Wednesday, 2016-02-17

rm_worksbalukoff: are your L7 changes rebased on top of those other db model changes?00:02
rm_workit doesn't ... look like it? I don't know, i can't freaking tell anymore, the new gerrit layout is not clear about commit chains00:02
sbalukoffWell.. they're... mostly rebased. ;)00:02
rm_worklol k00:02
sbalukoffThe only difference is a commit message00:02
sbalukoff(Adding the Closes-Bug: tag)00:02
bloganreedip: ping00:03
rm_workwell when you fix that I can +2 that one... i wonder if that affects anything else tho00:03
sbalukoffrm_work: I'm looking. :p00:05
sbalukoffthat was a really good catch.00:05
rm_workthat's one of the kinds of things my brain flags immediately just scanning through00:06
johnsomrm_work Nova and Neutron client factories can return incorrect clients on second call00:07
openstackLaunchpad bug 1543269 in octavia "Nova and Neutron client factories can return incorrect clients on second call" [High,New]00:07
rm_workjohnsom: yeah?00:07
johnsomrm_work What is the impact of this?  Something we need to target for Mitaka?00:07
*** jorgem is now known as jorgem[away]00:07
rm_workprobably ASAP if anyone is using admin tenants for anything00:07
rm_workwhich I think we are internally...00:07
rm_workand probably most deployers will be...00:07
rm_workor rather, admin + non-admin00:07
rm_workIMO the quick fix is to just ... not cache00:08
rm_workthe more in-depth fix is to cache on a lookup dict based on a tuple00:08
bloganjohnsom: for Newton, we need to start trying to allow a better way to manage the amphorae than the single mgmt net00:08
rm_workhonestly it will probably be fine just not cacheing00:09
rm_workcan't have cache invalidation issues if you don't cache00:09
johnsomblogan Don't disagree.  Want to write up an RFE with a proposal?00:09
johnsomrm_work Ok, I am going to tag for mitaka for now00:09
bloganyeah i don't think we need to cache, if we're using the keystone session thats doing a good amount of caching for us, the ones we care about really00:09
rm_workso that's a super easy one if someone wants to pick it up00:10
bloganjohnsom: i have one idea but i may have more by the end of the week00:10
rm_workanyone still looking for ATC? :P00:10
johnsomOk, cool.00:10
johnsomme, me, oh just kidding00:10
bloganjohnsom: actually, we can talk about it at tomorrows meeting00:11
blogangot an agenda up yet that i can add to?00:11
johnsomblogan Sure.  The agenda might be a bit full, but go ahead.  Yes, I posted one today00:11
bloganjohnsom: our meetings hardly ever go the full time, if i'm a betting man we'll have time00:12
johnsomThere is at least one shed involved00:12
rm_workI vote blue00:12
rm_workdamnit johnsom, i don't know if we can be friends anymore00:13
rm_workmaybe PURPLE! :P00:13
rm_worklol, these docstrings:
rm_worktotally avoiding the whole point of the pep8 requirements for docstrings00:14
rm_worksupposed to be enforcing a *summary line* followed by a blank line followed by details00:15
rm_worknot ... "whelp it's forcing me to make a new line with a blank"00:15
johnsomRight.  It is consistent with the above _transform_pool though00:15
rm_workbut T_T00:15
rm_worki'm not going to -1, but i am going to be sad00:16
johnsomI want the crack of a ruler on someone's wrist too00:17
rm_workah he also forgot to add "Bluebox, an IBM company" to a bunch of copyright headers :P00:19
* rm_work trolls00:19
* johnsom throws the eraser at rm_work00:20
bloganwe should put a specific hacking check for that00:21
bloganjohnsom: added 2 items to the agenda, if we dont get to htem, we dont get to them00:21
blogani actually have another thing to talk about too, which kind of ties into all of this00:22
xgermanI really hate seeing two (c_ lines00:22
bloganwhat the hell is (c_ lines?00:22
johnsomblogan Add away00:22
xgermanyep (c_00:22
* xgerman can’t type00:22
xgermanbut we had that discussion...00:23
bloganjohnsom: before i add that 3rd one, mind if we talk about it here fist?00:24
johnsomSure.  I added another option too00:25
blogani can totally see it being either 1) i'm a moron and not thinking correctly or 2) a real problem00:25
* johnsom thinks that is too easy....00:25
bloganso we have the possibility to have many controller workers00:25
bloganhow do we get those controller workers on the mgmt net wihtout tieing ourselves to a specific neutron plugin?00:26
xgermanassume they can recieve routed traffic from there?00:26
xgermanhow is an implementation detail00:27
johnsomNot following on the tying to neutron plugin00:27
bloganxgerman: well the controller worker needs an ip on that mgmt net, so theoretically, we'd need to create a port for it00:27
xgermanyou only need an ip for the health stuff00:28
rm_workthat's ... an administration level thing, right?00:28
rm_worklike, the deployer would set up a box/vm/whatever and connect it up00:28
bloganjohnsom: kind of like right now, for single host, to get the heartbeats working i had to create an ovs interface that is on the mgmt network00:28
*** yuanying_ has joined #openstack-lbaas00:28
rm_workand then run the controller on it?00:28
xgerman(and even that’s debatable since you could just sue DNS, VIP, etc.)00:28
xgermanblogan, that’s an implementation detail… say I have a provide rent, connect a box and install the controller on that00:29
johnsomYeah, whoever sets up the controllers will need to wire them up the mgmt lan they create for the conf00:29
bloganso we're saying deployers need to create those ports00:29
xgermanblogan, right00:29
johnsomYes sir00:29
bloganand then they just create an interface with that port mac address and get the ip (dhcp or static) and thats all they need to do correct?00:30
johnsomcables don't plug themselves in the real world, so....00:30
xgermanand if they like to use a UDP LB for the health messages more power to them00:30
bloganyes they do!00:30
bloganso i was going down the road of us having a mechanism that automatically creates all that for deployers00:31
*** yuanying has quit IRC00:31
rm_workAnsible? :P00:31
xgermanI think that’s what Octavia ansible is trying to do :-)00:31
johnsomfanatical support?00:32
blogani'd like to actually test this out to make sure it works00:32
xgermanwell, we are building a “reference architecture” internally00:33
bloganbut that leads into my other option for not having the single mgmt met, having a mgmt net per lb, which has its flaws, one being the controller would need to also plug itself into each mgmt network00:33
johnsomRight, not good for scale or HA00:33
bloganxgerman: and interally we're basically just using a provider network so its not the same00:33
xgermanyeah, I think the mgmt net issue is something we need to think about though we also use a provider net00:34
blogannope but we're not doing vrrp either, and scale isn't going to be a problem (if you were talking about us doing provider net)00:34
xgermanbut we can always give a lit of nets and pick one at random00:35
xgermanI think the original plan was to be able to configure N mgmt nets which all would ropiute to controllers00:35
xgermanI am pretty sure one mgmt net per Lb will cause other scale problems00:35
bloganxgerman: well i think the better option would to maybe start thinking about clusters of controllers00:36
bloganso that every controller won't need to be hooked into every mgmt net00:36
johnsomblogan my comments were about mgmt net per lb00:36
xgermanwell, if you use clusters, you can just have N complete Octavia insytalls00:36
bloganjohnsom: ah okay, why wouldn't that be good for hA then?00:36
xgermanlike 10 devtsacks ;-)00:36
bloganxgerman: lol i guess, but why have multiple apis?00:37
johnsomblogan you would have to plug those all into each and every controller.00:37
bloganthere'd be a scheduling need to decide what cluster should be used00:37
johnsomSo, what is wrong with a controller net that routes with amphora networks?00:37
bloganjohnsom: you mean the current setup?00:38
xgermanblogan you can always schedule with LBaaS V2; provider octavia1, octavia200:38
bloganxgerman: do flavors schedule or something?00:38
*** ajmiller_ has joined #openstack-lbaas00:38
xgermantoday you can schedule on different providers00:39
johnsomblogan more than one "amphora" network that has a route to a "controllers" network.  Plug controllers in controllers net, at boot time distribute amphora across the "amphora networks"00:39
bloganxgerman: yeah but there's not automatic scheduling, just what the user tells the api00:39
xgermanflavor just gives them a name — so instead of octavia1 you can say “gold"00:39
xgermanyep, wouldn’t be automatic00:39
johnsomHousekeeping could even add "amphora networks" as needed00:40
bloganjohnsom: are you using amphora network in a generic way to encapsulate mgmt nets and ha nets?00:40
bloganjohnsom: then whats an amphora network?00:41
*** ajmiller has quit IRC00:41
johnsomNew concept, network segments we plug amphora mgmt interface into, that via router has a path to "controllers network"00:42
johnsomPreferably IPv6, but that is probably just my mantra00:42
bloganah okay i think i see what you mean now, using routers we'd be able to have one controller network to many amp networks00:42
johnsomIntended to get around to many instances on one neutron network00:43
johnsomblogan Bingo00:43
bloganwell one network per controller00:43
xgermanwho is using more than 200 lbs?00:43
xgermanand wouldn’t neutron up that number anyway?00:43
johnsomWhy one network per controller?  You can plug a bunch of them in the same network00:43
bloganxgerman: since yall got out of the public cloud space, just us00:43
johnsomYeah, we had one customer with more than 200 lbs00:44
xgermanyeah, exactly...00:44
* johnsom thinks because it was in beta and cheap00:44
bloganjohnsom: yeah true about the controllers being on one network, i was not thinking that out correctly00:45
xgermanI just don’t want us to over-engineer that bit00:45
bloganwe'd never over-engineer anything, what do you mean???00:46
blogani kid00:46
johnsomYeah, this is a bit of frosting on the cake.  But if it's important (N)ow-ish, we (octavia team) could look at it.00:47
johnsomI thought you were just doing the provider network and calling it a day00:47
blogannah i'd like to just have a plan or at least options00:47
xgermanI could get behind a lit of mgmt networks you can configure00:47
xgermanand they all plug into the controller net...00:47
bloganjohnsom: we are but we also have a private cloud side that can't use that, and they bring up good points about the single mgmt net not being a good solution, which we all agreed in the beginning, we just needed something working00:48
xgermanbut spinning up clusters and then scheduling between them...00:48
bloganyeah i think the controller network -> router -> amp netowrks is a good idea, though we've tied it to the cloud its being deployed on having L300:49
*** yuanying_ has quit IRC00:49
bloganxgerman: i wasn't suggesting we do the clusters right away!00:49
xgermanok, glad we are on the same page ;-)00:49
bloganjust a possible end goal00:49
*** yuanying has joined #openstack-lbaas00:49
bloganbut this sounds like a better plan00:49
johnsomblogan This should be optional, so if no L3 routers you can still run like devstack or provider.  I think that should be do-able00:50
bloganjohnsom: agreed00:50
*** SumitNaiksatam has quit IRC00:50
xgermanwell, if the interface is a int of networks the  operator set up… it’s generic enough00:50
johnsomSide note, does anybody know what we should do with this?
openstackLaunchpad bug 1541986 in octavia " Implementing EventStreamer" [Undecided,New]00:51
johnsomIt seems like it just assigned it back to us....00:51
xgermanisn’t it done?00:51
rm_workdid the neutron piece merge?00:51
johnsomYeah, but look at the content of the bug00:51
rm_workit did, right?00:51
rm_workoh right yeah that00:51
johnsomDear bug triager. This bug was created since a commit was marked with DOCIMPACT.00:51
xgermanyeah, no docs00:52
* xgerman wags finger00:52
johnsomThose conf options are documented somewhere I think.  Should I just change the title on this?00:53
bloganah shit i forgot to take that DocImpact flag off since i made all the config options have a default00:53
bloganand i think if you default the config option you odn't need a DocImpact but i'm not entirely surea bout that00:54
johnsomblogan you want to close it?00:54
bloganjohnsom: do you know00:54
bloganall these rules i can't keep track of00:54
johnsomI'm not either.  Like I said, I think there is a doc with the options somewhere.00:54
bloganlet me do some research real quick00:54
johnsomThank you!00:55
*** yuanying has quit IRC00:59
bloganjohnsom: isn't that autogenerated?00:59
johnsomIt should be, but I'm pretty sure we updated it for liberty00:59
bloganwell do we have an octavia equivalent?01:00
johnsomoctavia.conf is at the bottom of that page01:00
bloganson of a, i see01:01
*** doug-fish has quit IRC01:01
*** doug-fish has joined #openstack-lbaas01:02
johnsomblogan, I am going to close this one invalid and open a "update docs for mitaka" bug01:03
bloganjohnsom: okay01:03
johnsomOr dup it01:03
bloganwhen do we get to hire our technical writer?01:03
johnsomI do however know who updated it for Liberty and we might be able to bribe her01:04
bloganbribery is the best policy01:04
*** yuanying has joined #openstack-lbaas01:06
johnsomOk, need to step away.  I tagged 16 bugs with mitaka that we can discuss tomorrow.01:07
openstackgerritFranklin Naval proposed openstack/octavia: Octavia: Basic LoadBalancer Scenario Test
*** allan_h has quit IRC01:09
*** fnaval has quit IRC01:09
*** allan_h has joined #openstack-lbaas01:11
*** ducttape_ has joined #openstack-lbaas01:16
*** ducttape_ has quit IRC01:25
*** ducttape_ has joined #openstack-lbaas01:26
*** ducttape_ has quit IRC01:30
*** _cjones_ has quit IRC01:31
*** bana_k has quit IRC01:36
*** minwang2 has quit IRC01:44
*** doug-fish has quit IRC01:44
*** doug-fish has joined #openstack-lbaas01:44
*** doug-fis_ has joined #openstack-lbaas01:48
*** kev0 has quit IRC01:49
*** doug-fish has quit IRC01:49
*** doug-fis_ has quit IRC01:52
*** bana_k has joined #openstack-lbaas01:54
*** madhu_ak has quit IRC01:57
*** Aish has quit IRC01:59
*** yamamoto_ has joined #openstack-lbaas02:06
*** yamamoto_ has quit IRC02:09
*** ducttape_ has joined #openstack-lbaas02:23
*** fnaval has joined #openstack-lbaas02:28
*** Bjoern has joined #openstack-lbaas02:32
*** doug-fish has joined #openstack-lbaas02:38
*** doug-fish has quit IRC02:39
*** doug-fish has joined #openstack-lbaas02:40
*** ajmiller_ has quit IRC02:44
*** Aish has joined #openstack-lbaas02:51
*** Aish has left #openstack-lbaas02:51
*** doug-fish has quit IRC02:51
*** ducttape_ has quit IRC02:52
*** Bjoern has quit IRC03:04
*** Bjoern has joined #openstack-lbaas03:05
*** Aish has joined #openstack-lbaas03:06
*** Aish has quit IRC03:06
*** Bjoern has quit IRC03:09
*** yamamoto has joined #openstack-lbaas03:10
*** ducttape_ has joined #openstack-lbaas03:15
*** SumitNaiksatam has joined #openstack-lbaas03:20
openstackgerritFranklin Naval proposed openstack/octavia: Octavia: Basic LoadBalancer Scenario Test
*** allan_h has quit IRC03:25
*** yuanying has quit IRC03:30
*** yuanying has joined #openstack-lbaas03:32
*** yuanying has quit IRC04:05
*** yuanying has joined #openstack-lbaas04:06
*** yuanying has quit IRC04:07
*** yuanying_ has joined #openstack-lbaas04:07
*** TrevorV has joined #openstack-lbaas04:18
*** ducttape_ has quit IRC04:30
*** ducttape_ has joined #openstack-lbaas04:33
*** links has joined #openstack-lbaas04:41
*** ducttape_ has quit IRC04:42
*** ducttape_ has joined #openstack-lbaas04:45
openstackgerritStephen Balukoff proposed openstack/octavia: Fix model update flows
*** ducttape_ has quit IRC04:52
*** TrevorV has quit IRC04:56
*** bana_k has quit IRC04:57
*** links has quit IRC04:57
*** neelashah has joined #openstack-lbaas04:58
openstackgerritStephen Balukoff proposed openstack/octavia: Add L7 database structures
openstackgerritStephen Balukoff proposed openstack/octavia: Update repos for L7 policy / methods
*** neelashah has quit IRC05:20
*** armax has quit IRC05:23
openstackgerritStephen Balukoff proposed openstack/octavia: Update repos for L7 rules / validations
openstackgerritStephen Balukoff proposed openstack/octavia: Add L7 api - policies
openstackgerritFranklin Naval proposed openstack/octavia: Octavia: Basic LoadBalancer Scenario Test
openstackgerritStephen Balukoff proposed openstack/octavia: Add L7 api - rules
openstackgerritStephen Balukoff proposed openstack/octavia: Add L7 controller worker flows and tasks
openstackgerritStephen Balukoff proposed openstack/octavia: Add L7 jinja template updates
openstackgerritStephen Balukoff proposed openstack/octavia: Add L7 documentation
sbalukoffjohnsom, blogan, rm_work: Ok! Fixes are in and rebased. The last patch in the chain is   Please run your smoke tests against that!06:12
sbalukoffIn the mean time, it looks like jenkins is randomly failing on tests in the chain. That's... annoying. I'mma look closer into this to see if it's something I did, or if it's jenkins acting up again.06:13
*** allan_h has joined #openstack-lbaas06:21 appears to be having issues... so packages aren't reliably downloading. And this is leading to the errors in Jenkins.06:22
sbalukoffYeah, all the errors I'm seeing here (except the ongoing session_persistence bug) seem to be caused by this.06:24
*** allan_h has quit IRC06:43
*** numans has joined #openstack-lbaas07:07
*** numans has quit IRC07:07
*** numans has joined #openstack-lbaas07:08
*** kobis has joined #openstack-lbaas07:21
*** bana_k has joined #openstack-lbaas07:31
*** kobis has quit IRC07:31
*** kobis has joined #openstack-lbaas07:37
*** kevo has joined #openstack-lbaas07:38
*** rcernin has joined #openstack-lbaas07:42
*** bana_k has quit IRC08:06
*** bochi-michael has joined #openstack-lbaas08:07
*** jwarendt has quit IRC08:25
*** kevo has quit IRC08:25
*** pcaruana has joined #openstack-lbaas08:42
*** jschwarz has joined #openstack-lbaas08:54
openstackgerritStephen Balukoff proposed openstack/octavia: Fix health monitor URL in API documentation
*** eranra has joined #openstack-lbaas09:16
*** paco20151113 has joined #openstack-lbaas09:32
*** barclaac has quit IRC09:52
*** ramishra has quit IRC09:52
*** ctracey has quit IRC09:52
*** dougwig has quit IRC09:52
*** _laco has quit IRC09:52
*** redrobot has quit IRC09:52
*** amitry has quit IRC09:52
*** xgerman has quit IRC09:52
*** blogan has quit IRC09:52
*** mestery has quit IRC09:52
*** redrobot has joined #openstack-lbaas09:52
*** barclaac has joined #openstack-lbaas09:52
*** blogan has joined #openstack-lbaas09:53
*** redrobot is now known as Guest5316209:53
*** mestery has joined #openstack-lbaas09:53
*** paco20151113 has quit IRC09:56
*** amitry has joined #openstack-lbaas10:00
*** xgerman has joined #openstack-lbaas10:01
*** ramishra has joined #openstack-lbaas10:01
*** ramishra has quit IRC10:03
*** jschwarz has quit IRC10:03
*** rcernin has quit IRC10:03
*** jschwarz has joined #openstack-lbaas10:03
*** rcernin has joined #openstack-lbaas10:04
*** bochi-michael has quit IRC10:05
*** dougwig has joined #openstack-lbaas10:06
*** _laco has joined #openstack-lbaas10:06
*** ctracey has joined #openstack-lbaas10:07
*** mestery has quit IRC10:08
*** SumitNaiksatam has quit IRC10:08
*** yamamoto has quit IRC10:08
*** cgross- has quit IRC10:08
*** clduser has quit IRC10:08
*** chlong has quit IRC10:08
*** ajo has quit IRC10:08
*** clduser has joined #openstack-lbaas10:08
*** cgross has joined #openstack-lbaas10:08
*** ajo has joined #openstack-lbaas10:08
*** mestery has joined #openstack-lbaas10:08
*** yamamoto has joined #openstack-lbaas10:08
*** chlong has joined #openstack-lbaas10:08
*** ramishra has joined #openstack-lbaas10:12
openstackgerritStephen Balukoff proposed openstack/octavia: Delete SSH amphora driver
openstackgerritStephen Balukoff proposed openstack/octavia: Delete SSH amphora driver
*** ctracey has quit IRC10:49
*** amitry has quit IRC10:49
*** hockeynut has quit IRC10:49
*** lmiccini has quit IRC10:49
*** logan- has quit IRC10:49
*** mhayden has quit IRC10:49
*** kfox1111 has quit IRC10:49
*** reedip has quit IRC10:49
*** hockeynut has joined #openstack-lbaas10:49
*** reedip has joined #openstack-lbaas10:50
*** logan- has joined #openstack-lbaas10:52
*** mariusv has quit IRC10:53
*** mhayden has joined #openstack-lbaas10:53
*** kfox1111 has joined #openstack-lbaas10:54
*** amitry has joined #openstack-lbaas10:56
*** lmiccini has joined #openstack-lbaas10:57
*** ctracey has joined #openstack-lbaas10:58
*** rcernin has quit IRC11:03
*** hockeynut has quit IRC11:04
*** ramishra has quit IRC11:04
*** xgerman has quit IRC11:04
*** blogan has quit IRC11:04
*** barclaac has quit IRC11:04
*** Guest53162 has quit IRC11:04
*** numans has quit IRC11:04
*** bedis_ has quit IRC11:04
*** prabampm has quit IRC11:04
*** thomasem has quit IRC11:04
*** kbyrne has quit IRC11:04
*** HenryG has quit IRC11:04
*** rm_work has quit IRC11:04
*** mugsie has quit IRC11:04
*** barclaac has joined #openstack-lbaas11:04
*** redrobot has joined #openstack-lbaas11:04
*** hockeynut has joined #openstack-lbaas11:04
*** bedis has joined #openstack-lbaas11:04
*** blogan has joined #openstack-lbaas11:04
*** redrobot is now known as Guest185611:04
*** thomasem has joined #openstack-lbaas11:04
*** mugsie has joined #openstack-lbaas11:05
*** prabampm has joined #openstack-lbaas11:05
*** numans has joined #openstack-lbaas11:05
*** HenryG has joined #openstack-lbaas11:05
*** rm_work has joined #openstack-lbaas11:06
*** kbyrne has joined #openstack-lbaas11:07
*** ramishra has joined #openstack-lbaas11:10
*** xgerman has joined #openstack-lbaas11:17
*** yamamoto has quit IRC11:32
*** kbyrne has quit IRC11:39
*** logan- has quit IRC11:39
*** ptoohill has quit IRC11:39
*** pck has quit IRC11:39
*** kbyrne has joined #openstack-lbaas11:40
*** ptoohill has joined #openstack-lbaas11:40
*** logan- has joined #openstack-lbaas11:42
*** pck has joined #openstack-lbaas11:43
*** thomasem has quit IRC11:45
*** thomasem has joined #openstack-lbaas11:46
*** rcernin has joined #openstack-lbaas11:49
*** HenryG has quit IRC11:50
*** clduser has quit IRC11:50
*** mdavidson has quit IRC11:50
*** pai15- has quit IRC11:50
*** crc32|znc has quit IRC11:50
*** clduser has joined #openstack-lbaas11:50
*** pai15 has joined #openstack-lbaas11:50
*** crc32|znc has joined #openstack-lbaas11:50
*** HenryG has joined #openstack-lbaas11:51
*** mdavidson has joined #openstack-lbaas11:53
*** doug-fish has joined #openstack-lbaas11:58
*** openstack has joined #openstack-lbaas12:05
*** johnsom has quit IRC12:07
*** pc_m has quit IRC12:07
*** pc_m has joined #openstack-lbaas12:07
*** johnsom has joined #openstack-lbaas12:08
*** piet has quit IRC12:12
*** piet has joined #openstack-lbaas12:13
*** amitry has joined #openstack-lbaas12:14
*** piet has quit IRC12:15
*** piet has joined #openstack-lbaas12:20
*** piet has quit IRC12:22
*** piet has joined #openstack-lbaas12:22
*** amitry has quit IRC12:24
*** numans has quit IRC12:24
*** ajo has quit IRC12:24
*** itsuugo has quit IRC12:24
*** markvan_ has quit IRC12:24
*** sbalukoff has quit IRC12:24
*** jorgem[away] has quit IRC12:24
*** jidar has quit IRC12:24
*** Kiall has quit IRC12:24
*** zigo has quit IRC12:24
*** markvan has joined #openstack-lbaas12:25
*** sbalukoff has joined #openstack-lbaas12:25
*** Kiall has joined #openstack-lbaas12:25
*** Kiall has quit IRC12:25
*** Kiall has joined #openstack-lbaas12:25
*** jorgem has joined #openstack-lbaas12:25
*** numans has joined #openstack-lbaas12:25
*** markvan has quit IRC12:25
*** markvan has joined #openstack-lbaas12:25
*** ajo has joined #openstack-lbaas12:25
*** numans has quit IRC12:25
*** numans has joined #openstack-lbaas12:25
*** mariusv has joined #openstack-lbaas12:25
*** ajo has quit IRC12:25
*** ajo has joined #openstack-lbaas12:25
*** itsuugo has joined #openstack-lbaas12:26
*** zigo has joined #openstack-lbaas12:26
*** mariusv is now known as Guest4522912:26
*** piet has quit IRC12:26
*** yamamoto has joined #openstack-lbaas12:27
*** jidar has joined #openstack-lbaas12:27
*** yamamoto_ has joined #openstack-lbaas12:29
*** yamamoto has quit IRC12:31
*** hockeynut has quit IRC12:32
*** eranra has quit IRC12:32
*** openstackgerrit_ has quit IRC12:32
*** openstackgerrit has quit IRC12:32
*** Dave has quit IRC12:32
*** davidlenwell has quit IRC12:32
*** eranra has joined #openstack-lbaas12:32
*** eranra has quit IRC12:32
*** eranra has joined #openstack-lbaas12:32
*** davidlenwell has joined #openstack-lbaas12:33
*** Dave has joined #openstack-lbaas12:33
*** yamamoto_ has quit IRC12:33
*** openstackgerrit has joined #openstack-lbaas12:34
*** openstackgerrit_ has joined #openstack-lbaas12:34
*** hockeynut has joined #openstack-lbaas12:36
*** amitry has joined #openstack-lbaas12:41
*** rtheis has joined #openstack-lbaas12:43
*** yamamoto has joined #openstack-lbaas12:50
*** barclaac has quit IRC12:51
*** dougwig has quit IRC12:53
*** barclaac has joined #openstack-lbaas12:53
*** piet has joined #openstack-lbaas12:59
*** alejandrito has joined #openstack-lbaas13:02
*** piet has quit IRC13:05
*** piet has joined #openstack-lbaas13:06
*** piet has quit IRC13:08
*** ducttape_ has joined #openstack-lbaas13:11
*** Kiall has quit IRC13:18
*** Kiall has joined #openstack-lbaas13:19
*** yamamoto has quit IRC13:19
*** yamamoto has joined #openstack-lbaas13:25
*** ducttape_ has quit IRC13:27
*** yamamoto has quit IRC13:38
*** yamamoto has joined #openstack-lbaas13:43
*** doug-fish has quit IRC13:44
*** yamamoto has quit IRC13:50
*** Kiall has quit IRC13:54
*** yamamoto has joined #openstack-lbaas13:57
*** Kiall has joined #openstack-lbaas13:59
*** neelashah has joined #openstack-lbaas13:59
mhaydeni drafted some docs for LBaaS v2 in the networking guide:
mhaydenif anyone has any cycles available to weigh in there, i'd appreciate it! :)14:05
*** dougwig has joined #openstack-lbaas14:11
*** jwarendt has joined #openstack-lbaas14:24
*** localloop127 has joined #openstack-lbaas14:26
*** doug-fish has joined #openstack-lbaas14:28
*** jwarendt has quit IRC14:29
*** dougwig has quit IRC14:33
*** yamamoto has quit IRC14:35
*** ducttape_ has joined #openstack-lbaas14:50
*** TrevorV has joined #openstack-lbaas14:55
*** piet has joined #openstack-lbaas15:01
*** Bjoern has joined #openstack-lbaas15:14
*** piet has quit IRC15:16
*** piet has joined #openstack-lbaas15:18
*** numans has quit IRC15:26
*** ajmiller has joined #openstack-lbaas15:26
*** neelashah has quit IRC15:27
*** neelashah has joined #openstack-lbaas15:30
*** piet has quit IRC15:52
*** piet has joined #openstack-lbaas15:53
*** piet has quit IRC16:06
*** _cjones_ has joined #openstack-lbaas16:08
*** _cjones_ has quit IRC16:08
*** piet has joined #openstack-lbaas16:12
*** fnaval has quit IRC16:13
TrevorVsbalukoff are you online right now?16:15
*** dougw2 has joined #openstack-lbaas16:15
*** piet has quit IRC16:17
*** piet has joined #openstack-lbaas16:17
*** dougwig has joined #openstack-lbaas16:18
*** dougw2 has quit IRC16:19
*** armax has joined #openstack-lbaas16:19
*** Guest46279 is now known as redrobot16:22
*** fnaval has joined #openstack-lbaas16:26
*** piet has quit IRC16:26
*** jwarendt has joined #openstack-lbaas16:27
*** pcaruana has quit IRC16:35
*** rm_work has quit IRC16:37
*** rm_work has joined #openstack-lbaas16:37
*** allan_h has joined #openstack-lbaas16:53
*** minwang2 has joined #openstack-lbaas17:09
*** madhu_ak has joined #openstack-lbaas17:18
rm_worksbalukoff: your l7 chain ready?17:22
*** localloo1 has joined #openstack-lbaas17:24
*** localloop127 has quit IRC17:26
*** jschwarz has quit IRC17:28
johnsomrm_work From a message last night he said it was, top of chain:
*** xgerman has quit IRC17:31
*** xgerman has joined #openstack-lbaas17:32
*** davidlenwell has quit IRC17:34
*** zigo has quit IRC17:34
*** hockeynut has quit IRC17:35
*** sbalukoff has quit IRC17:35
*** kevo has joined #openstack-lbaas17:35
*** dougw2 has joined #openstack-lbaas17:36
*** zigo has joined #openstack-lbaas17:36
*** davidlenwell has joined #openstack-lbaas17:37
*** hockeynut has joined #openstack-lbaas17:39
*** sbalukoff has joined #openstack-lbaas17:39
*** Aish has joined #openstack-lbaas17:47
sbalukoffHey folks!17:49
sbalukoffrm_work: Yep, L7 chain is ready.17:49
sbalukoffTrevorV: I'm online now.17:50
*** kobis has quit IRC17:50
johnsomPatching the libc fun on my devstack, then restacking with that L7 to kick the tires17:52
*** eranra has quit IRC17:56
*** localloo1 has quit IRC17:57
rm_workleaving libc bug in place, stacking with l7 patches17:59
johnsomWell, previous testing with shared pools left a bunch of ports/sec groups laying around and I was out of quota, so instead of bumping those I just decided to start fresh.18:00
rm_worki reformat every single run, so i don't care so much if someone hacks one of my random VMs lol18:01
mhaydenadded octavia to the docs patch if someone wants to verify my work ->
*** woodster_ has joined #openstack-lbaas18:08
*** bana_k has joined #openstack-lbaas18:10
johnsommhayden Excellent, thanks!18:11
mhaydenno problem18:11
sbalukoffmhayden: Having a look now!18:15
* mhayden woots!18:15
mhaydeni enjoy writing documentation, but i'm still wrapping my head around octavia18:16
mhaydenbut i plan on being in the meeting today -- i heard blogan was bringing punch and pie18:22
johnsomMmm, pie18:23
johnsomWell, L7 created the haproxy.cfg ok this time, so improvement!18:31
*** rcernin has quit IRC18:36
*** neelashah has quit IRC18:38
bloganmhayden: i'm bringing piss and vinegar18:40
blogani mean honey and sugar18:40
bloganalways get those mixed up18:40
mhaydenhmm, i guess everyone has their own tastes18:40
*** eranra has joined #openstack-lbaas18:40
sbalukoffmhayden: You have earned your first "-1" from me. :) Please note, this is how I show my love.18:43
johnsomsbalukoff Does putting the VIP on the public network under devstack now work?  There was some ACL funniness that caused it to not work previously.18:46
sbalukoffjohnsom: I'll give it a shot and let you know for sure. In theory it should work, but you're right that we might have a bug lurking in there which prevents it.18:47
johnsomIt was something strange about the way that network is created, outside of our stuff.18:47
sbalukoffAah! Ok, well, if it breaks in my test, I'll file and bug and get on it.18:47
johnsomWe just always said, you need floats if you want an IP on devstack's "public"18:48
sbalukoffI know that we have some people internally here that want Octavia to work without floating IPs, so it's going to be important from my perspective for that to work.18:48
johnsomIf you create your own external network it works fine18:48
sbalukoffjohnsom: Did you see my note about the listener stats interface bug?18:49
johnsomNope, can you give me a pointer?18:50
sbalukoffSure... hold on...18:50
*** localloo1 has joined #openstack-lbaas18:50
openstackLaunchpad bug 1543356 in octavia "Errors found for octavia api----get listener stats" [Medium,Confirmed] - Assigned to Stephen Balukoff (sbalukoff)18:50
*** eranra has quit IRC18:50
johnsomOk, sure, we can talk about changing it.  I will pull the sharedpools tag from that bug.  I bet the lbaasv2 sync puts those over in the neutron db and that is why the scenario tests/CLI work18:53
sbalukoffAah, ok.18:53
sbalukoffI just figured that if we've never had working stats before, it's a good time to revisit that and see if we want to alter how it works, since we wouldn't be breaking any existing installations per se.18:55
sbalukoffBut it sounds like Neutron LBaaS is a precedent that may have a pattern we want to follow there with Octavia.18:55
sbalukoffSo... it may be a short discussion. :)18:55
johnsomYeah, it has been working.  I guess no one noticed that Octavia API part wasn't18:55
mhaydensbalukoff: ah, okay -- i think i addressed some of your concerns18:55
*** doug-fis_ has joined #openstack-lbaas18:55
mhaydensome of that could be done in a follow-up patch by someone who knows octavia better than i do ;)18:56
*** localloop127 has joined #openstack-lbaas18:56
*** localloo1 has quit IRC18:59
sbalukoffmhayden: i'll have a look later today, eh!18:59
*** doug-fish has quit IRC18:59
*** allan_h has quit IRC18:59
*** doug-fis_ has quit IRC19:00
*** allan_h has joined #openstack-lbaas19:03
*** doug-fish has joined #openstack-lbaas19:03
*** neelashah has joined #openstack-lbaas19:04
*** Bjoern has quit IRC19:06
*** rcernin has joined #openstack-lbaas19:06
*** Bjoern has joined #openstack-lbaas19:07
*** doug-fish has quit IRC19:07
*** eranra has joined #openstack-lbaas19:08
*** doug-fish has joined #openstack-lbaas19:13
johnsomdougwig Can you kick this through?
dougwigjohnsom: aye19:18
johnsomThank you!19:18
TrevorVHey sbalukoff can I steal your attention for a minute19:22
*** eranra has quit IRC19:24
sbalukoffTrevorV attention thief!19:25
TrevorVThat's me19:25
sbalukoffYou have my attention. You bastard.19:26
TrevorVSo, I'm still, STILL working on the get-me-a-loadbalancer change, but I'm working on it such that it includes L719:26
TrevorVBasically, I need to rebase my 2 commits on top of your stack of changes, or in this case, to make things simple, I want to base my API change on top of your L7 doc review, and then my CW change on top of my API review19:26
TrevorVHowever, to do this, I need your entire stack to be up-to-date to complete that.19:27
TrevorVIS that the case right now?19:27
TrevorV(I'm speaking in redundant fashion because my thoughts aren't 100% complete right now....)19:27
sbalukoffYes-- that is to say, the first patch in the series is behind master by about 5 minor commits, but none of those conflict, so I was just going to leave it that way (unless something does conflict) so that people don't lose their place if they're actively reviewing and commenting.19:28
sbalukoffThe nice part about L7 is it's almost entirely additive. So conflicts ought to be rare.19:28
*** piet has joined #openstack-lbaas19:29
sbalukoffOf course, the longer it goes without being merged as other stuff gets merged in the mean time, the more likely it is that something will conflict. :/19:29
TrevorVRight.  For example, the single-create for octavia is ready... except it needs L7 ha ha19:30
TrevorVSo its not 100% ready19:30
TrevorVAnway, I'm pulling down up-to-date l7-documentation review, and putting my reviews on top.  Okay with you?19:30
sbalukoffWorks for me!19:30
TrevorVSee you in the dependency chain!19:31
TrevorVJust a note... seeing "Do you really want to submit the above commits?" as an option really really scares me, even though I literally just did "git review -d <l7 doc review number>" and then cherry-picked my review19:32
TrevorVsbalukoff ^^19:32
sbalukoffHaha! Indeed-- there are lot of things in that chain, eh.19:33
openstackgerritTrevor Vardeman proposed openstack/octavia: WIP - Get me a Load Balancer API
*** neelashah1 has joined #openstack-lbaas19:41
*** neelashah has quit IRC19:41
*** armax has quit IRC19:42
*** ducttape_ has quit IRC19:42
*** piet has quit IRC19:46
*** allan_h has quit IRC19:47
*** evgenyf has joined #openstack-lbaas19:47
openstackgerritTrevor Vardeman proposed openstack/octavia: WIP - Get Me A Load Balancer Controller
openstackgerritTrevor Vardeman proposed openstack/octavia: Update repos for L7 policy / methods
openstackgerritTrevor Vardeman proposed openstack/octavia: Add L7 controller worker flows and tasks
openstackgerritTrevor Vardeman proposed openstack/octavia: Add L7 api - rules
openstackgerritTrevor Vardeman proposed openstack/octavia: Add L7 database structures
openstackgerritTrevor Vardeman proposed openstack/octavia: WIP - Get me a Load Balancer API
openstackgerritTrevor Vardeman proposed openstack/octavia: Add L7 api - policies
openstackgerritTrevor Vardeman proposed openstack/octavia: Fix model update flows
openstackgerritTrevor Vardeman proposed openstack/octavia: Add L7 jinja template updates
openstackgerritTrevor Vardeman proposed openstack/octavia: Update data model conversion to be more complete
openstackgerritTrevor Vardeman proposed openstack/octavia: Add L7 documentation
openstackgerritTrevor Vardeman proposed openstack/octavia: Update repos for L7 rules / validations
TrevorVHOLY CRAP19:49
TrevorVI broke it.19:49
TrevorVSee, I knew something was going to break.19:49
TrevorVIdk what I broke, but I think something did.19:50
TrevorVDid it?  Or did it just output all the changes in the chain for some reason?19:50
rm_workglad i already stacked19:50
TrevorVNow I'm confused.19:51
rm_worknope if it output them it made a new patchset19:51
rm_workwhich means it did something19:51
rm_worksbalukoff: gl;hf19:51
bloganif you didnt make any changes to the previous reviews, you just pushed up old PSes19:51
TrevorVYeah... idk how though19:51
blogansometimes its unavoidable with long dependency chains, especially if htey're not all rebased on each other's new PS'es19:51
TrevorVI did git review -d L7-docs and cherry-picked my API review19:52
TrevorVThen did a git review -d api-review and cherry picked the cw review19:52
TrevorVIt shouldn't have broke anything....19:52
rm_workyeah so if that CR wasn't based on the newest of EVERYTHING below it19:52
rm_workit will overwrite the CRs below it with older versions of themselves19:52
rm_workso like, if you have commits A/B/C chained, and you do like19:53
*** allan_h has joined #openstack-lbaas19:53
rm_workthen commit B219:53
*** allan_h has quit IRC19:53
rm_workyou have A1->B2; B1->C119:53
rm_workthen if you commit C2 it'll do:19:53
sbalukoffEr... I think I can just un-do this by committing my latest l7-doc patch set?19:54
rm_workA1->B3->C2 where B3==B119:54
blogansbalukoff: yep19:54
sbalukoff(ie. the one from last night)19:54
rm_worksbalukoff: probably19:54
sbalukoffOk: TrevorV: Mind if I do that?19:54
TrevorVUm, no?19:54
TrevorVBut it has no changes19:54
rm_workyeah it should be fine19:54
rm_workit won't change his19:54
rm_workhe'll just have to rebase again later19:54
TrevorVSorry sbalukoff19:55
TrevorVIf you can just push the change up and it fix the stack, then that'd be great.19:55
TrevorVI'm hoping your l7 changes get merged soon anyway19:55
bloganTrevorV: looks like you have the CW review directly on top of the L7 reviews19:56
blogannope nvm19:56
bloganit just didn't push up a new API review19:56
sbalukoffSo, this is weird:
bloganjust do a git review -d L7-docs, cherry-pick api, cherry-pick controller, git review19:57
TrevorVwoah, what?19:57
sbalukoffYeah... we're about to go into a meeting anyway. I'll have a closer look afterward.19:57
bloganTrevorV: sure you didn't do a bunch of cherry-picks or a rebase that would change those commit hashes?19:58
TrevorVI mentioned what I did already19:58
TrevorVI can gist my history19:58
blogani know and if you ust cherry-picked on top of L7 that wouldn't change the L7 commit hashes, but they got changed19:58
TrevorVOkay, once again19:58
*** blallau has joined #openstack-lbaas19:59
TrevorVI did a review -d <l7-docs>19:59
TrevorVCherry pick API review19:59
TrevorVreview -d <api-review>19:59
TrevorVcherry pick cw review19:59
TrevorVand THAT broke it19:59
bloganand I am asking are you sure that is all you did?19:59
bloganoh wait19:59
johnsomOctavia meeting starting on #openstack-meeting-alt19:59
bloganoh wait after you did the cherry-pick of the API review youd didn't push it up?20:00
bloganbc review -d <api-review> would give you the old api review20:00
*** ducttape_ has joined #openstack-lbaas20:02
TrevorVblogan I added a few more lines.  The CW review needed a rebase, which made sense, but maybe was what caused the problem?  The gist above is the history20:04
bloganand thats what changed the hashes20:04
*** localloop127 has quit IRC20:05
TrevorVThat doesn't make sense to me.20:05
TrevorVRebasing CW made the preview reviews change?20:05
rm_workwas gonna say, the first one was fine, i cheered (internally) for it just working20:05
blogana rebase plays the commit back, but they're basically new commits20:05
bloganwhen did it tell you it needed a rebase?20:06
TrevorVWhen I did git-review20:06
*** localloop127 has joined #openstack-lbaas20:06
rm_workthat's what -R is for20:06
rm_worknow I remember20:07
TrevorVin my gist, line 530 it said "rebase", line 531 I identified the conflict, and then made the change.20:07
bloganyeah -R or --no-rebase would have solved that but usually you dont need to use those20:07
bloganbut it comes in handy with dependency chains sometimes20:07
rm_worki was trying to explain the use of -R earlier but it wasn't doing anything20:07
rm_workit's because it's for THIS problem, not the one i was trying to solve before20:07
TrevorVOnce again, I don't follow.20:08
TrevorVI was rebasing my CW review, why does it update everything else?20:08
TrevorVI've never seen that before20:08
blogani can't tell you why right now without seeing it happen honestly20:08
*** Bjoern has quit IRC20:10
TrevorVI see.  Just incredibly confusing to me.  API review went up, no problem.  CW review goes up, changes all over?  Just doesn't make sense20:10
bloganyeah sometimes you just dont know honestly, until you dive into and figure out why20:11
*** Bjoern has joined #openstack-lbaas20:14
*** Bjoern is now known as Bjoern_zZzZzZzZ20:14
*** localloop127 has quit IRC20:15
*** localloop127 has joined #openstack-lbaas20:17
rm_workthere's some workflow guidelines you can follow to try to keep it from ever happening though20:20
rm_workmy rules are: one commit at a time; git review -d <something>; git cherry-pick <something>; git review20:20
rm_workif you don't deviate AT ALL20:20
rm_work*usually* you will be ok20:20
rm_workbut literally any little change to that and boom20:21
rm_workeverything explodes and it's forensics time :P20:21
*** allan_h has joined #openstack-lbaas20:25
*** allan_h has quit IRC20:27
*** allan_h has joined #openstack-lbaas20:27
*** Bjoern_zZzZzZzZ is now known as Bjoern20:40
*** allan_h has quit IRC20:45
bloganmy rule is if i think something might explode, check my local commit hashes with the ones in gerrit to be certain20:53
TrevorVrm_work that's EXACTLY what I did... look at the gist I sent blogan.20:54
TrevorVJust to save you the time20:55
rm_workwhat is 528  git commit --amend20:57
rm_workthat remakes the commit hash for the commit above you20:57
rm_workthough it should have been your OWN commit20:57
rm_workso not a huge issue20:57
rm_workjust odd20:57
TrevorVOh sorry, that commit --amend was me checking a message20:58
rm_workI am guessing the git review on 530 didn't work?20:58
TrevorVI was looking for the topic in the commit message, and instead of typing "git log20:58
TrevorV" I was prepared to update it20:58
TrevorVha ha20:58
TrevorVAnyway, the 530 told me I had a rebase to do20:59
*** neelashah has joined #openstack-lbaas20:59
*** neelashah1 has quit IRC20:59
TrevorVhence me doing the status, and then rebasing20:59
rm_workonce it told you to do a rebase, something was f'd20:59
rm_worki'd have to see the full output of that git status20:59
rm_workand of the fetch+cherry-pick20:59
TrevorVOf course I don't have it, because I ran tox21:00
TrevorVwait, I do have it21:00
sbalukoffOk, I'mma try re-basing my commit chain to HEAD and re-upping.21:01
bloganok i gottag et lunch before my next meeting21:04
*** evgenyf has quit IRC21:06
TrevorVsorry again sbalukoff21:07
TrevorVI'm still trying to wrap my head around what went wrong...21:08
TrevorVI still don't see any problems :(21:08
rm_work mean i see right where it redid all the commits, but not sure WHY21:10
rm_workthough, you would have had some issues possibly because your --amend did change THAT commit hash21:10
rm_workwhich MAYBE could have made git-review look harder at rebasing? though it shouldn't have still21:11
TrevorVwhich again doesn't quite make sense rm_work21:11
TrevorVSince its only the top level commit hash that would have changed.21:11
TrevorVIdk, this was wonky.21:11
rm_workthe git-review plugin may look deeper in the chain if you commit two new things instead of one? not sure21:11
rm_workbasically, if it says anything to do with a rebase, ABORT MISSION EJECT EJECT21:12
TrevorVYeah, next time I'll do that.21:12
sbalukoffTrevorV: Ok, so it's not letting me push out a new patchset (it's saying the existing patchset is the same).  Glancing this over somewhat, I think that's the case.21:21
sbalukoffSo... I'm guessing your code base actually was up to date with master.21:21
neelashahjohnsonm: it looks like LBaaS v2 is not the default in M yet even though v1 was deprecated in L - is the plan to make v2 the default in M?21:21
TrevorVsbalukoff I'm not sure I follow.21:21
sbalukoffSo... I *think it's OK.21:21
sbalukoffTrevorV: I'm not sure I do either.21:21
TrevorVHa ha ha21:22
neelashahjohnsom ^21:22
TrevorVsbalukoff you're saying there are no changes to make, even though there was a complaint from gerrit originally?21:22
sbalukoffBut, I'll look at each of these patches on gerrit to see if there's any difference between the last two patschsets. thankfully, the web interface makes this somewhat easy.21:22
sbalukoffTrevorV: Yes.21:22
TrevorVsbalukoff I was doing that, I did see one file before I waited to see you update.21:22
sbalukoffTrevorV: At least, I *think* that's what I'm saying. ;)21:22
sbalukoffWhich file?21:22
TrevorVsbalukoff lemme verify21:23
johnsomneelashah Where do you see that LBaaSv2 is not the default?21:23
TrevorVconstants and controller_worker had changes21:23
TrevorVbetween the last two patch sets21:23
sbalukoffOk, I'll have a look.21:23
TrevorVconstants changed21:24
TrevorVconstants changed21:24
sbalukoffI think that's the result of a rebase from HEAD.21:25
TrevorVconstants changed.21:25
sbalukoffSo... it's not disruptive.21:25
sbalukoffSo, indvertently, you rebased all my patches to HEAD. XD21:25
TrevorVYeah, but I think it pushed a new patch set to everything....21:25
sbalukoffSo, not a bad thing, per se.21:25
TrevorVYeah, well, hopefully that doesn't break anything.21:25
sbalukoffNo worries. I'm not sure you did anything wrong.21:25
*** rtheis has quit IRC21:26
sbalukoffYeah, everything I'm seeing here is just the result of a rebase to HEAD. So... I think we're completely fine.21:28
TrevorVI hope...21:28
TrevorVNow I'll re-focus and double my efforts for the single-create.21:28
sbalukoffSounds good!21:29
TrevorVdougwig sorry for not having anything to report on the neutron lbaas side... :(21:29
*** rtheis has joined #openstack-lbaas21:29
neelashahjohnsom: nm, looks like I need to explicitly enable lbaas (whether v1 or v2) since its a separate plugin21:31
johnsomneelashah Yep, ok, no problem21:31
*** Guest45229 is now known as mariusv21:31
neelashahjohnsom but all the lbaas v2 tests are non voting at present, right?21:31
*** mariusv has quit IRC21:32
*** mariusv has joined #openstack-lbaas21:32
johnsomThe API tests are voting.  We are trying to track down a new issue that is causing some tests to fail, so we have made them non-voting21:32
sbalukoffIt's a Heisenbug-- seems to go away whenever we look closely. Which obviously makes troubleshooting it really annoying. :P21:33
neelashahjohnsom - oh I see, thanks21:34
johnsomYes, I noticed they have cleared up.  I'm wondering it it wasn't a bug in some other part of OpenStack that was impacting us21:34
sbalukoffIt could still be a timing issue. :/21:35
sbalukoffblogan: Can I get you to revisit your -1 on this:  :)21:37
johnsomI was just looking at that one myself21:37
*** armax has joined #openstack-lbaas21:41
xgermanblogan, johnsom:
openstackLaunchpad bug 1496628 in octavia "Amphora agent reload fails with socket in use" [High,Fix committed] - Assigned to German Eichberger (german-eichberger)21:46
johnsomYeah, looking21:46
xgermanwell, it works… so closed it21:48
johnsomxgerman Yeah, ok.  I can repro that it works too21:48
johnsomThanks for the followup21:48
xgermany. w.21:49
xgermanI would have rather spend that hour bike shedding why I think the L7 flow design is wrong ;-)21:50
openstackgerritStephen Balukoff proposed openstack/octavia: Delete SSH amphora driver
doug-fishHas anyone successfully created listeners using TERMINATED_HTTP protocol?21:52
doug-fishI'm getting this error: and can't make sense of it21:52
johnsomxgerman sadly, the service amphora-agent reload triggers that issue still21:52
doug-fishTERMINATED_HTTPS of course21:52
johnsomYeah, I demo'd that in Vancouver, so it worked at one time21:53
doug-fishjohnsom: what should the certificate parameter look like?21:53
rm_workdoug-fish: i just did one a second ago21:53
doug-fishI've tried the url and the id21:53
xgermanwell, if curl works… then the flow is broken?21:53
rm_workjust the ID21:53
doug-fishI'm going through neutron client ... rm_work that's good to know!21:54
johnsomSomething like this: neutron lbaas-listener-create --loadbalancer lb1 --protocol-port 443 --protocol TERMINATED_HTTPS --name listener1 --default-tls-container=$(barbican secret container list | awk '/ tls_container / {print $2}')21:54
rm_workdoug-fish: check out:
rm_workyou can see how i get the IDs and such21:55
rm_workand at the bottom are two commented commands that I run verbatim21:55
doug-fishrm_work: excellent! I'll take a look21:55
rm_workthat error though ... looks like maybe it can't communicate properly with Barbican?21:58
rm_workwhat user created the certificates in barbican, and what user are you trying to create the LB with?21:59
rm_workright now the ACLs don't work right in Barbican21:59
rm_work... and even if they did, you'd need another step to do that21:59
doug-fishrm_work: I created the certificate in Barbican using their user guide example22:06
doug-fishand I've verified that I get the same failure with a command line similar to yours. I think I must have the certificate wrong in Barbican22:07
doug-fishI'm new to barbican -- do you have any suggestions for checking my container?22:09
rm_worktry just making a new one using my commands22:10
rm_workand see if IT works22:10
rm_workthats a trash but valid cert22:10
doug-fishyeah that's a good idea22:10
*** blallau has quit IRC22:11
sbalukoffBy the way, for anyone smoke-testing the L7 stuff, this might come in handy:
johnsomHandy, thanks22:18
*** neelashah has quit IRC22:19
minwang2thank you sbalukoff22:21
sbalukoffNo problem. Obviously, please try out stuff I'm not doing in this as well...  :)22:22
TrevorVsbalukoff no way, we only want test results from happy path22:23
sbalukoff(For what it's worth, as I've been testing this, I've been closely watching the haproxy config that gets generated with each change.)22:23
sbalukoffTrevorV: Well, the happy path should work, too. ;)22:23
johnsomYeah, me too, I have a watch going on the file22:24
*** rtheis has quit IRC22:25
TrevorVsbalukoff just found a nit.22:26
TrevorVWasn't even reviewing :P22:26
rm_workyeah i've got happy-path for non-l7 pretty much all tested and looks good22:26
rm_workabout to do actual L722:26
TrevorVl7policy types.  L7PolicyResponse.  "action" is just a string, not an enum like you've marked in POST and PUT classes.22:27
TrevorVsbalukoff ^^22:27
sbalukoffTrevorV: That's actually just fine.22:27
sbalukoffWe don't really need to type-check our responses.22:27
TrevorVIf you insist :D22:28
sbalukoff(And we don't, elsewhere)22:28
TrevorVWell if no precedent, then I shut up :P22:28
*** diogogmt has joined #openstack-lbaas22:28
*** doug-fish has quit IRC22:33
*** doug-fish has joined #openstack-lbaas22:34
*** doug-fish has quit IRC22:34
*** doug-fish has joined #openstack-lbaas22:34
*** localloop127 has quit IRC22:40
*** alejandrito has quit IRC22:42
*** allan_h has joined #openstack-lbaas22:47
*** armax has quit IRC22:51
*** diogogmt has quit IRC22:54
doug-fishrm_work: I issued commands based on your script and I was able to successfully create a listener based with TERMINATED_HTTPS  -- thanks!22:54
doug-fishI'm wondering if we should be filtering the list of certificates a bit more in the UI22:54
*** diogogmt has joined #openstack-lbaas22:54
openstackgerritMadhusudhan Kandadai proposed openstack/octavia: Implement tempest-plugin for octavia
doug-fishso that we only show certicates from containers where there is a private key and a certificate22:55
madhu_akjohnsom, fnaval : ^^22:55
*** doug-fish has quit IRC22:57
*** doug-fish has joined #openstack-lbaas22:57
fnavalcool thanks madhu_ak - i'll take a look.22:57
*** diogogmt has quit IRC22:59
openstackgerritMadhusudhan Kandadai proposed openstack/octavia: Implement tempest-plugin for octavia
*** TrevorV has quit IRC22:59
*** allan_h has quit IRC22:59
madhu_akfnaval: if the above patch goes in, we just need to rebase
fnavaldoes running the tests stay the same?23:01
fnavalline 28 on might have been correct the previous way?23:02
*** doug-fish has quit IRC23:02
*** dougw2 has quit IRC23:02
madhu_akyes. the same way we run the tests23:03
*** allan_h has joined #openstack-lbaas23:03
madhu_akno confusion. The above plugin will be useful for setting up gate job in octavia23:03
fnavalok ok, what benefit does it provide?23:03
fnavalah, so it's for the gate job23:03
johnsommadhu_ak I see what fnaval is saying, should the return be in the for loop?23:03
fnavalmaking the tests themselves plug inable23:03
madhu_akyeah johnsom if I understand correctly, will we be having api tests for octavia in future?23:04
johnsomYes, we will have both tests, but return inside a for loop will only run once23:05
madhu_akah, I see. got you johnsom23:05
madhu_akwill resubmit and correct it23:06
openstackgerritMadhusudhan Kandadai proposed openstack/octavia: Implement tempest-plugin for octavia
*** allan_h has quit IRC23:11
*** Bjoern has quit IRC23:20
*** ducttape_ has quit IRC23:36
*** armax has joined #openstack-lbaas23:42
openstackgerritStephen Balukoff proposed openstack/octavia: Update data model conversion to be more complete
sbalukoffPlease note that the above commit (and the ones that follow) are literally just a comment update.23:56
sbalukoffTrevorV: I found that gerrit wouldn't let me create further dependent patches at the end of the chain because the first one failed to upload because it was exactly same as the one already up there.23:56
sbalukoffSo, I made an inconsequential change to the first in the chain, and will bubble the rebases through that chain so that I can add a new dependent patch.23:57
sbalukoff(The new dependent patch adds listener stats. :P )23:57
openstackgerritStephen Balukoff proposed openstack/octavia: Fix model update flows

Generated by 2.14.0 by Marius Gedminas - find it at!