| *** kzaitsev_mb has joined #openstack-kuryr | 00:01 | |
| *** hongbin has quit IRC | 00:10 | |
| *** gianpiet_ has quit IRC | 00:38 | |
| *** gianpietro has joined #openstack-kuryr | 00:38 | |
| *** gianpietro has quit IRC | 01:08 | |
| *** gianpietro has joined #openstack-kuryr | 01:09 | |
| *** gianpietro has quit IRC | 01:13 | |
| *** s1061123 has quit IRC | 01:24 | |
| *** yedongcan has joined #openstack-kuryr | 01:27 | |
| *** s1061123 has joined #openstack-kuryr | 01:34 | |
| *** gianpietro has joined #openstack-kuryr | 01:57 | |
| *** gianpietro has quit IRC | 02:01 | |
| *** hongbin has joined #openstack-kuryr | 02:09 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/kuryr master: Updated from global requirements https://review.openstack.org/462106 | 02:09 |
|---|---|---|
| openstackgerrit | OpenStack Proposal Bot proposed openstack/kuryr-libnetwork master: Updated from global requirements https://review.openstack.org/461376 | 02:09 |
| *** xiangxinyong is now known as edisonxianbg | 02:31 | |
| *** edisonxianbg is now known as edisonxiang | 02:31 | |
| *** gianpietro has joined #openstack-kuryr | 02:59 | |
| *** gianpietro has quit IRC | 03:03 | |
| *** edisonxiang has quit IRC | 03:10 | |
| *** yuanying has joined #openstack-kuryr | 03:24 | |
| *** vikasc has joined #openstack-kuryr | 03:31 | |
| *** hongbin has quit IRC | 03:32 | |
| *** gouthamr has quit IRC | 03:57 | |
| *** yamamoto has quit IRC | 04:13 | |
| *** yamamoto has joined #openstack-kuryr | 04:13 | |
| *** yamamoto has quit IRC | 04:17 | |
| *** janki has joined #openstack-kuryr | 04:32 | |
| *** kzaitsev_mb has quit IRC | 04:40 | |
| *** ltomasbo|away is now known as ltomasbo | 04:57 | |
| *** yamamoto has joined #openstack-kuryr | 05:01 | |
| *** kzaitsev_ws has joined #openstack-kuryr | 06:12 | |
| *** pcaruana has joined #openstack-kuryr | 06:20 | |
| *** dimak_ has joined #openstack-kuryr | 06:38 | |
| *** vikasc has quit IRC | 07:30 | |
| irenab | apuimedo, ping | 07:32 |
| apuimedo | irenab: pong | 07:39 |
| apuimedo | dmellado: rise and shine | 07:39 |
| irenab | apuimedo, I think I found the answer ... | 07:39 |
| irenab | good morning :-) | 07:40 |
| apuimedo | I'd say good if it were not that the answer is that you can't verify the patch | 07:40 |
| apuimedo | XD | 07:40 |
| apuimedo | kzaitsev_ws: any luck verifying the macvlan patch | 07:40 |
| apuimedo | ? | 07:40 |
| apuimedo | irenab: I had a busy night. Did you see the bug I reported? | 07:41 |
| kzaitsev_ws | apuimedo: not yet, sorry. I have a couple of meetings today, that I've been preparing for | 07:41 |
| irenab | apuimedo, yes, looks crazy | 07:42 |
| irenab | but seems you maid it work eventually | 07:42 |
| apuimedo | kzaitsev_ws: it's good to always go to meetings prepared | 07:42 |
| apuimedo | :-) | 07:42 |
| apuimedo | irenab: it is working indeed | 07:42 |
| apuimedo | but the matter is now how to solve the bug | 07:43 |
| apuimedo | I would probably do it in two steps | 07:43 |
| irenab | apuimedo, I am trying to verify the patch for devstack with pod host stuff | 07:43 |
| apuimedo | first step, we put a config option for adding bm host to the pod network (using kuryr-lib binding scripts) | 07:43 |
| apuimedo | second step, we put worker nodes in a separate net and we make a handler for liveness and readiness probes | 07:44 |
| *** xiangxinyong has joined #openstack-kuryr | 07:44 | |
| *** egonzalez has joined #openstack-kuryr | 07:44 | |
| *** xiangxinyong is now known as edisonxiang | 07:45 | |
| irenab | apuimedo, first step is to make controller and services like DNS to be able to talk to API? | 07:45 |
| apuimedo | no, that is taken care of by my patch | 07:48 |
| apuimedo | first step is to make the dns and such services not be constantly killed due to the kubelet failure to make http probes to them | 07:48 |
| irenab | apuimedo, dmellado any idea why unstack/clean and then stuck with updated kuryr-kubernetes branch ends up with issues with 'getting to identity versions ...'? | 07:49 |
| irenab | no openstack cli command is working | 07:49 |
| apuimedo | what gets stuck? the stacking? | 07:49 |
| irenab | apuimedo, maybe probes can be disabled, at least for devstack | 07:49 |
| irenab | stacking seems to be completed, but no cli command is working | 07:50 |
| irenab | I wonder if any of you had issues with unstack/clean and then stack again | 07:50 |
| apuimedo | irenab: I usually never unstack | 07:52 |
| apuimedo | but when I did lately it worked | 07:52 |
| apuimedo | 3/4 times last week | 07:52 |
| apuimedo | but previously once I got stacking to get stuck | 07:53 |
| apuimedo | hence my earlier question | 07:53 |
| irenab | I will try to create clean VM to run devstack with your patch | 07:53 |
| apuimedo | irenab: which? | 07:56 |
| irenab | the one with devstack changes to deploy LB | 07:57 |
| irenab | https://review.openstack.org/#/c/467241/5 | 07:58 |
| irenab | apuimedo, something more urgent? | 07:58 |
| apuimedo | irenab: uh | 07:59 |
| apuimedo | wait a sec | 07:59 |
| apuimedo | you are deploying a devstack with heat to run this patch atop? | 08:00 |
| apuimedo | so the kuryr deployment would be in second level of virtualization? | 08:00 |
| irenab | no, two separate verifications. One for heat and the second for the above patch | 08:01 |
| kzaitsev_ws | irenab: thanks for the spec review. I'll prep the comments today. Here are just some inititial thoughs. I do understand that SRIOV handler seem to violate the 1-k8s-entity 1-handler design. My initial goal was to minimize the changes to original handler/workflow | 08:08 |
| kzaitsev_ws | in this case one could argue, that an opaque-resource-request is an entity, that vif maps to | 08:10 |
| apuimedo | ok | 08:10 |
| apuimedo | kzaitsev_ws: irenab: Personally I prefer having multiple handlers for the same entity than having branch heavy handlers | 08:11 |
| apuimedo | since handlers can work in parallel | 08:11 |
| apuimedo | and simplicity of handlers is important | 08:11 |
| irenab | apuimedo, kzaitsev_pi at a meeting, will be back soon to the conversation | 08:11 |
| kzaitsev_ws | it might worth thinking about adding/extending a multi-vif handler as you mentioned, but the logic for requesting them still lies inside the pod object itself. i.e. we would have to inspect to determine whether it wants direct vifs or not. | 08:13 |
| kzaitsev_ws | irenab: sure =) | 08:14 |
| kzaitsev_ws | My meeting would start in approx. 4 hours, so I should be available | 08:16 |
| janonymous | a silly question: Can a container launched on different Baremetal machine use kuryr which have only container running only? | 08:32 |
| apuimedo | janonymous: can you rephrase? | 08:38 |
| janonymous | Ops sry, suppose we have a setup with Openstack and kubernetes and kubernetes schedules a pod on let's say a baremetal node | 08:40 |
| janonymous | Is it feasible? | 08:42 |
| *** garyloug has joined #openstack-kuryr | 08:44 | |
| irenab | apuimedo, kzaitsev_ws : back | 08:52 |
| apuimedo | janonymous: yes. Kuryr works well in baremetal | 08:57 |
| apuimedo | you only need to have ovs agent on the worker nodes | 08:57 |
| janonymous | oh cool, and if baremetal node is in different region/network? | 08:59 |
| apuimedo | I have not tried that | 09:00 |
| kzaitsev_ws | sorry, the meeting got transfered to now =( will be back in 1h or so | 09:00 |
| apuimedo | kzaitsev_ws: np | 09:00 |
| janonymous | apuimdeo: thanks! :) | 09:02 |
| irenab | ok | 09:03 |
| apuimedo | you're welcome | 09:06 |
| *** vikasc has joined #openstack-kuryr | 09:07 | |
| kzaitsev_ws | back for another 40 mins %) sorry for being sporadic | 09:10 |
| *** pmannidi__ has quit IRC | 09:10 | |
| kzaitsev_ws | well, my biggest fear with a separate handler was that k8s could have refused an upgrade with resource version | 09:11 |
| kzaitsev_ws | but in fact it swallows the thing nicely, since I'm annotating a different key in the pod | 09:11 |
| kzaitsev_ws | (which is also a bit controversial, but I can't find a way to do it without breaking/altering the main handler) | 09:11 |
| kzaitsev_ws | it may be worth adding multi-vif handler, however with current model I'm not sure if that could be practical | 09:12 |
| kzaitsev_ws | at least there should be a way to request ports from different networks | 09:13 |
| kzaitsev_ws | maybe with what danil is doing (with requesting neutron ports by id) it would be more meaningful | 09:14 |
| apuimedo | janonymous: ivc_: irenab: kzaitsev_ws: vikasc: https://blueprints.launchpad.net/kuryr-kubernetes/+spec/cni-split-exec-daemon | 09:25 |
| janonymous | @apuimedo: great! | 09:26 |
| janonymous | @apuimedo: I would like to work on that :) | 09:27 |
| apuimedo | janonymous: I pointed to some example code on the executable | 09:29 |
| apuimedo | I would recomment starting with that | 09:29 |
| apuimedo | and do the server part later | 09:29 |
| *** aojea has joined #openstack-kuryr | 09:44 | |
| kzaitsev_ws | kuryr-k8s is bursting with blueprints lately ) | 09:46 |
| kzaitsev_ws | oh, so you want cni daemon to be actual k8s daemonset, right? | 09:47 |
| apuimedo | kzaitsev_ws: that's exactly right | 09:49 |
| apuimedo | on both accounts | 09:49 |
| apuimedo | when I open my browser the first link is gmail, the second is blueprints | 09:49 |
| apuimedo | :'( | 09:49 |
| kzaitsev_ws | and it would cache info on all the alive pods it has ever seen and at the time request comes — would plug the vif in | 09:49 |
| kzaitsev_ws | or hang untill the pod is ready | 09:50 |
| apuimedo | kzaitsev_ws: right | 09:50 |
| apuimedo | and on delete time it would already have the info | 09:50 |
| apuimedo | or get it if it was restarted by the liveness probe | 09:50 |
| kzaitsev_ws | I think ivc_ mentioned that after the split cni-daemon would be more pluggable. | 09:54 |
| kzaitsev_ws | but that doesnt seem to be the case. | 09:54 |
| kzaitsev_ws | or maybe just the archeticture would be a bit more straightforward | 09:54 |
| apuimedo | kzaitsev_ws: do you remember where the pluggability was supposed to be gained? I can't imagine it more pluggable with the possibility of having multiple handlers | 09:56 |
| kzaitsev_ws | and it would have a single event-handling loop, which would be easy to plug into. rahter than current every-handler-ends-the-loop scenario | 09:56 |
| kzaitsev_ws | I'll be digging cni handlers soon to see how I can avoid my custom sriov-cni and not break the regular workflow | 09:59 |
| apuimedo | ok | 10:01 |
| *** yamamoto has quit IRC | 10:11 | |
| irenab | apuimedo, kzaitsev_ws : following the initial design concept, the idea was is that there is a Handler per k8s entity and the end result in SDN (Neutron). So there is a Handler for handle Pod events and provide VIF. SR-IOV port is just a type of vif, at least from the neutron perspective. I would like to hear the ivc_ point of view before approving the current direction of separate handler | 10:19 |
| apuimedo | ivc_ input would be great | 10:20 |
| apuimedo | personally I always wanted separate handlers running in parallel | 10:20 |
| apuimedo | :-) | 10:20 |
| apuimedo | (and namespacing annotations to avoid conflicts) | 10:20 |
| *** vikasc has quit IRC | 10:30 | |
| *** vikasc has joined #openstack-kuryr | 10:46 | |
| *** yamamoto has joined #openstack-kuryr | 10:58 | |
| *** yedongcan has left #openstack-kuryr | 11:12 | |
| *** aojea has quit IRC | 11:15 | |
| kzaitsev_ws | irenab: acknowledged =) would try to ping ivc_ on it =) | 11:17 |
| irenab | kzaitsev_ws, great. Thank you for your patience | 11:19 |
| kzaitsev_ws | np. actually from my perspective you guys are pretty fast with spec reviews. ;) | 11:20 |
| *** aojea has joined #openstack-kuryr | 11:20 | |
| irenab | apuimedo, got the same issue with clean devstack deployment: https://pastebin.com/Tms8qCGe | 11:23 |
| apuimedo | thanks | 11:28 |
| apuimedo | will check after lunch | 11:28 |
| *** gianpietro has joined #openstack-kuryr | 11:35 | |
| *** gianpietro has quit IRC | 11:40 | |
| irenab | ltomasbo, ping | 11:49 |
| *** gianpietro has joined #openstack-kuryr | 11:52 | |
| *** gianpietro has quit IRC | 11:53 | |
| *** atoth has joined #openstack-kuryr | 12:17 | |
| *** gianpietro has joined #openstack-kuryr | 12:24 | |
| *** roeyc has joined #openstack-kuryr | 12:25 | |
| *** roeyc has quit IRC | 12:25 | |
| *** roeyc has joined #openstack-kuryr | 12:25 | |
| *** gianpietro has quit IRC | 12:26 | |
| *** zengchen has quit IRC | 12:30 | |
| *** gianpietro has joined #openstack-kuryr | 12:32 | |
| *** gianpietro has quit IRC | 12:32 | |
| openstackgerrit | Merged openstack/kuryr-kubernetes master: Improve nested k8s devstack install mechanism https://review.openstack.org/467532 | 12:36 |
| ltomasbo | irenab, pong | 12:51 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/kuryr-kubernetes master: Updated from global requirements https://review.openstack.org/464394 | 12:52 |
| *** yamamoto has quit IRC | 12:56 | |
| irenab | ltomasbo, wanted to ask you about heat patch verification, but got assistance from apuimedo already | 12:56 |
| *** jchhatbar has joined #openstack-kuryr | 12:58 | |
| *** janki has quit IRC | 12:58 | |
| ltomasbo | irenab, ok, I got some problems with the openstackcli, but solved with virtualenv | 12:59 |
| irenab | ltomasbo, I didn't have image in glance .... hopefully get to the issues you had soon | 13:00 |
| ltomasbo | irenab, just for curiosity, what was the problem? I left the devstack stacking and I see it had failed | 13:00 |
| ltomasbo | irenab, xD | 13:00 |
| *** gouthamr has joined #openstack-kuryr | 13:01 | |
| *** yamamoto has joined #openstack-kuryr | 13:02 | |
| irenab | ltomasbo, seems it finished with success. I see once server created. | 13:03 |
| dmellado | heh, I'll try to verify the patch too, /me crosses fingers | 13:06 |
| irenab | dmellado, I am using your VM to verify, it is very helpful | 13:09 |
| dmellado | irenab: glad that it helps ;) | 13:09 |
| *** gianpietro has joined #openstack-kuryr | 13:11 | |
| irenab | apuimedo, should the sec. groups be ok to enable ssh to the created server? | 13:11 |
| irenab | actually I do not see that stackname.pem is created | 13:15 |
| *** jchhatbar is now known as janki | 13:15 | |
| *** atoth has quit IRC | 13:17 | |
| *** atoth has joined #openstack-kuryr | 13:18 | |
| *** atoth has quit IRC | 13:18 | |
| *** atoth has joined #openstack-kuryr | 13:19 | |
| *** roeyc has quit IRC | 13:27 | |
| *** roeyc has joined #openstack-kuryr | 13:28 | |
| *** roeyc has quit IRC | 13:28 | |
| openstackgerrit | Danil Golov proposed openstack/kuryr-kubernetes master: Allow setting specific ports for SRIOV handler https://review.openstack.org/468032 | 13:29 |
| openstackgerrit | Danil Golov proposed openstack/kuryr-kubernetes master: [WIP] Allow setting specific ports for SRIOV handler https://review.openstack.org/468032 | 13:31 |
| *** dimak_ has quit IRC | 13:33 | |
| apuimedo | irenab: I'm back | 13:37 |
| apuimedo | sorry, taking the kid to school | 13:37 |
| apuimedo | did you try to do the ./devstack-heat ssh thing? | 13:37 |
| apuimedo | ltomasbo: did you try my container celebdor/openstack:ocata for the openstack client or just used python venv? | 13:39 |
| ltomasbo | apuimedo, just virtualenv with pip | 13:43 |
| apuimedo | ok | 13:47 |
| *** gianpietro has quit IRC | 13:51 | |
| *** janki has quit IRC | 14:02 | |
| *** yamamoto has quit IRC | 14:24 | |
| *** yamamoto has joined #openstack-kuryr | 14:39 | |
| *** yamamoto has quit IRC | 14:44 | |
| *** hongbin has joined #openstack-kuryr | 14:45 | |
| *** x00350071_ has joined #openstack-kuryr | 14:59 | |
| *** edisonxiang has quit IRC | 14:59 | |
| *** kzaitsev_ws has quit IRC | 15:00 | |
| *** ltomasbo is now known as ltomasbo|away | 15:04 | |
| *** gsagie has joined #openstack-kuryr | 15:04 | |
| *** gianpietro has joined #openstack-kuryr | 15:07 | |
| *** roeyc has joined #openstack-kuryr | 15:19 | |
| *** yamamoto has joined #openstack-kuryr | 15:20 | |
| *** yamamoto has quit IRC | 15:20 | |
| *** pcaruana has quit IRC | 15:38 | |
| *** pcaruana has joined #openstack-kuryr | 15:39 | |
| *** ltomasbo|away is now known as ltomasbo | 15:39 | |
| *** leifmadsen has quit IRC | 15:45 | |
| *** gianpietro has quit IRC | 15:46 | |
| *** leifmadsen has joined #openstack-kuryr | 15:49 | |
| *** aojea has quit IRC | 15:54 | |
| *** gsagie has quit IRC | 16:00 | |
| *** egonzalez has quit IRC | 16:05 | |
| *** roeyc has quit IRC | 16:09 | |
| *** roeyc has joined #openstack-kuryr | 16:10 | |
| *** roeyc has quit IRC | 16:11 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/fuxi master: Updated from global requirements https://review.openstack.org/454474 | 16:19 |
| *** yamamoto has joined #openstack-kuryr | 16:21 | |
| *** pcaruana has quit IRC | 16:30 | |
| *** pcaruana has joined #openstack-kuryr | 16:30 | |
| *** yamamoto has quit IRC | 16:32 | |
| openstackgerrit | Antoni Segura Puimedon proposed openstack/kuryr-kubernetes master: devstack: Add configuration for kubelet probes https://review.openstack.org/468100 | 16:44 |
| apuimedo | ltomasbo: please review ^^ | 16:45 |
| apuimedo | kzaitsev_pi: ^^ | 16:45 |
| *** ltomasbo is now known as ltomasbo|away | 16:48 | |
| *** garyloug has quit IRC | 16:52 | |
| openstackgerrit | Merged openstack/kuryr-kubernetes master: get nested drivers opts out of neutron defaults https://review.openstack.org/450113 | 17:05 |
| *** yamamoto has joined #openstack-kuryr | 17:06 | |
| *** yamamoto has quit IRC | 17:10 | |
| *** gianpietro has joined #openstack-kuryr | 17:11 | |
| openstackgerrit | Merged openstack/kuryr-kubernetes master: Updated from global requirements https://review.openstack.org/464394 | 17:15 |
| *** gianpietro has quit IRC | 17:24 | |
| *** gianpietro has joined #openstack-kuryr | 17:25 | |
| *** gianpietro has quit IRC | 17:29 | |
| *** egonzalez has joined #openstack-kuryr | 17:55 | |
| *** vikasc has quit IRC | 19:00 | |
| *** aojea has joined #openstack-kuryr | 19:06 | |
| *** yamamoto has joined #openstack-kuryr | 19:08 | |
| *** pcaruana has quit IRC | 19:09 | |
| *** yamamoto has quit IRC | 19:13 | |
| *** gianpietro has joined #openstack-kuryr | 20:57 | |
| *** aojea has quit IRC | 20:59 | |
| *** gianpietro has quit IRC | 21:14 | |
| *** gianpietro has joined #openstack-kuryr | 21:54 | |
| *** gianpietro has quit IRC | 21:59 | |
| *** gianpietro has joined #openstack-kuryr | 22:21 | |
| *** gouthamr has quit IRC | 22:29 | |
| *** gianpietro has quit IRC | 22:30 | |
| *** gianpietro has joined #openstack-kuryr | 22:39 | |
| *** gianpietro has quit IRC | 22:40 | |
| openstackgerrit | Alan Pevec proposed openstack/kuryr-kubernetes master: Fix typo https://review.openstack.org/468190 | 22:42 |
| *** gianpietro has joined #openstack-kuryr | 22:42 | |
| *** gianpietro has quit IRC | 22:44 | |
| openstackgerrit | Alan Pevec proposed openstack/kuryr-kubernetes master: Fix typo https://review.openstack.org/468190 | 22:58 |
| *** gouthamr has joined #openstack-kuryr | 23:05 | |
| *** egonzalez has quit IRC | 23:06 | |
| *** ashishb has quit IRC | 23:10 | |
| *** hongbin has quit IRC | 23:58 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!