Monday, 2015-03-09

*** sbalukoff has joined #openstack-lbaas00:08
*** banix has quit IRC00:20
*** sbalukoff has quit IRC00:21
*** busterswt has joined #openstack-lbaas00:37
*** ajmiller has quit IRC00:49
*** banix has joined #openstack-lbaas00:56
*** woodster_ has quit IRC01:10
*** banix has quit IRC01:35
*** banix has joined #openstack-lbaas01:41
*** sbalukoff has joined #openstack-lbaas01:51
*** banix has quit IRC01:59
*** banix has joined #openstack-lbaas02:31
*** Krast has joined #openstack-lbaas02:32
*** enikanorov has quit IRC02:40
*** enikanorov has joined #openstack-lbaas02:40
*** sbalukoff has quit IRC03:04
*** banix has quit IRC03:29
openstackgerritBrandon Logan proposed openstack/neutron-lbaas: Pulling in tempest files
*** vivek-ebay has joined #openstack-lbaas03:50
*** sbfox has joined #openstack-lbaas03:53
*** woodster_ has joined #openstack-lbaas04:37
*** sbfox has quit IRC04:46
*** sbfox has joined #openstack-lbaas04:50
*** SumitNaiksatam has quit IRC05:15
*** Varun_Lodaya has joined #openstack-lbaas05:37
*** sbalukoff has joined #openstack-lbaas05:40
*** vivek-ebay has quit IRC05:40
*** Varun_Lodaya has quit IRC06:29
*** woodster_ has quit IRC06:40
openstackgerritBrandon Logan proposed openstack/neutron-lbaas: Pulling in tempest files
openstackgerritBrandon Logan proposed openstack/neutron-lbaas: Neutron LBaaS: Initial Load Balancer Tests
*** apuimedo has joined #openstack-lbaas07:07
*** sbfox has quit IRC07:42
*** Miouge has joined #openstack-lbaas07:44
*** chlong has quit IRC08:07
openstackgerritKobi Samoray proposed openstack/neutron-lbaas: VMWare Edge loadbalancer v1 driver
*** amotoki has joined #openstack-lbaas08:33
*** enikanorov_ has joined #openstack-lbaas08:35
*** vivek-ebay has joined #openstack-lbaas08:41
*** vivek-ebay has quit IRC08:45
*** BeardyMcBeards has quit IRC09:04
*** BeardyMcBeards has joined #openstack-lbaas09:06
*** chlong has joined #openstack-lbaas11:03
*** woodster_ has joined #openstack-lbaas11:18
*** kobis has joined #openstack-lbaas11:53
*** Miouge_ has joined #openstack-lbaas12:01
*** Miouge has quit IRC12:01
*** Miouge_ is now known as Miouge12:01
*** Miouge has quit IRC12:50
*** Miouge has joined #openstack-lbaas12:50
*** chlong has quit IRC13:04
*** Miouge has quit IRC13:04
*** amotoki has quit IRC13:04
*** amotoki has joined #openstack-lbaas13:14
*** jschwarz has joined #openstack-lbaas13:15
*** chlong has joined #openstack-lbaas13:17
*** woodster_ has quit IRC13:30
*** banix has joined #openstack-lbaas13:52
*** sballe__ has joined #openstack-lbaas14:01
*** woodster_ has joined #openstack-lbaas14:07
*** jamiem has joined #openstack-lbaas14:15
*** vivek-ebay has joined #openstack-lbaas14:33
*** vivek-ebay has quit IRC14:47
*** vivek-ebay has joined #openstack-lbaas14:54
*** juliancash_ has quit IRC14:54
*** vivek-ebay has quit IRC14:54
*** Varun_Lodaya has joined #openstack-lbaas15:04
*** xgerman has joined #openstack-lbaas15:20
*** Varun_Lodaya has quit IRC15:49
*** kobis has quit IRC16:03
*** woodster_ has quit IRC16:10
*** apuimedo has quit IRC16:11
*** mlavalle has joined #openstack-lbaas16:13
jschwarzdougwig, ping. can I bother you with an lbaas v1 question? :)16:28
dougwigjschwarz: shoot16:31
*** Varun_Lodaya has joined #openstack-lbaas16:32
jschwarzdougwig, what's the wanted behavior of the following scenario:16:32
jschwarzdougwig, create a network, create an lbaas pool attached to the network, remove the network16:33
dougwigoh, and it'll delete because it has no ports attached, right?16:33
jschwarzdougwig, should the pool still be there even if the network and subnet it was connected to no longer exist?16:33
jschwarzdougwig, yep - only a pool, no ports16:33
dougwigi would think the network delete should fail, if the pool is really associated to it.16:34
jschwarzdougwig, was afraid you'd say that :)16:34
jschwarzdougwig, alright, will have a look at the code and suggest a patch. thanks16:34
*** ajmiller has joined #openstack-lbaas16:35
*** sbfox has joined #openstack-lbaas16:38
*** fnaval has joined #openstack-lbaas16:42
bloganjschwarz: i believe the reason for that is when you create a pool a port is not actually created on that subnet/network, unless a driver does it16:45
bloganjscwarz: which i dont believe the ref impl driver creates a port either16:46
jschwarzblogan, yes, I agree. but logically the pool is attached to a subet16:46
bloganjschwarz: yeah but i dont believe there is even a foreign key to it16:47
jschwarzblogan, removing such a subnet (as I described) pops up errors on horizon's side16:47
jschwarzblogan, I'm fine with going for a horizon-fix, but am conflicted if the fix shouldn't be at the source16:47
bloganjschwarz: i havent been in v1 land in a while, but i bet horizon has its own validation16:48
bloganjschwarz: that type of thing should be done at the source16:48
jschwarzblogan, well, on my setup (and on a client's) this raises an error16:48
jschwarzblogan, let me pastebin, sec16:48
bloganjschwarz: what does? trying to delete the network with an attached pool through the API? or through horizon?16:49
jschwarzblogan, opening the pool's page gives this error on upstream16:49
jschwarzblogan, but the client says that on earlier versions the error causes the entire 'load balancer' tab to be unaccessible16:50
*** vivek-ebay has joined #openstack-lbaas16:50
bloganyeah thats an odd error, not really too informative other than its not expecting a list and getting it somewhere16:51
bloganwell in that location16:51
bloganwhich i have 0 experience with horizon unfortunately16:51
jschwarzblogan, client reported this in icehouse IMO16:51
jschwarzblogan, upstream allows deleting the pool as long as this is done from the load balancer screen16:52
jschwarzblogan, either way i agree that this should be treated at the source16:52
bloganwell that is failing on a GET16:52
bloganis that after a DELETE of a pool?16:52
jschwarzthat is when trying to enter the page of a pool16:52
jschwarzso 'pool show' probably16:53
jschwarz(CLI doesn't cause problems obviously)16:53
bloganyeah, so thats a horizon issue, but what actions did yout ake to get this to start?16:53
jschwarzblogan, I created a network, created a pool, attached a healthmonitor to the pool (probably not needed), deleted the network16:53
jschwarzblogan, then, tried to "view" the pool's information on the loadbalancer tab16:54
jschwarzie. press on the '<pool name>' link16:54
bloganjschwarz: i see, but the cli will show you the pool's information without error?16:54
bloganjschwarz: i wonder if horizion is trying to get the network/subnet information that you deleted and its failing there16:55
*** Krast has quit IRC16:55
*** Krast has joined #openstack-lbaas16:56
jschwarzblogan, yes re: cli question16:56
jschwarzblogan, neutron lb-pool-show shows that the pool's subnet_id field is non-empty (pointing to old subnet)16:57
jschwarzprobably horizon is trying to show the subnet's name or something... let me check one sec16:57
jschwarzblogan, yeah, horizion is displaying mininalistic data on the subnet16:58
jschwarzname and actual subnet CIDR16:58
jschwarz(with a link)16:58
bloganjschwarz: yeah there's the issue, of course the root of the issue is the pool registering itself in some form or fashion as being on that subnet16:59
bloganjschwarz: or lack of16:59
jschwarzblogan, yep.17:00
jschwarzblogan, so I'll look into patching db_base_plugin (on neutron's side) as it seems to be the code should go in there17:00
bloganjschwarz: are you going to check the pools table on delete or something?17:01
jschwarzblogan, hopefully - try to retrieve all the pools connected to the subnet which is deleted and if there is raise an exception17:02
bloganjschwarz: i'm not sure that is the right way to go as the pool technically has no links to that subnet, other than a db entry.  i think the two ways to solve it are make sure the pool creates a port on that subnet, or the pool's subnet_id has a foreign key to the subnets table17:03
bloganthe 2nd one i'm not sure about either bc we're at the point where we are starting to think about having neutron-lbaas being in its own db, which obviously can't have a foreign key to the neutron db17:04
jschwarzblogan, I also dislike the foreign key. backporting this would be hell and I don't think that's the correct solution here17:04
jschwarzblogan, also I think that adding a new port for each pool is a bit of an overkill.. why do you think that's needed?17:05
bloganjschwarz: well replace "create port" with "register with", some mechanism to do that17:06
jschwarzblogan, we could achieve the same functionality through simply checking if a pool is associated to a subnt17:06
bloganjschwarz: but if neutron-lbaas is in its own db that check may not be possible17:06
bloganthough v1 tables will most likely remain in neutron17:06
bloganbtw the thought of neutron-lbaas in its own db is something nto discussed much, and probably only applicable to v2, we'll be discussing it more in vancouver17:07
jschwarzI see17:07
bloganjschwarz: as you can tell, i'm uncertain about the way forward17:07
jschwarzblogan, I still don't see why creating a port for each pool is needed here though17:08
bloganjschwarz: i have a feeling that neutron wouldn't accept code that checks the pools table though.  i coudl be wrong17:08
jschwarzblogan, uhm...17:08
bloganjschwarz: oh i agree its overkill, but i said above, there needs to be some way to register, creating a port was a simplistic way17:08
jschwarzblogan, well there's an API to check exactly these kind of things17:09
jschwarzblogan, but I see your point now17:10
bloganjschwarz: i would say ask the neutron cores about modifying db_base_plugin to check the pools table17:11
jschwarzblogan, the file has checks to the VPNaaS17:11
jschwarzblogan, specifically "vpnservice.check_router_in_use"17:12
bloganjschwarz: i have a sneaking feeling they have some other way to do this17:12
jschwarzblogan, so it isn't unheard of17:12
bloganjschwarz: i have very limited knowledge of vpnaas, but i think there's something exceptional behind that, don't quote me on that though17:13
bloganjschwarz: pc_m would probably know about that17:13
jschwarzblogan, nevermind, I've been looking at an outdated code - armax added merged a patch 5 days that removed exactly this:
*** sbalukoff has quit IRC17:15
bloganah yeah thats what i was thinking about when i said "some other way to do this", the event registry17:15
jschwarzblogan, lol :)17:15
jschwarzwhat is this registry? where did it come from? <.<17:16
bloganjschwarz: i remember reading a spec and the code, it got merged recently17:16
jschwarzblogan, well, better do some digging17:17
jschwarzI'll report back when I have something, though it might take a few days17:17
jschwarzblogan, thanks :)17:20
*** Varun_Lodaya has quit IRC17:20
jschwarzblogan, so if I use that API, will that be an acceptable approach for you?17:21
jschwarzblogan, or do you think we should add a port regardless?17:22
bloganjschwarz: i need to read up on it in detail, but thats a form of registring that i believe works no matter where neutron-lbaas's db is located.  but i need to read in detail into it17:23
bloganjschwarz: TL;DR: yes, as long as i undersrtand it how i think i do17:23
*** madhu_ has joined #openstack-lbaas17:34
jschwarzblogan, the registry code seems very straightforward and I think we can use it17:48
*** madhu_ has quit IRC17:48
jschwarzblogan, it will however make backport a huge pain since this is new code :<17:49
*** madhu_ has joined #openstack-lbaas17:49
jschwarzblogan, I'm inclined to just create a port when attaching a pool to a subnet :)17:49
bloganjschwarz: lol coming full circle.  so for v1 that would work, though thinking about it more, I wonder if that would break existing deployments in some way17:51
jschwarzblogan, gah.17:51
*** banix has quit IRC17:52
bloganjschwarz: lol i know, never now if someone relies on that lack of functionality17:52
bloganjschwarz: if horizon handled it gracefully would that suit your needs?17:52
bloganjschwarz: but thats just duct tape, bleh17:52
jschwarzmaybe a db migrate fix? :)17:52
jschwarzcan add a port for each pool...17:53
bloganthat has a lot of different ramifications17:53
bloganquotas, can a subnet handle another port, things im not evne away of17:53
bloganaware of17:53
jschwarzoh yey!!!17:54
jschwarzwell we could duct tape it nicely ^_^17:55
bloganjschwarz: that might be the best bad idea we have17:55
jschwarzblogan, let me sleep on this for a few days. I'll get back to you17:57
bloganjschwarz: okay thanks, also keep in mind we're doing a v2 of lbaas, but that can use the event registry code17:57
jschwarzblogan, aye, I heard such a rumor17:58
jschwarzblogan, doubt though that the v2 will make the same mistake twice ;-)17:58
bloganjschwarz: haha famous last words17:59
*** sbfox has quit IRC18:00
*** crc32 has joined #openstack-lbaas18:05
*** chlong has quit IRC18:05
*** amotoki has quit IRC18:09
*** amotoki has joined #openstack-lbaas18:10
*** mwang2 has joined #openstack-lbaas18:11
*** sbfox has joined #openstack-lbaas18:13
*** jschwarz has quit IRC18:14
*** sbalukoff has joined #openstack-lbaas18:23
*** banix has joined #openstack-lbaas18:32
*** juliancash has joined #openstack-lbaas18:49
juliancash^ oops18:55
madhu_blogan: for this patch, where can I see in api/ directory in-order to run the tests?19:08
*** SumitNaiksatam has joined #openstack-lbaas19:09
*** busterswt has quit IRC19:11
madhu_or should I need to run the tempest tests in this way? 1. Pull this and pull on top of it?19:12
madhu_blogan, fnaval : ^^19:12
*** vivek-ebay has quit IRC19:13
madhu_xgerman, fnaval,juliancash, blogan: I am able to run the tests using blogan patch( and fnaval patch (
*** jorgem has joined #openstack-lbaas19:20
xgermanthanks madhu_19:25
*** Aish has joined #openstack-lbaas19:31
*** Varun_Lodaya has joined #openstack-lbaas19:37
*** woodster_ has joined #openstack-lbaas19:41
fnavalmadhu_: awesome to hear that19:43
fnavalsorry for the late reply19:43
bloganmadhu_: i may end up going a different route than that review, im finalizing it right now, looks like it'll work and be a lot easier19:49
Aishblogan: i tried the same. But, why should we run the lbaasv2 agent manually? and while doing that i am getting this error..
fnavalAish: I don't believe that it needs to be run manually now.  though blogan would know20:00
fnavalor ptoohill20:00
fnavaldoes the agent still need to be run manually?20:01
fnavalwith the latest code?20:01
ptoohillthe latest devstack?20:01
bloganit does until devstack automatically starts it, which i believe ajmiller has a patch up for20:02
madhu_I ran the lbaasv2 haproxy agent manually, then hitting the tests... (but I didnt want to pull ptoohills patch though as it is merged in master)20:02
bloganthat review will get the agent going automatically20:03
blogannot merged yet20:03
fnavalah ok.  then it has to run manually20:03
Aishfnaval blogan: thnks.. So it has to be run manually until this patch is merged20:03
bloganAish: correct20:03
bloganAish: or you can point your localrc to pull that review down and it'll do it for you20:03
ptoohillAnd no need to worry about the driver review either, thats been merged20:04
Aishoh.. thnks blogan ptoohill20:05
*** sbfox has quit IRC20:05
ajmillerblogan pthooil fnaval There are two neutron-lbaas patches and 1 devstack patch make things work automatically20:05
ajmillerThe neutron-lbaas patches are ready for merge (need some final reviews).20:05
ptoohillAh, so a few things need to be checked out/merged20:05
bloganajmiller: what two neutron-lbaas patches?20:05
ajmillerThe devstack patch will depend on a grenade patch that i'm working on.20:06
ajmillerjust a sec..20:06
bloganajmiller: oh you mean the devstack repo patch, and your service_provider neutron-lbaas patch and the agent patch20:06
*** sbfox has joined #openstack-lbaas20:07
bloganyeah that service_provider one will need to be added as well, so Aish it probably won't automatically start the v2 agent up wiht just that one patch20:07
bloganwell it'll start it up, but the neutron-server won't be loaded with the driver that would send data to the agent20:07
ajmillerBTW, Ive been traveling Friday - tomrrow, not very well connected to work, but trying to keep things moving along20:09
ajmillerblogan I think that with all of these patches, the server will load the driver.  If that isn't the case, let me know.20:10
bloganajmiller: i'll have to test it out to be sure, it looks like it will but whenever ig et time ill let you know, gotta focus on the tempest stuff first though20:11
ajmillerNP.  I'll be back full time tomorrow afternoon.  I will also be testing some of this as part of the grenade work too...20:11
*** vivek-ebay has joined #openstack-lbaas20:14
*** vivek-ebay has quit IRC20:19
*** sballe__ has quit IRC20:28
madhu_fnaval: how to run client tests?20:29
*** bharath has joined #openstack-lbaas20:29
fnavalthe clients are used in teh actual tests20:30
fnavalthey're used to make calls to the api20:30
madhu_ah, I see..thanks for correcting20:30
fnavalyep - btw, it's time20:30
juliancashfnaval - Are you up for diving into google hangout quickly on Mondays?  We'll keep it short.20:31
fnavalah yes20:32
fnaval1 sec20:32
fnavalwhats the room20:32
juliancashThanks!  :-)20:32
fnavalk thank20:32
*** sbfox has quit IRC20:45
*** sbfox has joined #openstack-lbaas20:50
*** apuimedo has joined #openstack-lbaas20:59
openstackgerritDoug Wiegley proposed openstack/neutron-lbaas: Empty hooks for testing experimental lbaas job
*** vivek-ebay has joined #openstack-lbaas21:04
*** apuimedo has quit IRC21:05
*** jorgem has quit IRC21:06
crc32is brandon around?21:08
*** apuimedo has joined #openstack-lbaas21:19
*** ajmiller has quit IRC21:28
openstackgerritmin wang proposed stackforge/octavia: change specs for amphora manager interface
openstackgerritBrandon Logan proposed openstack/neutron-lbaas: Neutron LBaaS: Initial Load Balancer Tests
openstackgerritBrandon Logan proposed openstack/neutron-lbaas: Added tox tempest environment
*** sbfox has quit IRC21:43
dougwigjuliancash: sorry i missed the hangout. anything that needs my attention or status or whatnot?21:45
*** vivek-ebay has quit IRC21:45
*** vivek-ebay has joined #openstack-lbaas21:45
*** vivek-ebay has quit IRC21:45
*** chlong has joined #openstack-lbaas21:45
*** vivek-ebay has joined #openstack-lbaas21:45
madhu_blogan: reviewed:
bloganmadhu_: did you cherry-pick or checkout that review?21:48
blogandid you run tox -e tempest?21:49
bloganand have devstack running? and have the correct tempest.conf?21:49
blogani just added a README21:49
bloganbout to push that up21:49
madhu_nope. I did run in this way in neutron_lbaas directory: nosetests -sv tests/tempest/v2/api/test_load_balancers.py21:49
madhu_btw the previous patch 21st set did work for me, checkedout 22nd patch and the ran the same I did for 21st set21:50
bloganyeah no more nosetests21:50
madhu_oh, why is that so?21:51
*** vivek-ebay has quit IRC21:51
*** vivek-eb_ has joined #openstack-lbaas21:51
blogantox -etempest will take care of pulling the right tempest down, and running the tests21:51
madhu_okay, can I see the readme.txt?21:51
blogangive me a second to push it up21:51
madhu_maybe I need to stop the devstack running, do some changes21:51
openstackgerritBrandon Logan proposed openstack/neutron-lbaas: Added tox tempest environment
openstackgerritBrandon Logan proposed openstack/neutron-lbaas: Neutron LBaaS: Initial Load Balancer Tests
*** chlong has quit IRC21:57
*** chlong has joined #openstack-lbaas21:57
*** vivek-eb_ has quit IRC22:00
*** vivek-ebay has joined #openstack-lbaas22:01
*** crc32 has quit IRC22:02
openstackgerritBrandon Logan proposed openstack/neutron-lbaas: Neutron LBaaS: Initial Listeners API tests
openstackgerritBrandon Logan proposed openstack/neutron-lbaas: Neutron LBaaS: Creating Pool API Tests
*** sbfox has joined #openstack-lbaas22:04
*** chlong has quit IRC22:04
openstackgerritBrandon Logan proposed openstack/neutron-lbaas: Neutron LBaaS: Adding test cases for Members.
openstackgerritBrandon Logan proposed openstack/neutron-lbaas: Neutron LBaaS: Creating Health Monitor API Tests
*** SumitNaiksatam has quit IRC22:11
bloganmadhu_: when you get a chance, could you do what that README says and let me know if it works or not for you?22:14
Aishblogan: I did a git checkout of 162772, and then tried running tox -etempest. but, it resulted in ERROR: InvocationError: '/bin/sh tools/ '22:19
bloganthat exists22:20
AishI could run the tests manually though..22:20
bloganhow so manually?22:20
Aishusing nosetests22:21
bloganwith these new reviews?22:21
AishI just tried the initial listeners test with the new patch. And it ran fine22:23
Aishusing nosetests22:23
*** openstack has joined #openstack-lbaas22:24
AishOh I am sry, i have tried with patch 11 not 12.22:27
Aishblogan: ^^22:27
madhu_blogan: will check that out22:28
madhu_Just reviewed22:30
openstackgerritBrandon Logan proposed openstack/neutron-lbaas: Neutron LBaaS: Load Balancer Basic Scenario Test
bloganmadhu_: it got added in that review you commented on22:31
madhu_oh, I see you changed the dev_tempconf to tempest.conf22:32
blogandidnt see the point in having it called dev22:33
bloganwould just be confusing22:33
madhu_ah I will recheck again and let you know then22:34
madhu_yes, its confusin to have that dev tempest config file22:35
bloganif i can just find an environment variable to point tempest to read a config from, we wont have to copy that over22:36
madhu_blogan: I have a question. Could you let me know what does tusker means?22:37
madhu_Is that a service?22:37
bloganlol that config file is the first ive ever heard of it too22:38
madhu_ha ha22:38
bloganah tripleo22:38
madhu_oh nice, thanks for the link22:39
openstackgerritmin wang proposed stackforge/octavia: change specs for amphora manager interface
*** apuimedo has quit IRC22:44
madhu_blogan: running tox taking longer than expected.. Is that so?22:45
bloganmadhu_: listener tests take me 5 seconds22:45
madhu_let me revisit22:46
*** bharath has quit IRC22:46
madhu_blogan: wow.. nicely put this after running the tests:   tempest: commands succeeded22:48
madhu_  congratulations :)22:48
madhu_it working for us22:48
bloganah good22:49
blogantox automatically puts that in :)22:49
Aishblogan: the initial listeners test ran fine with tox. :)22:59
openstackgerritGerman Eichberger proposed stackforge/octavia: Implements the haproxy amphora agent api server
madhu_blogan: I think we need to include "public_network_id = bb9e0fc9-0144-41a3-b17d-a63f829a1aba" in tempest.conf for scenarios tests to be successful23:21
bloganmadhu_: more complex than that, that id will be different every time you run devstack23:22
bloganmadhu_: when you run devstack with tempest, the tempest.conf is automatically built each time23:22
madhu_exactly.. maybe assigning a variable to neutron net-list and use awk for printing the first first columns and second row23:22
madhu_yeah, I see, even that works too23:23
bloganwill probably have to do that, but that should go in with the scenario review23:23
*** openstack has joined #openstack-lbaas23:24
madhu_let me know whether I can do the change or just wait for the changes to happen23:25
bloganwell i wish there was just an environment variable i could point to to say this is where you can find the tempest config23:25
bloganbut doesn't appear to be23:26
*** woodster_ has quit IRC23:30
bloganmadhu_, Aish: would yall say those tests (other than scenario) are good to go?23:31
madhu_it is good to go!23:31
madhu_I am bout to type this23:32
blogangreat, just need it to get through the large check queue now and it can be merged23:32
*** bharath has joined #openstack-lbaas23:33
*** chlong has joined #openstack-lbaas23:43

Generated by 2.14.0 by Marius Gedminas - find it at!