| *** aojea has joined #openstack-kuryr | 00:30 | |
| *** aojea has quit IRC | 00:35 | |
| *** dougbtv|laptop has joined #openstack-kuryr | 00:54 | |
| *** dougbtv|laptop has quit IRC | 01:00 | |
| *** yamamoto has joined #openstack-kuryr | 02:29 | |
| *** vikasc has joined #openstack-kuryr | 03:23 | |
| *** yamamoto has quit IRC | 03:37 | |
| *** yamamoto has joined #openstack-kuryr | 04:16 | |
| *** svinota has joined #openstack-kuryr | 05:14 | |
| *** janki has joined #openstack-kuryr | 05:14 | |
| *** svinota has quit IRC | 05:31 | |
| *** dimak_ has joined #openstack-kuryr | 05:40 | |
| *** gsagie has joined #openstack-kuryr | 05:56 | |
| *** kzaitsev_ws has joined #openstack-kuryr | 06:07 | |
| *** yuanying has quit IRC | 06:07 | |
| *** ltomasbo|away is now known as ltomasbo | 06:08 | |
| *** yuanying_ has joined #openstack-kuryr | 06:13 | |
| *** pcaruana has joined #openstack-kuryr | 06:57 | |
| *** neiljerram has joined #openstack-kuryr | 07:03 | |
| *** neiljerram has quit IRC | 07:19 | |
| *** aojea has joined #openstack-kuryr | 07:21 | |
| *** egonzalez has joined #openstack-kuryr | 07:49 | |
| *** apuimedo has quit IRC | 08:02 | |
| *** apuimedo has joined #openstack-kuryr | 08:02 | |
| *** openstackgerrit has quit IRC | 08:03 | |
| *** aojea_ has joined #openstack-kuryr | 08:21 | |
| *** aojea has quit IRC | 08:24 | |
| *** neiljerram has joined #openstack-kuryr | 08:54 | |
| *** garyloug has joined #openstack-kuryr | 08:56 | |
| *** neiljerram has quit IRC | 09:04 | |
| apuimedo | irenab: vikasc: please review https://review.openstack.org/#/c/376045/ | 09:05 |
|---|---|---|
| vikasc | apuimedo, sure! | 09:06 |
| apuimedo | thanks | 09:06 |
| *** pc_m has quit IRC | 10:08 | |
| *** pc_m has joined #openstack-kuryr | 10:10 | |
| *** gsagie has quit IRC | 11:15 | |
| *** neiljerram has joined #openstack-kuryr | 11:20 | |
| *** neiljerram has quit IRC | 11:47 | |
| *** dougbtv_ has quit IRC | 12:08 | |
| *** dougbtv_ has joined #openstack-kuryr | 12:33 | |
| *** dougbtv_ is now known as dougbtv | 12:36 | |
| *** limao has joined #openstack-kuryr | 12:36 | |
| *** neiljerram has joined #openstack-kuryr | 12:39 | |
| *** limao has quit IRC | 12:44 | |
| *** limao has joined #openstack-kuryr | 12:46 | |
| *** limao has quit IRC | 12:52 | |
| *** limao has joined #openstack-kuryr | 12:53 | |
| *** openstackgerrit has joined #openstack-kuryr | 12:57 | |
| openstackgerrit | Liping Mao proposed openstack/kuryr-libnetwork master: Add NetworkDriver api AllocateNetwork and FreeNetwork https://review.openstack.org/451475 | 12:57 |
| *** limao has quit IRC | 13:04 | |
| *** limao has joined #openstack-kuryr | 13:04 | |
| openstackgerrit | Liping Mao proposed openstack/kuryr-libnetwork master: [WIP]Update Kuryr-libnetwork Docker managed plugin related doc https://review.openstack.org/451479 | 13:06 |
| *** limao has quit IRC | 13:15 | |
| *** limao has joined #openstack-kuryr | 13:16 | |
| *** limao has quit IRC | 13:25 | |
| *** limao has joined #openstack-kuryr | 13:29 | |
| *** limao has quit IRC | 13:37 | |
| *** mattmceuen has joined #openstack-kuryr | 13:39 | |
| *** janki has quit IRC | 13:42 | |
| *** janki has joined #openstack-kuryr | 13:44 | |
| *** limao has joined #openstack-kuryr | 13:46 | |
| *** hongbin has joined #openstack-kuryr | 13:53 | |
| *** limao has quit IRC | 13:55 | |
| *** limao has joined #openstack-kuryr | 13:56 | |
| *** limao_ has joined #openstack-kuryr | 14:06 | |
| *** limao has quit IRC | 14:09 | |
| openstackgerrit | Merged openstack/kuryr master: Updated from global requirements https://review.openstack.org/451019 | 14:17 |
| openstackgerrit | Merged openstack/kuryr-libnetwork master: Updated from global requirements https://review.openstack.org/451020 | 14:20 |
| openstackgerrit | Liping Mao proposed openstack/kuryr-libnetwork master: [WIP]Update Kuryr-libnetwork Docker managed plugin related doc https://review.openstack.org/451479 | 14:42 |
| *** janki has quit IRC | 14:48 | |
| ltomasbo | so, moving the discussion, ivc_ apuimedo, should I adapt the patches to follow this approach: https://review.openstack.org/#/c/446462/ | 15:04 |
| ltomasbo | kind of more similar to what I had at the beginning | 15:05 |
| ltomasbo | vif-handler owning the pool-driver and triggering the ports creation? | 15:05 |
| apuimedo | I would obviously like that :-) | 15:05 |
| apuimedo | I find it a much simpler approach | 15:05 |
| apuimedo | :P | 15:05 |
| ltomasbo | :D | 15:05 |
| ltomasbo | I like the idea (with I did not include in my first patches) about having a pool driver always | 15:06 |
| ltomasbo | but just a Noop driver | 15:06 |
| ltomasbo | and a base driver | 15:06 |
| *** limao_ has quit IRC | 15:06 | |
| *** limao has joined #openstack-kuryr | 15:07 | |
| ltomasbo | my main concern is how to repopulate the pool not during container creation, once the pool was already populated before | 15:07 |
| ltomasbo | ie., not waiting until there is no ports there, but when there is a number below X | 15:07 |
| apuimedo | ltomasbo: can you rephrase, I'm not sure I got the concert | 15:09 |
| apuimedo | can you walk us through an example? | 15:09 |
| ltomasbo | https://review.openstack.org/#/c/446462/1/kuryr_kubernetes/controller/drivers/vif_pool.py L55-L65 | 15:10 |
| ltomasbo | at the begining, it is clear | 15:10 |
| apuimedo | right. | 15:11 |
| ltomasbo | there are no ports at the pool, so we trigger the rquests_vif to create | 15:11 |
| ltomasbo | and lets say we create 10 | 15:11 |
| apuimedo | right | 15:11 |
| ltomasbo | at some point, if we target a minimum, lets say 5 | 15:11 |
| ltomasbo | if enough containers have been created | 15:12 |
| ltomasbo | ideally, I would like to create more ports in the pool, once we go below 5, not when we empty the pool | 15:12 |
| ltomasbo | so that, if we ensure certain minimum at the pool, the containers will not have to wait for the creation (only the first time) | 15:12 |
| dmellado | that'd be cool, somehow mimicking the nodepool idea... | 15:13 |
| ltomasbo | or if a big number of them needs to be created at once | 15:13 |
| ltomasbo | so, perhaps, in those lines | 15:13 |
| ltomasbo | we could include another checking at the end of the function | 15:13 |
| apuimedo | right | 15:13 |
| apuimedo | mmm | 15:14 |
| ltomasbo | to trigger some repopulation (in another thread) so that the container can proceed | 15:14 |
| apuimedo | well, you could certainly do that | 15:14 |
| ltomasbo | without waiting for that to complete | 15:14 |
| ltomasbo | otherwise, if we say the bulk creation is 10 | 15:14 |
| ltomasbo | we will have 1 container taking 10 times (a bit better as it is a bulk creation) the current time | 15:15 |
| apuimedo | sure | 15:15 |
| ltomasbo | and the other 9 will be pretty fast | 15:15 |
| apuimedo | do you have some alternative to adding that check at the end? | 15:15 |
| ltomasbo | great, so, we agreed on that? Vif-handler triggering the pool-driver | 15:16 |
| ltomasbo | which in turns calls the vif-driver | 15:16 |
| apuimedo | not yet | 15:16 |
| apuimedo | :P | 15:16 |
| ltomasbo | :D | 15:16 |
| apuimedo | still thinking | 15:16 |
| ltomasbo | on a related topic, the devref | 15:17 |
| ltomasbo | I replied some time ago to irenab and ivc_ comments | 15:17 |
| ltomasbo | ivc_, are you ok with the replies? anything extra or not convincing? should I update the devref including the comments/replies? | 15:18 |
| apuimedo | ltomasbo: couldn't we just have the pool be an object that when you pop can notify somehow | 15:19 |
| apuimedo | ? | 15:19 |
| ivc_ | ltomasbo i'm on a meeting atm, will check in about an hour | 15:19 |
| ltomasbo | ivc_, sure, thanks! | 15:19 |
| ltomasbo | apuimedo, let me picture that in my mind... | 15:19 |
| ltomasbo | apuimedo, still need to fix the repopulation issue, right? | 15:22 |
| apuimedo | I was thinking that maybe when you pop and it's under limit, you could get either a tuple or low-level exception (that includes the vif) | 15:24 |
| ltomasbo | umm | 15:25 |
| ltomasbo | perhaps at the vif handler | 15:25 |
| *** dimak_ has quit IRC | 15:25 | |
| ltomasbo | then you can handle that exception | 15:25 |
| ltomasbo | so that you have the vif returned, and the repopulation action | 15:26 |
| ltomasbo | but still, I don't want to wait for that action to complete, and the vif_handler could not move to the next action until it process that exception, right> | 15:27 |
| ltomasbo | ? | 15:27 |
| ltomasbo | need to go to yet another meeting today. I'll be right back in 30 min or so... | 15:28 |
| apuimedo | I'd rather the vif handler not know anything about it | 15:32 |
| apuimedo | you should be able to pass it to another green thread | 15:32 |
| *** limao has quit IRC | 15:38 | |
| *** aojea_ has quit IRC | 15:45 | |
| *** pcaruana has quit IRC | 15:53 | |
| *** egonzalez has quit IRC | 16:15 | |
| *** garyloug has quit IRC | 16:33 | |
| *** ltomasbo is now known as ltomasbo|away | 17:12 | |
| hongbin | dims: hi dims, i have several questions for this proposal: https://github.com/kubernetes/kubernetes/pull/30285/commits/95197536fc5be791e474b952aa939ac7a2299d7e , are you the right person to answer it? | 18:00 |
| dims | hongbin : not really :) we should try to ask folks in sig-storage over on slack | 18:01 |
| hongbin | ok, let me go there | 18:01 |
| *** tonanhngo has joined #openstack-kuryr | 18:43 | |
| *** aojea has joined #openstack-kuryr | 19:01 | |
| *** vikasc has quit IRC | 19:05 | |
| *** vikasc has joined #openstack-kuryr | 19:46 | |
| hongbin | apuimedo: hi toni, need your w+1 on this one: https://review.openstack.org/#/c/431085/ | 20:36 |
| *** aojea has quit IRC | 21:20 | |
| *** mattmceuen has quit IRC | 21:20 | |
| *** aojea has joined #openstack-kuryr | 21:20 | |
| *** aojea has quit IRC | 21:25 | |
| *** tonanhngo has quit IRC | 23:24 | |
| *** tonanhngo has joined #openstack-kuryr | 23:37 | |
| *** tonanhngo has quit IRC | 23:41 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!