, how can network driver know of its existance?21:16
bloganoh calling the _core_plugin is fine then?21:16
blogansorry i have no context21:17
s3wongblogan: I think that is OK - though not "service insertion", but at least the driver know where in the network the pool is inserted21:17
s3wongdougwig: blogan: given the time conflict on us landing (if at all), I am OK with deferring making the change on lbaas21:18
dougwigblogan: oh, i missed it reaching in that deep.21:18
*** rm_work has joined #openstack-lbaas21:18
*** rm_work has quit IRC21:19
*** rm_work has joined #openstack-lbaas21:19
s3wongdougwig: blogan: like when enikanorov originally planned for flavor, perhaps we can first experiment with just fwaas21:19
blogans3wong: sorry still don't have the context, so what exactly is it that you need from lbaas v1?21:19
dougwigblogan: see the spec he posted a few screens up.  it deprecates the Port object.21:19
s3wongblogan: lbaas v1 has no load balancer object, right? just VIP and pool?21:19
dougwigs3wong: given the timing, i'd agree on holding off modifying lbaas, yes.21:20
sbalukoffs3wong: Correct. 'loadbalancer' is a v2 object.21:20
s3wongblogan: so just deprecating the port object reference from lbaas21:20
blogans3wong: ah instead of going through _core_plugin, just go through an abstraction layer21:20
dougwigi think it's on loadbalancers and members in v2, right?21:20
blogandougwig: yes21:21
s3wongblogan: yes21:21
s3wongdougwig: members also... OK, will take a look21:21
*** LouisF has quit IRC21:22
bloganokay another question then21:22
s3wongdougwig: just curious, members are VMs, right? why would LBaaS be managing the ports there?21:22
dougwigor if we don't have a port in there, we allow specify a subnet id ,which means at some point drivers will begin plumbing in ports.21:22
dougwigmembers are VMs or any IP.21:22
dougwigyou can lbaas, if you're so inclined.21:22
blogans3wong: the workflow for the haproxy driver is that it creates the port first, then send it to the agent, which will then in turn update that port to activate21:23
dougwigblogan: that's for the vip, right?21:23
blogandougwig: correct21:23
s3wongblogan: hmm... very tightly coupled with port21:25
blogans3wong: yes it appears21:25
*** orion_ has joined #openstack-lbaas21:26
blogans3wong: does service insertion only create the port once and not abel to update afterwards?21:26
s3wongblogan: so haproxy driver runs in a namespace; so it will create a Neutron port (upon knowing which network it should be plugged into), then inform agent to do whatever config; then after config, the driver will activate the port21:26
blogans3wong: yes21:27
s3wongblogan: update... meaning you want to unplug from a network and plug into a different network?21:27
blogans3wong: well the plugin creates the port21:27
s3wongblogan: in service insertion we allow you to add interface (or remove)21:28
blogans3wong: not that, but update device_ownder, device_id, and the bindings host_id21:28
blogans3wong: and admin_state_up, which I'm not sure if it'd be a problem if the port is active before the vip is actually active21:29
bloganbut its there21:29
*** orion__ has quit IRC21:30
s3wongblogan: I see, LBaaS manages port state in accordance to when an instance is ready21:30
blogans3wong: you mean the plugin does?21:30
s3wongblogan: actually that was my question - LBaaS controls the admin states / ownership of the port via plugin interface, right?21:31
*** orion_ has quit IRC21:31
blogans3wong: so the plugin will initially create the vip port, and set admin_state_up to False.  It passes that to the driver.  It is the driver's responsibility to activate it, in which they go through the neutron core_plugin to update said port21:32
blogans3wong: so it's not calling back to the plugin to update it, other than calling plugin._core_plugin.update_port21:33
blogans3wong: by it i mean the driver21:33
*** rm_work has quit IRC21:33
blogans3wong: well and the plugin too to create the port, it'll call _core_plugin21:33
*** rm_work has joined #openstack-lbaas21:33
blogans3wong: am I making sense?21:34
s3wongblogan: OK, I see. ServiceInterface allows you to insert into network, but currently we are NOT exposing methods to allow service to manipulate port states21:34
s3wongblogan: yes21:34
blogans3wong: okay good in that I am making sense21:34
blogans3wong: bad in that it won't work for SeviceInterface right now21:35
blogans3wong: so yeah leaving lbaas v1 out for now is probably the best way to go21:35
s3wongblogan: it is a good feedback though. As we are focusing on having ServiceInterface as an abstraction to allow service to insert in different ways21:35
blogans3wong: because then all the drivers would need to be updated21:35
*** rm_work has quit IRC21:35
s3wongblogan: but we aren't accounting for services' needs to manipulate ports directly21:36
*** rm_work has joined #openstack-lbaas21:36
s3wongblogan: when are you guys going to deprecate lbaas v1?21:36
blogans3wong: So is a requirement that all advanced services need to use service insertion?21:36
blogans3wong: that is an unknown right now, v2 will most likely go into the incubator21:36
s3wongblogan: "requirement" is a strong word :-)21:37
blogans3wong: and v1 will most likely not be deprecated until v2 gets graduated.21:37
s3wongblogan: but we hope that we can flush out the interfaces/APIs such that all advanced services have a unified way to insert to Neutron network21:37
s3wongblogan: markmcclain1 also presented the incubator idea during our GBP meeting21:38
dougwigs3wong: i've noticed you reference instances a few times.  lbaas backends can be hardware, not just nova instances plugged via neutron.  similar, the servers that are being balanced do not have to be openstack, today.21:38
blogans3wong: yes and an abstracted layer in which we need to communicate with neutron would be beneficial to remove all the tight coupling21:38
s3wongblogan: so within two releases then21:38
blogans3wong: lol maybe? I hope? need more details on the incubator and graduation process21:39
s3wongdougwig: yes. Sorry for the misnomer. We definitely accounted for 'external port' case where the backend is hardware21:39
s3wongblogan: OK. Makes sense. We aren't going to introduce service insertion to lbaas v1 then21:40
blogans3wong: probably best, sorry that I wasn't involved in that spec review, probably around the time I was learning about this21:40
s3wongblogan: we did talk about it a bit during San Antonio, if you remember :-)21:41
dougwigblogan: that makes almost all of us.21:41
blogans3wong: lol yeah but even then I was still missing a lot of details21:41
s3wongblogan: true21:41
blogans3wong: now having worked through it all I have a much clearer picture, but still there is some fuzziness21:42
s3wongblogan: please let us know (or ask now if you can articulate your concerns)21:42
blogans3wong: no i think I get the service insertion now, but things with lbaas such as would it be okay to have the port be active before the driver actually creates the loadbalancer21:43
s3wongblogan: dougwig: the framework certainly isn't set in stone - we would love to get feedback from all services21:43
blogans3wong: not sure if that makes a different, I would assume so since it has been coded that way21:43
blogans3wong: i need to attend the advanced services meetings for sure21:44
s3wongblogan: sounds good. If lbaas is to be split from Neutron in the future, a cleaner interface without invoking _core_plugin may be a good idea21:44
blogans3wong: ive been putting that off, too much to do21:44
blogans3wong: definitely! that was my exact thought21:44
s3wongdougwig: blogan: thanks for you guys' time21:45
blogans3wong: though if service insertion will allow me to update the port, but the API still won't allow the updating of some of the attributes needed, it won't help in that regard21:45
dougwiganytime, feel free anytime.21:45
blogans3wong: still a step in the right direction though21:45
s3wongblogan: that is good feedback, and I will look into how to unify port related state info on this abstraction21:45
blogans3wong: is service insertion using the _core_plugin?21:46
s3wongblogan: I think at the very least, admin state up/down needs to be supported by ServiceInterface21:46
s3wongblogan: for Neutron port, yes21:46
*** crc32 has quit IRC21:47
s3wongblogan: for external port, would have to defer to kevinbenton to know how that framework works21:47
s3wongblogan: for service (router, for example). it doesn't need to21:47
blogans3wong: there's still some other attributes that are updated on the port that we shoudl research more21:48
blogans3wong: see how much they are needed21:48
s3wongblogan: yes. please let me know. certainly we can't possibly satisfy all services now, but feedbacks will make the framework better21:49
bloganshould I join the advanced services meeting and send feedback there? or is there review with code up?21:49
dougwigby admin-state, you mean that there will be an interface to call into the service interface code for admin state, if it's a neutron port, right?  not that it'll handle it directly?21:50
s3wongblogan: yes (that would be great) and somewhat yes (WIP code just went up, though it is very preliminary)21:50
dougwigcan you link the review, or add it to ?21:50
s3wongdougwig: yes. and I would imagine it would also work for Neutron port21:50
s3wongdougwig: though depending on how external port is implemented, we can possibly manipulate the state there as well21:51
s3wongdougwig: sure21:51
s3wongdougwig: I will keep on updating the code leading up to next Thursday21:51
dougwigfor reference, i use admin_state to actually influence the state of the logical objects on our hardware appliance, which is more than just network access.21:51
s3wongdougwig: blogan: though from how contentious today's meeting look, this may also go to incubator first :-)21:51
blogans3wong: well it will live with v2 then21:52
dougwigs3wong: that was a surprisingly heated meeting.21:52
bloganadvanced services meeting?21:52
s3wongdougwig: Oh, OK - so for your driver (A10?), it also intercept port status?21:52
s3wongdougwig: I know. But since I am also in GBP, I kind of gotten used to it by now :-)21:53
blogans3wong: it would have to because the plugin creates the port in a down state21:53
dougwigthe drivers get admin_state changes, and i do stuff with it.21:53
bloganand every driver would have to update admin_state_up21:53
*** crc32 has joined #openstack-lbaas21:53
s3wongdougwig: blogan: OK - that is good to know. So transition on port status does notify LBaaS drivers... good to know21:54
s3wongblogan: dougwig: so in conclusion, we won't transition lbaas v1 to utilize service insertion framework; and will work with you guys to ensure we have enough features to help lbaas v2 to adopt to this framework21:56
blogans3wong: that sounds like the best option21:56
blogans3wong: and I will try to attend the advanced services meetings21:56
dougwigs3wong: +121:57
blogans3wong: when are those meetings again? i assume they are on wednesdays21:57
s3wongblogan: you will be most welcome there. dougwig already shows up here and there in that meeting21:57
dougwigblogan: wednesday at 11:30 mountain21:57
blogans3wong: dougwig is everywhere21:57
*** HenryG_ is now known as HenryG21:57
dougwigi was mostly flitting in and out for flavors, but i'll pay attention more.21:57
s3wongblogan: Wednesday 17:30 UTC21:57
bloganalright I shall set my calendar21:58
s3wongblogan: dougwig: yeah - I notice dougwig is also in tacker (serviceVM) meetings :-)21:58
dougwigblogan: #openstack-meeting-321:58
s3wongblogan: yes, as dougwig said, #openstack-meeting-321:58
dougwigha, that one is hard, since i'm mostly asleep for that one.21:58
s3wongblogan: dougwig: thanks for the discussion. Looking forward to discussing things with you guys moving forward21:59
s3wongdougwig: likewise, lbaas meeting is either 6am or 7am for me, so difficult to attend all the time :-)21:59
a2hilldid this go through?21:59
dougwigs3wong: totally agree.  and now my only meeting that wasn't either 1) super early, 2) super late, or 3) over lunch, which is neutron, is moving to a rotating schedule that is even EARLIER than lbaas once a week.  sigh.  :)22:00
dougwigevery other week, i mea.22:00
dougwiglong day.22:00
s3wongdougwig: :-)22:00
a2hillCan anyone see this message?22:02
dougwiga2hill: no, you're crazy.22:02
a2hillook, that's good :)22:02
a2hillHaving funky issues with freenode today it seems22:03
dougwigfreenode has been out back smoking a joint for about 2 hours now.22:06
a2hillheh, explains a lot.22:06
*** hemanth_ has quit IRC22:07
*** crc32 has quit IRC22:33
*** fnaval has quit IRC22:48
*** crc32 has joined #openstack-lbaas22:49
*** markmcclain1 has quit IRC22:52
*** markmcclain has joined #openstack-lbaas23:08
*** fnaval has joined #openstack-lbaas23:27
*** markmcclain1 has joined #openstack-lbaas23:56
*** markmcclain has quit IRC23:56
*** xgerman has quit IRC23:59

Generated by 2.14.0 by Marius Gedminas - find it at!