Friday, 2017-06-09

*** hongbin has quit IRC00:44
openstackgerritFrederick F. Kautz IV proposed openstack/kuryr-kubernetes master: Add support for running kuryr with ocata devstack.  https://review.openstack.org/47058601:04
*** limao has joined #openstack-kuryr01:09
*** limao has quit IRC01:20
*** limao has joined #openstack-kuryr01:22
*** aojea has joined #openstack-kuryr01:23
*** yedongcan has joined #openstack-kuryr01:23
*** aojea has quit IRC01:27
*** hongbin has joined #openstack-kuryr02:23
*** limao has quit IRC02:24
*** limao has joined #openstack-kuryr02:25
*** hongbin has quit IRC02:27
*** hongbin has joined #openstack-kuryr02:28
*** yamamoto has joined #openstack-kuryr03:07
*** salv-orlando has joined #openstack-kuryr03:21
*** salv-orlando has quit IRC03:26
*** limao has quit IRC03:29
*** limao has joined #openstack-kuryr03:30
*** limao has quit IRC03:34
*** limao has joined #openstack-kuryr03:35
*** limao has quit IRC03:37
*** longfei_zhang_ has quit IRC04:04
*** hongbin has quit IRC04:11
*** salv-orlando has joined #openstack-kuryr04:22
*** salv-orlando has quit IRC04:26
*** janki has joined #openstack-kuryr04:42
*** yamamoto_ has joined #openstack-kuryr04:43
*** salv-orlando has joined #openstack-kuryr04:44
*** yamamoto has quit IRC04:47
*** limao has joined #openstack-kuryr05:04
*** limao has quit IRC06:07
*** limao has joined #openstack-kuryr06:08
*** limao has quit IRC06:10
*** limao has joined #openstack-kuryr06:10
*** ltomasbo|away is now known as ltomasbo06:35
*** pmannidi has quit IRC06:35
*** pmannidi has joined #openstack-kuryr06:35
*** salv-orlando has quit IRC06:52
*** salv-orlando has joined #openstack-kuryr06:52
*** pcaruana has joined #openstack-kuryr07:05
*** aojea has joined #openstack-kuryr07:31
*** yamamoto_ has quit IRC07:38
*** egonzalez has joined #openstack-kuryr07:41
*** yamamoto has joined #openstack-kuryr07:57
*** yamamoto has quit IRC07:59
*** yamamoto has joined #openstack-kuryr07:59
*** limao has quit IRC08:02
*** limao has joined #openstack-kuryr08:04
*** kzaitsev_ws has joined #openstack-kuryr08:13
apuimedoltomasbo: ping08:19
ltomasboapuimedo, pong08:19
apuimedoltomasbo: I'll review the rest of the patches in the series today. But we'll still need Irena to merge on Sunday08:20
ltomasbook08:20
ltomasboI think I addressed her comments08:21
ltomasboI'll be traveling next week (sunday-wednesday)08:21
ltomasboso, not sure if I will be able to commit modifications during those days, but I'll do my best..08:21
apuimedoltomasbo: leisure or business08:22
ltomasboCloud Control Conference08:22
ltomasboI'll talk about kuryr there!08:22
apuimedoawesome08:23
apuimedoltomasbo: link?08:23
ltomasbohttp://cloudresearch.org/workshops/11th/08:24
ltomasboit is a research gathering08:24
ltomasbowhere a couple of universities join together a lot of companies doing research in cloud computing08:24
ltomasboso, I'm on behalf of Red Hat, but there are intel, ibm, google, microsoft, ericsson, ... going too08:25
apuimedocool08:28
apuimedoltomasbo: escaping from the heat in Madrid, eh?08:28
ltomasbocool and cold! :D08:28
ltomasbojejeje08:29
ltomasboI've been leaving for 4 years in the north of Sweden, so I guess the heat will be still welcome for a while...08:29
openstackgerritMerged openstack/kuryr-kubernetes master: Add support for running kuryr with ocata devstack.  https://review.openstack.org/47058608:35
apuimedodo we have somebody working on https://bugs.launchpad.net/neutron/+bug/1696051 ?08:35
openstackLaunchpad bug 1696051 in neutron "Missing binding details for bulk port creation requests" [Undecided,New]08:35
ltomasboapuimedo, I was planning to take a look, but didn't get the time yet08:36
ltomasbothis only affect us for the generic (neutron) vif case08:36
apuimedoltomasbo: you should bribe one of the guys in Madid08:36
apuimedotell them if they fix it today. That's one fuet08:37
ltomasboxD08:37
ltomasboI'll try!08:37
ltomasbonow that I'll go to Sweden, I can threaten them with this: https://en.wikipedia.org/wiki/Surstr%C3%B6mming08:40
dmelladohi fellows08:42
dmelladoIm going to merge all of the gerrithub's tempest plugin patches08:42
dmelladoin order to create the pipy package08:42
dmelladoFYI08:42
* dmellado sighs08:43
apuimedoltomasbo: you should get some salmiak08:43
* apuimedo adores salmiak08:43
dmelladowhat's that?08:43
apuimedolicorice + ammonium chloride08:43
apuimedodmellado: ok, merge away08:43
dmelladoapuimedo: I didn't recall that it was so tedious to create a new openstack project08:44
dmelladootherwise I would've voted against that goal xD08:44
dmelladopatch to project.config08:44
dmelladopatch to governance08:44
dmelladozuul08:44
dmelladopipy08:44
* dmellado sighs xD08:44
*** garyloug has joined #openstack-kuryr08:45
apuimedodmellado: that's what you get with self service infrastructure08:45
apuimedoif we would just say "create a project, pretty please" it would be probably faster08:45
*** yamamoto_ has joined #openstack-kuryr08:46
*** yamamoto has quit IRC08:47
dmelladoapuimedo: I'll create a project08:48
dmelladoself-service fuet08:48
dmelladobe prepared08:48
dmelladoxD08:48
dmelladoltomasbo: wasn't in sweden where they had some kind of weird cigarrettes?08:49
ltomasboyep, that too08:49
ltomasbodo you want some?08:49
dmelladono, please!08:50
dmelladoif those are the ones I recall08:50
dmelladothey had some kind of weird 'fish' inside08:50
dmelladoxD08:50
* dmellado recalling his short time in Karlstad08:50
dmelladoxD08:50
dmellado'the sunny city' they said, I do defy them to dare coming to Madrid08:50
apuimedoxD08:54
apuimedoI wish I had some salmiak08:55
dmelladoI wish I would recall where does enigmail stores his pubring08:58
* dmellado has forgotten his pipy creds xD08:58
apuimedolol08:59
dmelladooh, it was just so painfully slow xD08:59
apuimedowant me to create the project?08:59
apuimedo(in pypi. no way I'm doing the rest)08:59
dmelladonah, no worries08:59
dmelladoI just got the ack email08:59
dmelladoit just took a couple of minutes08:59
dmelladoso I can log in back now08:59
apuimedoltomasbo: a lot of these tests you did, I'd rather move them to functional in the future09:01
dmelladoapuimedo: ho ho ho09:01
apuimedoa lot of things are just testing that helper methods were called and such09:01
dmelladoso now I have an excuse of getting ltomasbo to improve testing xD09:01
apuimedoand this sets my spidey sense09:01
dmelladoapuimedo: but which spiderman are you09:01
dmelladoI even saw an indian one09:01
ltomasboapuimedo, what tests are you talking about?09:01
apuimedoltomasbo: a lot of the ports pool ones09:02
apuimedoI see a lot of assert_called09:02
ltomasboahh, ok09:02
apuimedofor non API methods09:02
apuimedoit's fine to have it now since there's no functional testing09:02
ltomasboyep, it will be great to have funcional tests...09:02
apuimedobut in general goes against my religion09:02
dmelladoTODO(ltomasbo): move foo to FT09:03
dmelladoxD09:03
ltomasbo:D09:03
*** yamamoto_ has quit IRC09:03
apuimedoltomasbo: can you make a patch on the tip of the branch that changes GenericVIFPool into NeutronVIFPool09:05
apuimedo(for consistency)09:05
*** yamamoto has joined #openstack-kuryr09:07
*** neiljerram has joined #openstack-kuryr09:07
*** limao has quit IRC09:08
*** limao has joined #openstack-kuryr09:13
*** limao has quit IRC09:14
*** limao has joined #openstack-kuryr09:14
*** limao has quit IRC09:16
*** yamamoto has quit IRC09:16
*** ltomasbo is now known as ltomasbo|away09:20
*** ltomasbo|away is now known as ltomasbo09:22
*** neiljerram has quit IRC09:24
*** yamamoto has joined #openstack-kuryr09:37
*** yamamoto_ has joined #openstack-kuryr09:38
*** yamamoto has quit IRC09:38
*** yedongcan has quit IRC09:45
*** yedongcan has joined #openstack-kuryr09:46
*** neiljerram has joined #openstack-kuryr09:54
*** yedongcan1 has joined #openstack-kuryr10:01
*** yedongcan has quit IRC10:03
apuimedoltomasbo: your branch is approved10:05
apuimedonow you need to nag other cores10:05
ltomasboapuimedo, thanks!10:11
dmelladoapuimedo: ar yu suar?10:14
dmelladoxD10:14
apuimedodmellado: I hope so10:15
*** dougbtv has joined #openstack-kuryr10:19
*** neiljerram has quit IRC10:25
*** salv-orlando has quit IRC10:42
*** salv-orlando has joined #openstack-kuryr10:43
*** salv-orlando has quit IRC10:47
openstackgerritKirill Zaitsev proposed openstack/kuryr-kubernetes master: Enforce E128 pep8 check  https://review.openstack.org/47224910:48
*** rwallner has joined #openstack-kuryr11:07
*** rwallner has quit IRC11:25
*** rwallner has joined #openstack-kuryr11:29
*** yamamoto_ has quit IRC11:42
*** yamamoto has joined #openstack-kuryr11:43
*** aojea has quit IRC11:53
openstackgerritKirill Zaitsev proposed openstack/kuryr-kubernetes master: Update installation documentation  https://review.openstack.org/47265312:05
*** yedongcan1 has left #openstack-kuryr12:31
*** aojea has joined #openstack-kuryr12:54
*** limao has joined #openstack-kuryr12:57
*** aojea has quit IRC12:59
*** limao_ has joined #openstack-kuryr13:01
*** limao has quit IRC13:02
*** yamamoto has quit IRC13:04
*** yamamoto has joined #openstack-kuryr13:04
*** limao_ has quit IRC13:08
*** limao has joined #openstack-kuryr13:09
*** limao has quit IRC13:10
*** rwallner has quit IRC13:10
*** limao has joined #openstack-kuryr13:10
*** rwallner has joined #openstack-kuryr13:13
*** aojea has joined #openstack-kuryr13:22
*** limao has quit IRC13:24
*** limao has joined #openstack-kuryr13:25
*** aojea has quit IRC13:26
openstackgerritLuis Tomas Bolivar proposed openstack/kuryr-kubernetes master: Nested vlan vif pool driver extension to precreate reusable subports  https://review.openstack.org/43689413:29
ltomasboapuimedo, ping13:29
ltomasboI sent a new version of the last patch for the ports pool13:29
ltomasbocan you check it13:29
ltomasboit is just a small modification to ensure proper behavior after kuryr-controller reboot13:29
*** rwallner has quit IRC13:33
*** aojea has joined #openstack-kuryr13:33
*** limao has quit IRC13:36
*** limao has joined #openstack-kuryr13:37
*** janki has quit IRC13:47
*** limao has quit IRC13:54
*** limao has joined #openstack-kuryr13:55
*** limao has quit IRC14:06
*** limao has joined #openstack-kuryr14:07
*** limao has quit IRC14:13
*** rwallner has joined #openstack-kuryr14:14
ltomasboapuimedo, I'm trying to push the patchset with the modification to NeutronVIFPool instead of GenericVIFPool14:22
ltomasbobut getting some internal server error...14:22
ltomasboso, I'll try it later...14:22
mchiapperohi guys14:27
mchiapperosorry I've been super busy these days14:28
mchiapperoltomasbo: I'm not too sure on how to answer your question14:28
dmelladoo/ hi mchiappero np! ;)14:29
mchiapperohi :)14:29
ltomasbohi!14:29
mchiapperobecause I'm not sure there is a way to check it from the switch14:29
mchiapperoI should check the code in Neutron14:29
mchiapperodo you think you might have an idea already?14:33
mchiapperobecause I'm not sure Neutron applies and return or successfully enqueues the request and passes it down to the agent14:34
ltomasbowell, I assume this is some rule that the neutron-server needs to include in the DB14:39
ltomasboand then te agent where the port is, needs to include in ovs14:40
ltomasbojust not sure if the call will return before the agent applies the rule, or it will wait for it14:40
ltomasbonot sure if this could work, but, perhaps you can try to disable the neutron agent, and trigger the command14:41
ltomasboand see if it returns or gives you an error14:41
ltomasbomchiappero, ^^14:41
*** salv-orlando has joined #openstack-kuryr14:42
*** salv-orl_ has joined #openstack-kuryr14:44
mchiapperook, but the problem is: do we have a way to get that feedback back from the agent?14:47
mchiapperootherwise we have no choice anyway14:48
mchiappero:(14:48
*** salv-orlando has quit IRC14:48
kzaitsev_wsas far as I understand the feedback is indirect for ovs ports at least14:48
kzaitsev_wsi.e. if the vif is not active — controller keeps pinging neutron14:49
mchiapperoyou mean that you may get a 200 from neutron but checking the port later would work?14:50
kzaitsev_wswhen neutron says that port is ready controller says ok. marks vif as ready. then the CNI reads an update in the annotation and says OK pod is ready14:50
mchiapperothe port will never become active as there is no actual pluggin14:50
mchiappero*plugging14:50
mchiapperoI'll try to make some tests, I'll let you know14:51
kzaitsev_wsyou mean neutron would always think that it's down?14:51
*** egonzalez has quit IRC14:52
kzaitsev_wsI think what ltomasbo wants is to be sure, that you can use network on the pod the moment CNI says 'ok'14:52
ivc_mchiappero kzaitsev_ws what kind of feedback are you discussing?14:53
mchiapperoyeah sure, but the port will stay down14:53
mchiapperoit's only used for the ip address14:53
kzaitsev_wsivc_: I just jumped into mchiappero and ltomasbo discussion on macvlan nested ports =)14:54
mchiapperoivc_: on the allowed address pairs14:55
ltomasbomchiappero, yep, the port status it does not matter14:56
ltomasbobut the status of the allowed address pairs does14:56
ltomasboif the neutron-agent has not include it the rule by the time the container starts14:56
ltomasbothen we end up in the same problem we could have if not doing the pooling for the port to become active in the other cases14:56
mchiapperono, I understand perfectly14:58
ivc_ltomasbo mchiappero thats another reason not to use allowed address pairs then :)14:58
mchiapperoI'm just thinking of: 1) how to correctly check it 2) whether having that check that slows the process down is a good tradeoff anyway14:59
ltomasbomchiappero, ivc_, I was just wondering if the neutron call returns once the actions is applied15:01
ivc_mchiappero i'm not sure if it's even possible to check at all15:01
ltomasboor just once it is in the DB15:01
ivc_ltomasbo its just the DB15:01
ltomasbothat was my feeling, and thus my concern15:01
mchiapperoivc_: that is what I was trying to understand15:03
openstackgerritLuis Tomas Bolivar proposed openstack/kuryr-kubernetes master: Nested vif driver extension to enable ports reuse  https://review.openstack.org/43689315:04
mchiapperoso what then? :)15:04
ltomasbowell15:06
ltomasboyou can do some pooling15:06
ltomasboand request the port info and ensure the allowed pairs is updated15:06
ltomasboor that information is updated upfront too?15:06
mchiapperoI think ivc_ implied the latter15:07
ltomasboI guess we need to see in neutron agent triggers any notification, and how that one is handled by the neutron-server (i.e, what is changed upon notification)15:08
ivc_ltomasbo mchiappero i'm greping the code now, but if you have an env you can prolly check it (by running a vm, killing neutron agent after vm is ready and then updating allowed_address_pair)15:09
openstackgerritLuis Tomas Bolivar proposed openstack/kuryr-kubernetes master: Generic vif pool driver extension to precreate reusable ports  https://review.openstack.org/43687715:09
openstackgerritLuis Tomas Bolivar proposed openstack/kuryr-kubernetes master: Nested vif driver extension to enable ports reuse  https://review.openstack.org/43689315:10
mchiapperoI'll try soon15:10
*** hongbin has joined #openstack-kuryr15:11
openstackgerritLuis Tomas Bolivar proposed openstack/kuryr-kubernetes master: Nested vlan vif pool driver extension to precreate reusable subports  https://review.openstack.org/43689415:12
ltomasboumm, not sure why my patches Ids got somehow corrupted, I had to manually do a rebase, one by one...15:13
ltomasboivc_, yep, that was my idea on how to test that15:15
ltomasboif it gives you an error, or at least following commands on openstack port show does not show that information15:15
ltomasbothen we are good15:15
*** salv-orl_ has quit IRC15:16
*** salv-orlando has joined #openstack-kuryr15:17
apuimedoI'm deep in devstack-overcloud-heat hell15:17
apuimedoman... Writing deployment tools sucks15:18
openstackgerritLuis Tomas Bolivar proposed openstack/kuryr-kubernetes master: Change from GenericVIFPool to NeutronVIFPool  https://review.openstack.org/47273415:18
apuimedoit's very nice to use once it exists though15:18
apuimedodoesn't help that fscking openstack cli does > /dev/null with a lot of info15:18
*** salv-orlando has quit IRC15:20
ltomasboapuimedo, check the last ports pool patch (I lost the +2 after some minor update)15:22
ltomasboapuimedo, and the new patch to keep aligned with NeutronVIF15:22
apuimedookey dokey15:22
ltomasboapuimedo, thanks!15:24
*** rwallner has quit IRC15:37
*** kzaitsev_ws has quit IRC15:49
*** aojea has quit IRC15:51
*** egonzalez has joined #openstack-kuryr15:54
*** salv-orlando has joined #openstack-kuryr15:55
mchiapperoivc_: ltomasbo: it seems it's stored in the port right away16:08
ltomasbomchiappero, you mean in the db? and in the port show info?16:09
ltomasbowith the agent down?16:09
mchiapperoyes16:10
ltomasbomchiappero, that's bad...16:10
ltomasbowe need to think in a way to get that info then16:11
mchiapperoit's how it works in kuryr-libnetwork though16:11
ivc_mchiappero *works most of the time maybe? :)16:12
ivc_if you are lucky and its faster than app startup you'll be fine :)16:12
mchiapperoivc_: I meant that it's the same logic we implemented in kuryr-libnetwork16:13
mchiapperodoes the timing make any difference?16:14
mchiapperoI mean, the biggest problem is not having the flows16:14
mchiapperohaving them is still a problem but less severe16:15
ivc_mchiappero i mean when you tell neutron to add allowed pair it takes time on neutron/agent side, but at the same time you update annotation and cni needs some time to detect it16:17
ivc_mchiappero so as long as it takes more time on container side than on neutron side, your app will have access16:17
ivc_mchiappero and i'm not trying to say its ok :)16:18
mchiapperoI understood already16:18
ivc_its (works sometimes) vs (just works) :)16:19
mchiapperono, ok, I used a bad wording16:20
mchiapperowhat I meant to say is that, having a delay in getting the connectivity is a problem but it would work in the end16:21
mchiapperoif for some reason there is a failure it's impossible to detect it, and it's fairly severe. Although in that case probably you would be out of luck anyway16:22
mchiapperook, then?16:26
mchiapperoactually it's something that in theory should be fixed in neutron16:27
*** ltomasbo is now known as ltomasbo|away16:32
*** pcaruana has quit IRC16:34
*** rwallner has joined #openstack-kuryr16:37
*** rwallner has quit IRC16:42
*** pmannidi has quit IRC16:57
*** pmannidi has joined #openstack-kuryr16:57
*** rwallner has joined #openstack-kuryr17:01
openstackgerritAntoni Segura Puimedon proposed openstack/kuryr-kubernetes master: [WIP] contrib: devstack-overcloud-heat  https://review.openstack.org/47276317:13
*** x00350071_ has joined #openstack-kuryr17:15
*** aojea has joined #openstack-kuryr17:15
*** edisonxiang has quit IRC17:19
*** aojea has quit IRC17:19
*** tonanhngo has joined #openstack-kuryr17:21
*** garyloug has quit IRC17:26
*** aojea has joined #openstack-kuryr17:45
*** salv-orlando has quit IRC17:50
*** salv-orlando has joined #openstack-kuryr17:51
*** salv-orlando has quit IRC17:55
*** rwallner has quit IRC17:56
*** rwallner has joined #openstack-kuryr17:59
*** rwallner has quit IRC18:22
*** rwallner has joined #openstack-kuryr18:34
*** aojea has quit IRC19:05
*** aojea has joined #openstack-kuryr19:20
*** rwallner has quit IRC19:21
*** aojea has quit IRC19:25
*** egonzalez has quit IRC19:47
*** salv-orlando has joined #openstack-kuryr19:57
*** aojea has joined #openstack-kuryr20:21
*** salv-orl_ has joined #openstack-kuryr20:44
*** salv-orlando has quit IRC20:46
*** limao has joined #openstack-kuryr21:56
*** limao_ has joined #openstack-kuryr21:59
*** limao has quit IRC22:00
*** limao_ has quit IRC22:06
*** limao has joined #openstack-kuryr22:07
*** limao_ has joined #openstack-kuryr22:18
*** limao__ has joined #openstack-kuryr22:19
*** limao has quit IRC22:21
*** limao_ has quit IRC22:22
*** limao__ has quit IRC22:29
*** limao has joined #openstack-kuryr22:30
*** limao has quit IRC22:30
*** limao has joined #openstack-kuryr22:31
*** aojea has quit IRC22:33
*** aojea has joined #openstack-kuryr22:34
*** aojea has quit IRC22:39
*** limao has quit IRC22:42
*** limao has joined #openstack-kuryr22:43
*** limao has quit IRC22:52
*** limao has joined #openstack-kuryr22:53
*** salv-orl_ has quit IRC22:56
*** salv-orlando has joined #openstack-kuryr22:57
*** salv-orlando has quit IRC23:01
*** limao has quit IRC23:13
*** limao has joined #openstack-kuryr23:14
*** limao has quit IRC23:35
*** aojea has joined #openstack-kuryr23:35
*** limao has joined #openstack-kuryr23:39
*** aojea has quit IRC23:40

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!