| openstackgerrit | Jake Yip proposed openstack/magnum master: Upgrade cloud-provider-openstack to v1.14.0 https://review.opendev.org/655875 | 00:29 |
|---|---|---|
| *** dave-mccowan has joined #openstack-containers | 00:32 | |
| *** dave-mccowan has quit IRC | 00:35 | |
| *** ricolin_ has joined #openstack-containers | 00:44 | |
| *** itlinux has joined #openstack-containers | 00:55 | |
| *** itlinux has quit IRC | 01:37 | |
| *** itlinux has joined #openstack-containers | 01:44 | |
| *** hongbin has quit IRC | 01:51 | |
| *** itlinux has quit IRC | 01:56 | |
| *** goldyfruit has quit IRC | 02:03 | |
| openstackgerrit | Feilong Wang proposed openstack/magnum master: Add cluster upgrade to the API https://review.opendev.org/514959 | 02:03 |
| *** itlinux has joined #openstack-containers | 02:25 | |
| *** threestrands has joined #openstack-containers | 03:31 | |
| *** ramishra has joined #openstack-containers | 03:32 | |
| *** altlogbot_2 has quit IRC | 03:44 | |
| *** altlogbot_0 has joined #openstack-containers | 03:45 | |
| *** ykarel|away has joined #openstack-containers | 03:53 | |
| *** ykarel|away has joined #openstack-containers | 03:56 | |
| *** ykarel|away is now known as ykarel | 03:58 | |
| *** ykarel has quit IRC | 04:05 | |
| *** ykarel has joined #openstack-containers | 04:08 | |
| *** ykarel has quit IRC | 04:14 | |
| *** sapd1_x has joined #openstack-containers | 04:17 | |
| *** ykarel has joined #openstack-containers | 04:34 | |
| *** ivve has quit IRC | 04:37 | |
| *** altlogbot_0 has quit IRC | 04:38 | |
| *** altlogbot_2 has joined #openstack-containers | 04:39 | |
| *** trident has quit IRC | 04:39 | |
| *** altlogbot_2 has quit IRC | 04:40 | |
| *** janki has joined #openstack-containers | 04:40 | |
| *** altlogbot_3 has joined #openstack-containers | 04:41 | |
| *** trident has joined #openstack-containers | 04:41 | |
| *** absubram has joined #openstack-containers | 04:49 | |
| *** absubram has quit IRC | 04:54 | |
| *** sapd1_x has quit IRC | 05:00 | |
| *** absubram has joined #openstack-containers | 05:00 | |
| *** pcaruana has joined #openstack-containers | 05:15 | |
| *** ivve has joined #openstack-containers | 05:41 | |
| *** absubram has quit IRC | 05:50 | |
| *** rcernin has quit IRC | 06:01 | |
| *** rcernin has joined #openstack-containers | 06:01 | |
| *** lpetrut has joined #openstack-containers | 06:23 | |
| *** ykarel is now known as ykarel|afk | 06:51 | |
| *** rcernin has quit IRC | 06:52 | |
| *** ykarel|afk is now known as ykarel | 07:30 | |
| *** itlinux has quit IRC | 07:42 | |
| *** ykarel_ has joined #openstack-containers | 07:55 | |
| *** ykarel has quit IRC | 07:56 | |
| *** ykarel_ is now known as ykarel | 07:58 | |
| *** ricolin_ has quit IRC | 08:07 | |
| *** ricolin has joined #openstack-containers | 08:07 | |
| *** inovas has joined #openstack-containers | 08:11 | |
| *** ykarel is now known as ykarel|lunch | 08:30 | |
| *** ramishra has quit IRC | 08:30 | |
| zigo | jrosser: Do I understand correctly that, if I want to setup a proxy for the discovery URL, I have to do this as a http_proxy / https_proxy variable in the systemd unit of the magnum-conductor? | 08:32 |
| jrosser | zigo: good morning | 08:32 |
| jrosser | it depends how you have magnum deployed i expect | 08:32 |
| jrosser | if you are using bare metal or machine conatiners like LXC / openstack-ansible then you can set the proxy for a process by adding settings to the systemd unit file and defining environment variables | 08:33 |
| jrosser | and i believe there are similar mechanisms for docker and podman, so as i say it depends on how you have deployed magnum to a degree | 08:33 |
| zigo | jrosser: I'm using my own tool at https://salsa.debian.org/openstack-team/debian/openstack-cluster-installer | 08:34 |
| zigo | This uses puppet-openstack internally. | 08:34 |
| zigo | That's bare-metal deployment (ie: no containers). | 08:34 |
| jrosser | however, there is currently code in magnum which forces the proxy for the discovery URL for the magnum service to be the one that is defined in the cluster template | 08:34 |
| zigo | jrosser: You mean without reverting as per https://review.opendev.org/#/c/660316/ ? | 08:35 |
| zigo | You can consider that I applied the patch reversion ... :) | 08:35 |
| jrosser | ok :) | 08:36 |
| zigo | (I can do this, since that's my oen Debian packages ...) | 08:36 |
| zigo | jrosser: So, I was right then, just setup http_proxy / https_proxy as env variables in my systemd .service file, and I'm good to go, no? | 08:36 |
| jrosser | if you need the proxy settings to be in any way different for the service and VM then yes you must revert that | 08:36 |
| *** ramishra has joined #openstack-containers | 08:36 | |
| jrosser | yeah, let me just look what i have as an example | 08:36 |
| zigo | jrosser: I probably will setup a Environment=/etc/default/magnum, so that can work for everyone in Debian. | 08:37 |
| zigo | Then setup that proxy URL there... | 08:37 |
| zigo | That way, no need to edit the .service file in puppet also. | 08:37 |
| zigo | I also probably can contribute the feature to puppet-magnum. | 08:38 |
| zigo | By all means, yes, give me some examples ! :) | 08:38 |
| zigo | jrosser: Though, isn't there is a way to setup my own etcd cluster, so that I don't need an external discovery URL ? | 08:38 |
| zigo | What would be the etcd discovery URL then? | 08:39 |
| jrosser | http://paste.openstack.org/show/752219/ | 08:39 |
| jrosser | zigo: i wish i had an answer to that - i've had to make specific provision for magnum conductor to hit the discovery URL, and i really don't like having to do that | 08:40 |
| jrosser | i'm not sure why all of that can't be done on the deployed master for example, there must be a good reason | 08:40 |
| zigo | Thanks. | 08:40 |
| zigo | This definitively is very helpful to me ! :) | 08:40 |
| jrosser | the no_proxy is for my internal VIP endpoint | 08:41 |
| zigo | Your what? | 08:41 |
| zigo | Oh ... | 08:41 |
| zigo | To tell: my API VIP IP doesn't need a proxy. | 08:41 |
| zigo | Right? | 08:41 |
| jrosser | loadbalancer for all the internalURL things | 08:41 |
| zigo | Got it. | 08:42 |
| jrosser | yes and i don't want internal traffic to go to the proxy | 08:42 |
| zigo | It wouldn't work for me without it either, so, thanks. | 08:42 |
| zigo | I'll try similar setup, but using an env file, as I wrote. | 08:42 |
| jrosser | no worries - glad it's not just me fighting with these things | 08:42 |
| jrosser | you can test it with cat /proc/<pid>/environ after you adjust the settings to make sure the process has the environment you expect | 08:44 |
| *** ramishra has quit IRC | 08:50 | |
| *** ramishra has joined #openstack-containers | 08:51 | |
| *** strigazi has joined #openstack-containers | 09:04 | |
| strigazi | flwang: ping | 09:05 |
| *** ykarel|lunch is now known as ykarel | 09:11 | |
| *** flwang1 has joined #openstack-containers | 09:11 | |
| flwang1 | strigazi: around? | 09:11 |
| strigazi | yes, I pinged you to flwang without '1' | 09:12 |
| flwang1 | strigazi: that's my workstation ;) | 09:12 |
| flwang1 | welcome back | 09:13 |
| flwang1 | strigazi: do you want to discuss the rolling upgrade patch? | 09:14 |
| strigazi | kubecon, was really cool, OS Austoi size | 09:14 |
| strigazi | kubecon, was really cool, OS Austin size | 09:14 |
| flwang1 | strigazi: any chance you will be in shanghai next month? or anyone from your team? | 09:14 |
| flwang1 | i will be in shanghai next month | 09:14 |
| strigazi | give me a sec to push, something in our gerrit | 09:15 |
| flwang1 | strigazi: based on your patch, i have made almost everything working now | 09:16 |
| strigazi | ok, so I don't push | 09:16 |
| flwang1 | strigazi: no, you can just push on your current patch | 09:16 |
| strigazi | so it is done? | 09:16 |
| flwang1 | so we can compare | 09:16 |
| flwang1 | the main part i was adding is the drain and make it work for master | 09:17 |
| strigazi | Is something missing? | 09:18 |
| openstackgerrit | Feilong Wang proposed openstack/magnum master: [DO NOT MERGE][k8s][fedora atomic] Rolling upgrade support https://review.opendev.org/660851 | 09:18 |
| flwang1 | https://review.opendev.org/#/c/660851/3/magnum/drivers/common/templates/kubernetes/fragments/upgrade-kubernetes.sh | 09:18 |
| flwang1 | this is the main change i did | 09:18 |
| flwang1 | other than that, i just follow your way to add the $ssh_cmd for those master node scripts | 09:18 |
| flwang1 | there are still some minor parts may need polish, but based on my testing, it works overall | 09:19 |
| flwang1 | the only thing, i haven't figured out is how to do drain when doing operating system upgrade :( | 09:20 |
| strigazi | ok. I can test and we merge if it works? | 09:20 |
| flwang1 | strigazi: yep, feel free test it | 09:20 |
| strigazi | do we need this then ? 514960 | 09:21 |
| strigazi | I think not, it is included in 660851 | 09:22 |
| flwang1 | i can push all my changes to 514960, i don't want to take the credit, it's all based on your awesome work :) | 09:22 |
| strigazi | it is ok. :) What do we miss? the OS upgrade/drain issue? | 09:23 |
| flwang1 | for OS upgrade, there is no chance to call 'kubectl drain' to drain the node, as a result, which may cause downtime for user services | 09:24 |
| strigazi | we can merge like this as a first step. no? | 09:25 |
| flwang1 | strigazi: sure, we can | 09:27 |
| flwang1 | i will submit another document patch to document this as a limitation | 09:27 |
| flwang1 | strigazi: so if you don't mind, we can work together on 660851 and get it done | 09:30 |
| strigazi | I'll test and we work on it on gerrti | 09:31 |
| flwang1 | strigazi: cooool | 09:31 |
| flwang1 | i'm standing by to address any comments | 09:31 |
| strigazi | cool | 09:32 |
| strigazi | initially, I moved ssh_cmd="ssh -F /srv/magnum/.ssh/config root@localhost" to start_heat_agent as a script so that we don't copy around the same thing | 09:32 |
| flwang1 | strigazi: that's a good idea, do yo mean the 'start_container_agent.sh'? | 09:36 |
| strigazi | yrs | 09:37 |
| strigazi | yes | 09:37 |
| flwang1 | cool | 09:39 |
| flwang1 | btw, one thing i haven't tested is, if any other heat stack update will also trigger the upgrade_kubernetes script? so I just added the check if the new_kube_tag equals the old one | 09:41 |
| strigazi | makes sense | 09:45 |
| flwang1 | strigazi: can we merge the api patch now? | 09:46 |
| strigazi | without tests? | 09:47 |
| flwang1 | as long as the functional patch working for you | 09:47 |
| strigazi | ok | 09:47 |
| flwang1 | another minor change we may need is, wait a second or check if all the pods has been successfully evicted, before doing the atomic commands | 09:50 |
| flwang1 | wait a second == give enough headroom | 09:50 |
| *** threestrands has quit IRC | 09:54 | |
| strigazi | sounds like an optimization, we can do it | 09:54 |
| brtknr | flwang1: jakeyip: sorry i was travelling yesterday during the time of the meeting | 09:57 |
| flwang1 | brtknr: no worries | 10:00 |
| flwang1 | brtknr: just wanna asking if you're still working on the client patch about hidden | 10:00 |
| brtknr | yes, I am trying to find out what the pep8 error its complaining about is | 10:00 |
| flwang1 | strigazi: can you do a quick review for https://review.opendev.org/#/c/648317/ ? | 10:00 |
| flwang1 | brtknr: cool, cheers | 10:01 |
| brtknr | http://logs.openstack.org/98/659298/5/check/openstack-tox-pep8/3eb5845/ara-report/ | 10:03 |
| brtknr | a lot of these pep8 warnings dont even seem related to my patch | 10:03 |
| strigazi | flwang1: I'll test | 10:04 |
| flwang1 | brtknr: seems we need the bandit fix for client | 10:05 |
| flwang1 | brtknr: like this https://review.opendev.org/#/c/659027/ | 10:05 |
| * strigazi is going for lunch | 10:07 | |
| brtknr | flwang1: also please take a look here... I dont understand why cover test is failing: http://logs.openstack.org/98/659298/5/check/openstack-tox-cover/4bd9667/ara-report/result/a9d42a93-2b88-4182-b787-30964549ce9e/ | 10:16 |
| brtknr | thanks for the bandit fix, it works | 10:16 |
| flwang1 | brtknr: sure | 10:16 |
| brtknr | flwang1: any insights? | 10:25 |
| jakeyip | hi brtknr, I can help take a look too | 10:28 |
| brtknr | thanks jakeyip | 10:28 |
| *** dave-mccowan has joined #openstack-containers | 10:31 | |
| jakeyip | brtknr: works after blacklisting bandit | 10:35 |
| jakeyip | I don't see a patch for bandit, you want me to send that up? | 10:36 |
| openstackgerrit | Bharat Kunwar proposed openstack/python-magnumclient stable/queens: Conditional hidden arg for backward compatibility https://review.opendev.org/661937 | 10:37 |
| *** dave-mccowan has quit IRC | 10:37 | |
| openstackgerrit | Bharat Kunwar proposed openstack/python-magnumclient master: Conditional hidden arg for backward compatibility https://review.opendev.org/659298 | 10:41 |
| brtknr | yeah works for me too | 10:42 |
| brtknr | its the cover test which is failing | 10:42 |
| jakeyip | yeah sorry I'm testing cover now | 10:42 |
| jakeyip | looks like something changed in coverage output, parsing of that is failing. | 10:57 |
| jakeyip | also seems like a few projects doing away with cover.sh and changing to a standard tox cover | 10:58 |
| jakeyip | https://review.opendev.org/#/c/601144/1/tox.ini | 10:58 |
| *** janki has quit IRC | 11:11 | |
| *** mkuf_ has quit IRC | 11:24 | |
| *** mkuf_ has joined #openstack-containers | 11:25 | |
| *** mkuf_ has quit IRC | 11:28 | |
| *** mkuf has joined #openstack-containers | 11:28 | |
| *** ykarel_ has joined #openstack-containers | 11:39 | |
| *** ykarel has quit IRC | 11:43 | |
| *** ykarel_ is now known as ykarel | 11:55 | |
| *** nfakhir_ has quit IRC | 11:58 | |
| *** goldyfruit has joined #openstack-containers | 11:59 | |
| *** udesale has joined #openstack-containers | 12:25 | |
| *** ykarel is now known as ykarel|afk | 13:16 | |
| *** slivik has joined #openstack-containers | 13:43 | |
| slivik | Hello, we are developing new magnum driver for k8s based on ubuntu, I would like to ask whether there are some constraints or rules in order to merge it into the project | 13:45 |
| *** ykarel|afk has quit IRC | 14:07 | |
| *** itlinux has joined #openstack-containers | 14:11 | |
| *** ricolin has quit IRC | 14:20 | |
| *** absubram has joined #openstack-containers | 14:20 | |
| *** absubram has quit IRC | 14:26 | |
| *** ricolin has joined #openstack-containers | 14:26 | |
| *** ivve has quit IRC | 14:29 | |
| *** absubram has joined #openstack-containers | 14:30 | |
| *** lpetrut has quit IRC | 14:30 | |
| *** ykarel|afk has joined #openstack-containers | 14:35 | |
| *** ykarel|afk is now known as ykarel | 14:35 | |
| *** itlinux has quit IRC | 14:41 | |
| *** slivik has quit IRC | 15:19 | |
| *** altlogbot_3 has quit IRC | 15:35 | |
| *** altlogbot_0 has joined #openstack-containers | 15:35 | |
| *** irclogbot_0 has quit IRC | 15:36 | |
| *** irclogbot_0 has joined #openstack-containers | 15:37 | |
| *** itlinux has joined #openstack-containers | 15:54 | |
| *** ykarel is now known as ykarel|away | 15:59 | |
| *** ivve has joined #openstack-containers | 16:02 | |
| *** lpetrut has joined #openstack-containers | 16:06 | |
| *** absubram has quit IRC | 16:06 | |
| *** lpetrut has quit IRC | 16:14 | |
| *** ykarel|away has quit IRC | 16:35 | |
| *** dave-mccowan has joined #openstack-containers | 16:39 | |
| *** ricolin has quit IRC | 16:42 | |
| *** ramishra has quit IRC | 16:57 | |
| *** absubram has joined #openstack-containers | 17:44 | |
| *** inovas has quit IRC | 18:36 | |
| *** inovas has joined #openstack-containers | 18:37 | |
| *** flwang1 has quit IRC | 18:44 | |
| *** sapd1_x has joined #openstack-containers | 19:06 | |
| *** hongbin has joined #openstack-containers | 20:00 | |
| *** sapd1_x has quit IRC | 20:41 | |
| *** itlinux has quit IRC | 21:05 | |
| *** itlinux has joined #openstack-containers | 21:13 | |
| *** pcaruana has quit IRC | 21:23 | |
| openstackgerrit | Lingxian Kong proposed openstack/magnum master: [k8s_fedora_atomic] Make calico devices unmanaged in NetworkManager config for master node https://review.opendev.org/662098 | 21:39 |
| *** udesale has quit IRC | 21:43 | |
| *** inovas has quit IRC | 22:13 | |
| *** inovas has joined #openstack-containers | 22:17 | |
| *** ianychoi has quit IRC | 22:32 | |
| *** ianychoi has joined #openstack-containers | 22:33 | |
| *** rcernin has joined #openstack-containers | 23:07 | |
| *** absubram has quit IRC | 23:18 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!