Friday, 2017-04-28

*** xgerman has joined #openstack-lbaas00:14
*** JudeC has quit IRC00:52
*** fnaval has joined #openstack-lbaas01:05
*** fnaval has quit IRC01:07
*** fnaval has joined #openstack-lbaas01:08
*** armax has quit IRC01:17
*** blogan_ has joined #openstack-lbaas01:53
*** blogan has quit IRC01:56
*** armax has joined #openstack-lbaas02:45
*** JudeC has joined #openstack-lbaas02:46
*** reedip has joined #openstack-lbaas03:12
*** armax has quit IRC03:18
openstackgerritcheng proposed openstack/octavia master: Fix set cookie_name to None if not in PUT content
*** aojea has joined #openstack-lbaas03:38
*** aojea has quit IRC03:45
*** fnaval has quit IRC04:02
*** amotoki has joined #openstack-lbaas04:14
openstackgerritJoshua Harlow proposed openstack/octavia master: Shutdown a endpoint workers executor on termination
*** armax has joined #openstack-lbaas04:48
*** reedip has quit IRC04:48
*** blogan_ has quit IRC05:07
openstackgerritYAMAMOTO Takashi proposed openstack/neutron-lbaas master: Remove get_namespace from API extensions
*** yuanying_ has quit IRC05:57
*** voelzmo has joined #openstack-lbaas06:24
*** sticker has quit IRC06:24
*** yuanying has joined #openstack-lbaas06:36
*** pcaruana has joined #openstack-lbaas06:36
*** gcheresh_ has joined #openstack-lbaas06:38
*** yuanying_ has joined #openstack-lbaas06:42
*** yuanying has quit IRC06:42
*** yuanying_ has quit IRC06:46
*** yuanying has joined #openstack-lbaas06:47
*** tesseract has joined #openstack-lbaas07:05
*** aojea has joined #openstack-lbaas07:18
openstackgerritJude Cross proposed openstack/python-octaviaclient master: [WIP] Add listener commands
*** yamamoto has quit IRC07:40
*** armax has quit IRC07:55
*** aojea has quit IRC08:08
*** crazik has joined #openstack-lbaas08:31
openstackgerritJude Cross proposed openstack/python-octaviaclient master: [WIP] Add listener commands
openstackgerritJude Cross proposed openstack/python-octaviaclient master: [WIP] Add pool commands
crazikhi, I have some issues with communication neutron-> octavia08:33
crazikgetting 401 on lb create08:33
crazikbut have no idea which enpoint it's trying to reach08:33
crazikit's not octavia-api, looks like keystone08:34
crazikis there any method to debug this more? (I already have debug turned on)08:34
crazikreturned info: Driver error: The request you have made requires authentication. (HTTP 401)08:34
*** JudeC has quit IRC08:36
crazikmaybe is there any running configuration example other than devstack?08:38
*** ipsecguy_ is now known as ipsecguy08:45
*** aojea has joined #openstack-lbaas08:59
*** gcheresh has joined #openstack-lbaas09:17
*** gcheresh_ has quit IRC09:18
*** tesseract has quit IRC10:04
*** tesseract has joined #openstack-lbaas10:08
rm_workjohnsom: ah yeah would be good to merge if you had that test :P10:30
*** yamamoto has joined #openstack-lbaas11:05
*** yamamoto_ has joined #openstack-lbaas11:06
*** gcheresh has quit IRC11:10
*** yamamoto has quit IRC11:10
*** yamamoto_ has quit IRC11:18
*** csomerville has joined #openstack-lbaas11:40
*** cody-somerville has quit IRC11:40
*** tesseract has quit IRC11:50
*** aojea has quit IRC12:01
*** tesseract has joined #openstack-lbaas12:05
*** coreycb has quit IRC12:09
*** yamamoto has joined #openstack-lbaas12:15
*** yamamoto has quit IRC12:15
*** yamamoto has joined #openstack-lbaas12:18
*** chlong has quit IRC12:21
*** yamamoto has quit IRC12:47
*** catintheroof has joined #openstack-lbaas13:01
*** catintheroof has quit IRC13:04
*** catintheroof has joined #openstack-lbaas13:04
rm_workthe octaviaclient-devstack thing isn't working in the gate, still not sure exactly what is wrong13:08
rm_worksince it work in barbican13:08
rm_workand it *works* outside the gate13:08
xgermanmmh — neither am I13:15
*** catinthe_ has joined #openstack-lbaas13:18
*** catintheroof has quit IRC13:21
rm_workaaah got it13:25
rm_worknot my problems13:25
rm_workone sec13:25
rm_workneed that13:28
rm_workand this13:29
openstackgerritAdam Harwell proposed openstack/python-octaviaclient master: Version name correction to allow installing
rm_workxgerman: as soon as you can +2 that, should be good13:30
xgermanwell, I can +2 right now ;-)13:30
rm_workit's been tested by the LB patch13:30
rm_workI just need it sooner as a gatefix13:30
xgermanthough the gate fix I can only +!13:31
rm_workI stole it from here13:31
rm_workthey should be able to do that quickly though13:31
rm_workusually they are responsive13:31
johnsomrm_work: why do we need the openstack client for these test gates?  Just thinking about it more I am not sure why it is needed13:35
rm_workso, i'd like it to be in the devstack plugin13:36
*** aojea has joined #openstack-lbaas13:36
rm_workand it can't be13:36
rm_workunless it's configured in the gates13:36
rm_workbecause the gates just run the devstack plugin13:36
rm_workso basically, when you install octavia with devstack, you should also get our client13:36
rm_workbarbican does exactly this13:37
rm_workjohnsom: :)13:37
rm_workjohnsom: if that makes sense to you, needs more +1 :)13:38
johnsomAh, so just have it there...  hmm, I thought you could just do that via localrc13:38
rm_workLIBS_FROM_GIT only works for the native stuff13:38
rm_workor stuff that is installed by other means13:38
rm_workand you can't enable_plugin a client because it doesn't have a devstack plugin13:38
rm_worki think this is a common pattern13:39
johnsomHmm, ok.  I will poke that when I get in the office13:39
rm_workalright, so when you walk 20 feet :P13:39
rm_workor are you getting coffee13:40
rm_workmy bed is my office so it's a bit easy for me13:40
johnsomAlso, could you poke the tempest plugin repo cookiecutter patch?13:40
johnsomIt's still 6:40am,  I don't trust myself to review yet.  Grin13:40
rm_worktrue why are you up13:41
johnsomGood question13:41
rm_worki should be asleep but i'm dumb13:41
rm_workmaybe same reason? :P13:41
rm_workmight be merged by the time you get to your office ^_^13:44
*** gcheresh has joined #openstack-lbaas13:46
*** yamamoto has joined #openstack-lbaas13:48
rm_workjohnsom: and +A plz :)13:52
*** chlong has joined #openstack-lbaas13:53
rm_workjohnsom: did we say the commands were going to all be subcommands of loadbalancer? so "openstack loadbalancer listener list"13:54
johnsomAll except quota which is centralized13:55
johnsomI put in pretty descriptive bugs with the tag octavia-osc if you want to take a look13:56
johnsomOh, hmm, in my rush I might have entered those wrong...  hmmm13:57
rm_worklol no clue but commented on Jude's patches13:57
johnsomI have to check the etherpad from the PTG13:57
rm_workpretty sure it's required13:58
rm_workesp since pool is failing the osc-dupe check13:58
rm_workbecause some other project already has it13:58
rm_workremember i was complaining about how long the commands were going to be :P13:58
johnsomYeah, I think I put that in the bug wrong (though I haven't done pool bugs yet)13:58
rm_worktabcomplete doesn't seem to be working for me yet either >_<13:58
rm_workwell, listener as well13:58
rm_workcommented on both patches13:59
johnsomOk.  Tab complete should work.  We need to look into that13:59
rm_workmight just be my issue13:59
rm_workyeah something is wrong on my VM14:00
rm_workdoesn't work for anything14:00
rm_work(i mean, anything inside "openstack"14:00
rm_workthe client needs to register something with BASH?14:00
xgermantab completion needs sometging14:01
rm_workit works if I just run "openstack" and use the prompt14:01
rm_workso yeah it's just something i'm missing prolly14:01
johnsomShould just be a readline library14:01
rm_workoh, no, readline would be at the openstack> prompt14:01
rm_worki mean just *in bash*14:01
rm_workthe neutron client registers some autocomplete magic14:02
rm_workso I can just opens<tab> lb<tab>lo<tab>c<tab>14:02
johnsomYeah, but I thought that had the hook for bash.  Ah I don't know.  Still coffee free and not in the office yet.  Grin14:02
rm_worknot opens<tab> :P14:02
rm_workwell, i'm gonna try to nap, but there's a couple things awaiting your +A :)14:03
*** chlong has quit IRC14:07
*** yamamoto has quit IRC14:16
*** chlong has joined #openstack-lbaas14:20
*** armax has joined #openstack-lbaas14:33
*** voelzmo has quit IRC15:02
*** gcheresh has quit IRC15:03
*** openstackgerrit has quit IRC15:17
*** pcaruana has quit IRC15:28
*** chlong has quit IRC15:34
*** fnaval has joined #openstack-lbaas15:40
*** aojea has quit IRC16:06
*** aojea has joined #openstack-lbaas16:07
*** aojea has quit IRC16:11
*** openstackgerrit has joined #openstack-lbaas16:16
openstackgerritMerged openstack/python-octaviaclient master: Version name correction to allow installing
*** chlong has joined #openstack-lbaas16:16
*** aojea has joined #openstack-lbaas16:27
*** tesseract has quit IRC16:45
*** gcheresh has joined #openstack-lbaas16:48
*** gcheresh has quit IRC16:59
*** harlowja has quit IRC17:06
*** aojea has quit IRC17:08
openstackgerritAdam Harwell proposed openstack/octavia master: Set up octaviaclient as part of devstack
*** blogan has joined #openstack-lbaas17:16
rm_workjohnsom: ok, so looks like after talking to clarkb in infra, i picked the wrong project to model after for clientlib17:24
rm_workswitched to the pattern Magnum uses, and it does use the LIBS_FROM_GIT17:24
johnsomAh, so I was right, we don't need the project entries?17:25
*** amotoki has quit IRC17:25
rm_workcorrect, LIBS_FROM_GIT does add them17:25
rm_workso now I'm tempted to go redo Barbican's impl17:26
rm_workthe problem was I needed pieces of both approaches17:26
rm_workI tried one, then the other17:27
rm_workanyway, testing this now17:27
*** aojea has joined #openstack-lbaas17:37
*** aojea_ has joined #openstack-lbaas17:38
*** aojea has quit IRC17:41
rm_workyep works, sweet17:41
*** aojea_ has quit IRC17:43
*** catintheroof has joined #openstack-lbaas18:04
*** catinthe_ has quit IRC18:08
*** gcheresh has joined #openstack-lbaas18:15
openstackgerritAdam Harwell proposed openstack/octavia master: Set up octaviaclient as part of devstack
*** voelzmo has joined #openstack-lbaas18:23
*** harlowja has joined #openstack-lbaas18:38
*** KeithMnemonic1 has joined #openstack-lbaas18:40
*** JudeC has joined #openstack-lbaas18:40
*** KeithMnemonic has quit IRC18:41
rm_workwtf happened below here:18:43
rm_workhow does AmphoraePostVIPPlug get a 404 on the amp18:43
johnsomSigh, this has been bugging me too18:44
johnsomI don't know if it is a gunicorn startup thing or some kind of address overlap issue18:45
johnsomI put in the second retry, but I don't want to go farther as 404 is valid for other calls18:45
*** crazik has left #openstack-lbaas18:49
rm_worki mean19:12
rm_workthat isn't the first call19:12
rm_workthe LB went active19:13
rm_workwhich means the rest API was already reachable19:13
rm_workI really should finish removing flask from this...19:16
rm_workI just didn't know how to access the request object without it19:16
rm_worki wonder if that could be related -- would it get a 404 if something broke in one of the handler functions?19:17
rm_workrandom scenario breakage is already starting to be annoying with these jobs voting T_T19:25
*** voelzmo has quit IRC19:33
openstackgerritJude Cross proposed openstack/python-octaviaclient master: [WIP] Add rest of load balancer commands
johnsomrm_work actually no, this is the very first contact with the REST API on the amphora-agent20:45
johnsomI think I messed up those listener bugs21:10
JudeCYeah I was discussing it with rm_work21:11
johnsomYeah, he noticed my mistake.  Sorry about that21:11
JudeCthey need to be  "loadbalancer listener create --options" right?21:12
johnsomI will update them now21:12
johnsomI will try to fill out the rest over the weekend21:12
JudeCkk when ever you get to it, not critical I am already planning on making the changes21:12
JudeCI plan to work on this stuff over the weekend also :)21:12
johnsomOk, I will be around if you have questions21:13
*** catintheroof has quit IRC21:14
johnsomOk, I am going to look at the multi-node gate issue21:16
*** harlowja has quit IRC21:17
*** harlowja has joined #openstack-lbaas21:17
*** aojea has joined #openstack-lbaas21:21
*** cody-somerville has joined #openstack-lbaas21:27
*** cody-somerville has quit IRC21:27
*** cody-somerville has joined #openstack-lbaas21:27
*** csomerville has quit IRC21:29
*** gcheresh has quit IRC21:29
JudeCjohnsom: What should I do about the unimplemented --cascade option. I am not sure how the API consumes the cascade action so I dont know if I should just stub out the option or just leave it out for now and implement it as soon as the API supports it.21:29
johnsomI think the plan is to have a URL parameter for that, but for now do whatever makes sense to you (stub/leave out)21:30
*** aojea has quit IRC21:35
*** harlowja has quit IRC21:36
*** chlong has quit IRC21:39
openstackgerritMerged openstack/octavia master: Fixes Octavia not working with Barbican
xgermanthere was also a thought to have all deletes be cascade ;-)21:41
openstackgerritJason Niesz proposed openstack/octavia master: blueprint: l3-active-active
*** blogan_ has joined #openstack-lbaas21:59
*** blogan has quit IRC22:01
openstackgerritJude Cross proposed openstack/python-octaviaclient master: Add loadbalancer commands to client
xgermanok, I am out for now22:16
johnsomHave a good weekend22:16
*** catintheroof has joined #openstack-lbaas22:17
*** blogan__ has joined #openstack-lbaas22:24
*** blogan_ has quit IRC22:26
openstackgerritJude Cross proposed openstack/python-octaviaclient master: Add loadbalancer commands to client
JudeCrequirements >.>22:37
JudeCmade a mistake with requirements...22:37
johnsomWell, I figured out why our wsgi wasn't working right with multinode, fixing that22:39
openstackgerritMichael Johnson proposed openstack/octavia master: Run Octavia API in a WSGI server
*** fnaval has quit IRC22:55
rm_workjohnsom: oh, cool22:58
rm_worki stared at that for a while to no avail22:58
johnsomYeah, me too22:58
johnsomThen I figured it out22:58
johnsomPort conflict22:58
rm_workJudeC: my recommendation is what xgerman said -- all deletes are cascade22:58
johnsomReally, I thought you un-did the automatic cascade delete22:59
rm_worki did22:59
rm_workbecause it was not a real cascade delete22:59
rm_workand it was broken22:59
rm_workso i just "fixed" it22:59
johnsomSo I recommended we add an option to cascade22:59
rm_workbut I think everything should be a TRUE cascade delete22:59
rm_workit's there22:59
rm_workyou can pass a param22:59
johnsomAh, so it's implemented now?22:59
rm_workmy recommendation to jude is to just *make the client always do that*23:00
rm_workbecause non-cascade-delete is dumb23:00
johnsomWell, I spec'd out a --cascade option23:00
rm_workok, well23:00
rm_workhe can pass that23:00
rm_workbut a regular delete seems to possibly still be misbehaving23:00
rm_workI think it's that I fixed it in v123:00
rm_workbut it probably got reintroduced in v223:01
rm_workah actually might be thinking of v1 for the param too, checking23:01
johnsomNo, I just saw it's there23:01
johnsomHmm, not consistent on the objects though.  Pool doesn't have cascade in the api23:02
JudeCSo wait is cascade working in the API right now?23:03
johnsomYeah, it looks like LB is the only one23:03
JudeCoh cool ok I will add that in today then23:03
rm_workyes it's supposed to be an LB thing23:03
rm_workthough I don't actually see how you specify23:03
rm_work(in v2)23:03
JudeCits just a url param I am assuming right?23:03
rm_workyeah that's not exposed23:04
rm_workwe expose delete() below it23:04
rm_workbut don't take a param23:04
johnsomOh, darn, you are right23:04
rm_workso uhh23:04
johnsomSorry, I have had a long week23:04
rm_workwell the other thing is23:04
rm_workit still DOES cascade23:04
rm_workso obviously still broken23:04
rm_workbut i think it does it badly23:05
rm_workhonestly my opinion is just23:05
rm_workchange that False to True23:05
rm_workand call it a day23:05
rm_workthe cascade delete paths work23:05
rm_workah i guess really it's validation23:07
johnsomI am too tired to know if that is really the right answer.  The nagging in the back of my head is with shared pools23:08
rm_workI got it23:08
rm_workone sec23:08
johnsomBut at LB it seems like it doesn't matter23:08
openstackgerritMichael Johnson proposed openstack/octavia master: Add new PROXY protocol to lbaas pool
johnsom^^^ rebased23:09
rm_work"Cannot delete Load Balancer %s - it has children"23:09
rm_workI kinda want to make that say23:09
rm_work"Cannot delete Load Balancer %s - think of the children"23:09
JudeCI was literally in the middle of typing that23:09
*** catintheroof has quit IRC23:11
*** JudeC has quit IRC23:21
*** JudeC has joined #openstack-lbaas23:22
*** harlowja has joined #openstack-lbaas23:30
openstackgerritJoshua Harlow proposed openstack/octavia master: Shutdown a endpoint workers executor on termination
*** harlowja has quit IRC23:55
*** harlowja has joined #openstack-lbaas23:59

Generated by 2.14.0 by Marius Gedminas - find it at!