Tuesday, 2014-10-07

*** VijayB_ has quit IRC00:03
*** sbfox has joined #openstack-lbaas00:07
*** sbfox has quit IRC00:10
*** vivek-ebay has quit IRC00:13
*** markmcclain1 has quit IRC00:20
*** mlavalle has quit IRC00:36
*** Youcef has quit IRC01:00
*** ptoohill-oo has joined #openstack-lbaas01:24
*** fnaval has quit IRC02:04
*** fnaval has joined #openstack-lbaas02:17
*** vivek-ebay has joined #openstack-lbaas02:19
*** sbalukoff has quit IRC02:26
*** sbfox has joined #openstack-lbaas03:23
*** vivek-eb_ has joined #openstack-lbaas03:59
*** vivek-ebay has quit IRC04:00
*** amotoki has joined #openstack-lbaas04:19
*** vivek-ebay has joined #openstack-lbaas05:49
*** vivek-eb_ has quit IRC05:52
*** vivek-ebay has quit IRC05:53
*** ksamoray has joined #openstack-lbaas06:18
*** woodster_ has quit IRC06:51
*** ksamoray1 has joined #openstack-lbaas07:05
*** rm_you|wtf has joined #openstack-lbaas07:06
*** ksamoray has quit IRC07:06
*** rm_you| has quit IRC07:09
*** sbfox has quit IRC07:41
*** jschwarz has joined #openstack-lbaas07:48
*** sbalukoff has joined #openstack-lbaas08:07
*** openstackgerrit has quit IRC08:11
*** ksamoray1 has quit IRC08:58
*** ksamoray has joined #openstack-lbaas09:10
*** ksamoray has quit IRC10:13
*** jschwarz has quit IRC10:54
*** busterswt has quit IRC10:54
*** mordred has quit IRC10:54
*** jschwarz has joined #openstack-lbaas11:03
*** busterswt has joined #openstack-lbaas11:03
*** mordred has joined #openstack-lbaas11:03
*** ksamoray has joined #openstack-lbaas11:12
*** amotoki has quit IRC11:52
*** woodster_ has joined #openstack-lbaas12:26
*** ajmiller has joined #openstack-lbaas12:30
*** ksamoray has quit IRC12:35
*** ksamoray has joined #openstack-lbaas12:37
*** ksamoray has quit IRC12:42
*** openstackgerrit has joined #openstack-lbaas12:49
*** vivek-ebay has joined #openstack-lbaas13:28
*** ksamoray has joined #openstack-lbaas13:35
*** vivek-ebay has quit IRC13:49
*** markmcclain has joined #openstack-lbaas14:00
*** fnaval has quit IRC14:10
*** vivek-ebay has joined #openstack-lbaas14:30
*** dboik has joined #openstack-lbaas14:32
*** dboik has left #openstack-lbaas14:32
*** vivek-ebay has quit IRC14:34
*** fnaval has joined #openstack-lbaas14:52
*** markmcclain has quit IRC14:58
*** jschwarz_ has joined #openstack-lbaas15:07
*** jschwarz has quit IRC15:09
rm_workwho owns this? https://review.openstack.org/#/c/123487/215:09
rm_workdougwig: any idea?15:09
rm_workI need whoever it is to do their rebase15:09
rm_workfollowed by https://review.openstack.org/#/c/123491/ I guess, all the way up the chain to mine15:11
rm_workdougwig: maybe try hitting the rebase button and see if it works? >_>15:11
rm_workthat'd be magical15:12
*** mlavalle has joined #openstack-lbaas15:25
*** ajmiller_ has joined #openstack-lbaas15:28
*** jorgem has joined #openstack-lbaas15:29
*** ajmiller has quit IRC15:30
rm_workdougwig: dooooouglaaaas T_T15:45
dougwigwas driving to work.  sec.  :)15:49
rm_worknow https://review.openstack.org/#/c/123491/15:52
rm_workpress button, receive rebase15:52
*** jschwarz_ has quit IRC15:59
bloganrm_work: do you need some particular code from that rebase?16:04
dougwigrm_work: done16:10
blogandougwig: thats like giving a stray cat food16:10
bloganyou'll never get rid of him now16:10
rm_workblogan: i needed it rebased so I could push an update to *my* CR16:23
rm_workwe talked about this on the phone :P16:23
bloganrm_work: well i didnt get a reason16:33
bloganrm_work: but you can push without rebase to your review with -R16:33
blogangit review -R16:33
bloganin case it happens again in the future16:33
*** ksamoray has quit IRC16:35
*** ksamoray has joined #openstack-lbaas16:39
*** crc32 has joined #openstack-lbaas16:42
*** crc32 has quit IRC16:43
*** ajmiller has joined #openstack-lbaas16:43
*** ajmiller_ has quit IRC16:47
*** sbfox has joined #openstack-lbaas16:48
*** clev has joined #openstack-lbaas16:52
*** jschwarz has joined #openstack-lbaas16:56
*** ksamoray has quit IRC16:59
*** VijayB has joined #openstack-lbaas17:28
*** regXboi has joined #openstack-lbaas17:32
regXboigreetings - I'm looking for a pointer to documentation for the LBaaSv2 API - specifically how members are specified when adding them to pools.  I've poked around the wiki and some of the etherpads and nothing is jumping out at me17:33
*** jschwarz has quit IRC17:33
regXboiA pointer to the current code base would also be nice :-)17:33
*** ksamoray has joined #openstack-lbaas17:37
dougwigregXboi: hiya.  code, pull this review: https://review.openstack.org/#/c/123491/17:37
dougwigdocs, one sec.17:37
regXboidougwig: danke for the gerrit link17:38
dougwigthe specs got nuked when they slipped out of juno, and haven't been resubmitted for kilo yet, but you can find them here: https://github.com/openstack/neutron-specs/tree/29b84bfc3ea03c0d1ecbb4bbdc80b54551a9e972/specs/juno-incubator17:40
regXboiok, that's good enough for the moment17:40
regXboijust needs something I can review :)17:40
regXboiand familiarize myself with17:40
dougwigjust holler if you have any questions.17:41
*** ajmiller_ has joined #openstack-lbaas17:45
regXboithe biggest one that jumps out at me is why there isn't a way to identify members by neutron port...17:46
regXboior by (getting crazier) something like security group17:47
regXboibut let's just stay with the port question for now ;)17:48
*** ajmiller has quit IRC17:48
rm_workok, new barbican TLS code is out17:53
*** markmcclain has joined #openstack-lbaas17:53
rm_workit doesn't "work" yet17:53
rm_workbecause there are still changes necessary in the barbican client17:53
rm_workand I didn't update the config yet17:54
dougwigsuggest WIP, then.17:56
rm_workyeah go for it17:56
* rm_work still can't WIP17:56
bloganregXboi: lbaas v1 did not store a neutron port association with members, so V2 kind of followed suit17:57
rm_workhilariously though, I *can* +1 my own review17:57
bloganregXboi: I do believe it was up to the drivers to create a neutron port for the members or not17:57
*** jorgem1 has joined #openstack-lbaas17:57
*** jorgem has quit IRC17:57
*** jorgem1 is now known as jorgem17:58
bloganregXboi: I haven't looked into the code in a while so my memory is not so fresh, but I do not remember why it is that way17:58
regXboiblogan: that doesn't quite make sense - if I consider horizontal scaling, the first thing I'm going to do is deploy a new workload that will be added to the member list17:58
regXboiand then the second thing I'd do is add it to the member list and using a port uuid strikes me as more "the openstack way"17:58
regXboinow... honestly, I'm not 100% sure it *matters*17:59
regXboibut it just struck me as .... odd17:59
bloganim sure it won't be the last thing that will strike you as odd17:59
bloganbut I should talk to some of the original people who designed it that way as to the reasons why it is that way18:00
*** ajmiller_ has quit IRC18:05
*** ajmiller_ has joined #openstack-lbaas18:06
bloganregXboi: looks like what you want should be added as its own feature, so it should be its own blueprint and spec18:08
*** sbfox has quit IRC18:09
rm_workdougwig: let me know if that review looks better to you now18:09
rm_workat least, the parts I updated (barbican_utils)18:09
rm_workactually I am not sure about the filename barbican_utils.py anymore, was tempted to merge the TLSInfo class into tls_utils.py because, well, obvious18:09
dougwigi started it, and it was looking better.  tls_utils.TLSParseUtils bugged my redundancy filter.  :)18:09
rm_workyeah, though you can18:10
rm_workfrom tls_utils import TLSParser, TLSInfo18:11
rm_workthe Utils at the end is maybe a bit bleh18:11
rm_worknow I am rethinking TLSInfo >_>18:11
rm_worktls_utils.BarbicanCertificate.get() ?18:11
rm_workneed the whole cadre of jesters here to have a naming circus18:12
* regXboi hears echoes of "the merry go round broke down" in the background18:14
dougwigno, openstack hacking rules mean you can't import like that.18:15
*** VijayB has quit IRC18:18
*** VijayB_ has joined #openstack-lbaas18:20
rm_workdougwig: err, i don't think they do prohibit that...18:21
rm_workit's done in quite a few places IIRC18:21
rm_workyou can't do splat imports...18:21
rm_workok, so I see H30218:21
rm_workmaybe they don't enforce it and it is just ignored by a lot of people? :/18:22
*** ksamoray has quit IRC18:24
*** sbfox has joined #openstack-lbaas18:26
*** ksamoray has joined #openstack-lbaas18:27
*** ksamoray has quit IRC18:29
*** ksamoray has joined #openstack-lbaas18:30
*** ksamoray has quit IRC18:36
*** ksamoray has joined #openstack-lbaas18:37
*** ksamoray has quit IRC18:37
*** markmcclain has quit IRC18:38
*** ksamoray has joined #openstack-lbaas18:45
*** ksamoray has quit IRC18:51
*** TrevorV_ has joined #openstack-lbaas19:01
*** TrevorV_ has quit IRC19:02
*** VijayB_ has quit IRC19:10
*** HenryG has quit IRC19:28
dougwigit used to be enforced.19:46
*** markmcclain has joined #openstack-lbaas19:54
*** HenryG has joined #openstack-lbaas20:06
*** sbfox has quit IRC20:11
rm_worksballe: so if you're my new boss, will you send me to Paris? :P http://www.cbronline.com/news/hp-split-will-it-now-buy-rackspace-439686320:16
sballe:-) Off course20:20
sballeBut remember HP cut down on its attendance a lot too this time. xgerman and I are just lucky to go20:21
rm_workah, thought they were still sending everyone :)20:22
*** VijayB_ has joined #openstack-lbaas20:26
dougwigfloating ips - why are we worrying about them at the lbaas vip layer?  isn't that a function of neutron to map a FIP to any internal neutron IP port?  (not counting mapping members fip's; different topic.)20:28
blogandougwig: the reason is we (and I believe HP) is planning on having the floating IP be the customer facing VIP20:29
bloganand if lbaas creates the vip first, then the customer will see that IP first20:30
dougwigright, but that's a neutron function, not an lbaas function.  it's two steps: vip ip, then map a fip to that ip.  just like how you do it today.20:30
dougwigwhy wouldn't your UI just make two api calls?20:30
bloganbut how will neutron lbaas show the fip to the user?20:30
dougwigit's in the floating ip list, and horizon can optionally display it if it wants, just like nova.20:31
*** VijayB_ has quit IRC20:31
*** dboik has joined #openstack-lbaas20:31
bloganwell what i mean, strictly through the API, the vip_address comes back after a load_balancer create20:31
*** dboik has quit IRC20:31
bloganvip_address is already populated with the neutron port's fixed ip20:32
*** dboik has joined #openstack-lbaas20:32
bloganthe fip association with the neutorn port would happen at the driver layer, and since this is async the driver will not be able to change the vip_address shown to the user20:33
blogannow the driver could do teh fip association before it starts the async portion, but then we'd need to have the plugin show the fip address20:35
dougwiguse nova instances as a model; you get back the internal ip.  whether you choose to publish those to users is up to you, but nova show will have that internal IP.  the FIP mapping is separate.20:35
bloganand not the neutorn port's fixed ip20:35
dougwigi'm not sure why we'd deviate from the nova model here.20:35
bloganalright so you're saying that the user will be responsible for creating the FIP afterwards?20:36
dougwigor the operator's UI or wrapper, yes20:36
bloganI feel like that should automatically done for the user if teh user wants it20:37
bloganhowever, that is a reasonable starting point20:37
dougwigthere is a setting for that in nova, and i could see us having one as well.20:38
bloganbecause that requires no changes to lbaas20:38
dougwigyeah, that can go into the plugin at any time.20:38
bloganokay so now when we want it done automatically20:39
bloganwhat are you saying should happen?20:39
bloganthe operator UI or wrapper?20:39
bloganwhat do you mean by that20:39
*** sbfox has joined #openstack-lbaas20:40
dougwigi mean you'd hide two operations in one: neutron lb-vip-create blah; neutron floatingip-associate blah blah20:41
dougwigor we'd add a config item to auto-allocate and associate as the final step of lb-vip-create.20:41
bloganhow will the neutron lbaas api show the user what fip was createdf or them?20:42
dougwigneutron lb-vip-show would have tow IPs; the internal one and the fip.20:42
dougwigor, neutron floatingip-list would also contain the mapping.20:42
bloganokay so changes to lbaas then would be required20:42
bloganwell i think its better to have neutron lbaas show the fip if it auto-allocated20:42
*** dboik has quit IRC20:43
dougwigfor nova parity in features, yes, though today it happens by getting the vip internal ip and then parsing the floatingip-list manually.20:43
dougwigbecause *you can do this today*; it's just the display issue that doesn't happen.20:43
*** markmcclain has quit IRC20:50
*** TrevorV_ has joined #openstack-lbaas20:57
*** TrevorV_ has quit IRC21:04
*** dboik has joined #openstack-lbaas21:10
*** dboik has quit IRC22:00
*** VijayB has joined #openstack-lbaas22:08
*** jorgem has quit IRC22:08
*** sbfox has quit IRC22:17
*** sbfox has joined #openstack-lbaas22:27
*** VijayB has quit IRC23:00
*** VijayB_ has joined #openstack-lbaas23:04
*** mlavalle has quit IRC23:15
*** sbfox has quit IRC23:20
*** VijayB_ has quit IRC23:31
*** VijayB has joined #openstack-lbaas23:39
*** regXboi has quit IRC23:43
*** VijayB has quit IRC23:51
*** VijayB_ has joined #openstack-lbaas23:52

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!