09:59:57 <strigazi> #startmeeting containers 09:59:57 <openstack> Meeting started Tue Apr 17 09:59:57 2018 UTC and is due to finish in 60 minutes. The chair is strigazi. Information about MeetBot at http://wiki.debian.org/MeetBot. 09:59:59 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 10:00:00 <strigazi> #topic Roll Call 10:00:01 <openstack> The meeting name has been set to 'containers' 10:00:47 <flwang1> o/ 10:01:11 <strigazi> hi flwang1 10:01:21 <slunkad> hi 10:01:24 <flwang1> good to see you 10:01:29 <flwang1> "see" 10:01:47 <strigazi> yeah see is relative :) 10:02:05 <strigazi> #topic Blueprints/Bugs/Ideas 10:02:41 <strigazi> From my side, 10:03:34 <strigazi> I pushed a patch for moving all SoftwareConfigs to SoftwareDeployments, I'll do the same for master 10:03:52 <strigazi> #link https://review.openstack.org/#/c/561858/ 10:03:57 <flwang1> strigazi: awesome, love it 10:04:36 <strigazi> the only not great part is the to install the syscontainer we need to be in the same fs 10:04:38 <ricolin> o/ 10:04:50 <strigazi> hi ricolin 10:05:00 <flwang1> ricolin: waves from NZ 10:05:01 <strigazi> so I did what ansible does... 10:05:05 <strigazi> ssh 10:05:27 <flwang1> hmm.... 10:05:30 <strigazi> https://review.openstack.org/#/c/561858/1/magnum/drivers/common/templates/kubernetes/fragments/configure-kubernetes-minion.sh@16 10:06:10 <flwang1> strigazi: i will pay attention on that 10:06:11 <strigazi> It must be either ssh or a systemd unit to do atomic install 10:07:27 <strigazi> alternatively we can use ansible, https://docs.ansible.com/ansible/2.5/modules/atomic_container_module.html 10:08:04 <strigazi> ansible is the same principle, it executes stuff over ssh 10:08:15 <flwang1> strigazi: i see 10:08:25 <strigazi> see here the config https://review.openstack.org/#/c/561858/1/magnum/drivers/common/templates/kubernetes/fragments/start-container-agent.sh 10:08:38 <flwang1> i mean when I review the patch, i will do more research about that 10:09:10 <brtknr> Hi all 10:09:52 <strigazi> The tricky point is that, atomic creates hard links from /var/lib/containers/atomic to ostree, hard link means, same fs, means no mount namespace 10:09:56 <strigazi> hi brtknr 10:10:11 <strigazi> Also from my side, 10:10:53 <strigazi> I tested the eventlet fix: https://github.com/eventlet/eventlet/commit/1d6d8924a9da6a0cb839b81e785f99b6ac219a0e for the periodic task 10:11:10 <strigazi> flwang1: we will have an eventlet release soon I guess 10:12:04 <strigazi> That's it from me, this week I owe, api-ref for upgrades and the rest of the functionality 10:12:07 <flwang1> strigazi: cool, is there any relationship between doug's patch with our issue? 10:12:25 <strigazi> flwang1: coincidence :) 10:12:33 <flwang1> strigazi: gotcha 10:12:56 <strigazi> flwang1: And I hope no troubles when we want to inc the eventlet version 10:13:16 <flwang1> strigazi: finger cross ;) 10:13:36 <strigazi> flwang1: do you want to go next? I saw authz and authn :) 10:13:45 <flwang1> strigazi: oh, yes 10:14:15 <flwang1> for my side, i'm testing the keystone authN/authZ, hopefully i can finish the patch this week 10:14:47 <flwang1> and internally, Catalyst Cloud is planning the production release for Magnum, so I need to get the calico patch done ASAP 10:15:00 <flwang1> strigazi: sorry for all the pushing ;) 10:15:36 <strigazi> flwang1: yeah, i found an issue with the tolerations to spawn on the master 10:15:59 <flwang1> strigazi: with my calico patch? 10:16:03 <strigazi> yes 10:16:21 <flwang1> can you add more details on my patch or we can discuss offline? 10:16:31 <brtknr> strigazi: sorry i didnt mean to explain kubelet to you, misunderstood the question 10:17:07 <strigazi> flwang1: ok 10:17:45 <flwang1> and i'm starting to work on the cluster monitoring 10:18:28 <flwang1> that's all 10:18:57 <strigazi> brtknr: np, we can talk about it in a bit 10:19:40 <strigazi> slunkad? you have something? 10:20:14 <slunkad> yes 10:20:52 <slunkad> I was testing ricolin patch, it seems to have broken something, still looking into it. ricolin have you tired it out? 10:21:20 <slunkad> https://review.openstack.org/#/c/557337/ 10:22:04 * strigazi will test it too 10:22:11 <slunkad> tried* 10:22:25 <strigazi> slunkad: what was the problem? 10:22:37 <strigazi> slunkad: What did you try? 10:22:39 <ricolin> slunkad yeah, I tested it against master 10:23:33 <slunkad> The broker has blocked the connection: connection blocked, see broker logs but I'm not sure yet that this is caused by the pacth, but happened after I applied the patch so was wondering if anyone tested it already 10:23:53 <slunkad> that was in the heat logs 10:23:56 <ricolin> If you can provide the test scenario that break, I can try that in my environment 10:24:59 <slunkad> I just tried to create a cluster after applying the changes, but if it worked for you I will check again maybe just something to do with my env 10:25:43 <ricolin> Let me know if any more information:) 10:25:51 <slunkad> sure! 10:26:32 <slunkad> other than that I think this patch is ready https://review.openstack.org/#/c/552099/ 10:26:50 <slunkad> the glossary I will take another look before we can merge it. 10:26:59 <strigazi> slunkad: cool 10:27:02 <strigazi> ricolin: 10:27:30 <strigazi> The trust id is in the db right? 10:27:51 <ricolin> Yes 10:27:58 <strigazi> After a stack update we should expect that it is changed 10:28:07 <strigazi> right? 10:28:41 <ricolin> Yep 10:28:54 <ricolin> That’s the scenario 10:29:08 <strigazi> ricolin: And this will happen everytime? 10:29:50 <strigazi> ricolin: User A creates stack, userB does stack update (trust rotate), userA does stack update (trust rotate) 10:30:02 <ricolin> If the provided user changed 10:30:43 <slunkad> so we just need to call stack update to renew trusts? 10:30:57 <strigazi> slunkad: fo the heat trusts ye 10:30:59 <strigazi> slunkad: fo the heat trusts yes 10:31:24 <ricolin> Yes 10:31:27 <strigazi> for magnum you need to generate a new trust and pass it to heat 10:31:57 <slunkad> I see 10:32:04 <strigazi> pass it to heat: do a stack update passing the new trust as a stack parameter 10:33:07 <ricolin> Use another user to update stack, Heat will generate the new trust and replace the old 10:37:35 <strigazi> slunkad: anything else? 10:37:51 <slunkad> no not for this topic 10:38:48 <strigazi> cool, thanks 10:39:57 <strigazi> brtknr: I tried your fix in two different nodes. in one it worked in the other no. 10:40:41 <strigazi> Can you do atomic host status in the node it worked for you? 10:41:12 <strigazi> @all I'm talking about this one: https://review.openstack.org/#/c/561605/1/magnum/drivers/common/templates/kubernetes/fragments/configure-kubernetes-minion.sh,unified 10:42:33 <strigazi> brtknr: are you still here? 10:42:48 <brtknr> strigazi: yes im here 10:46:57 <strigazi> brtknr: let's take https://review.openstack.org/#/c/561605 offline 10:46:59 <brtknr> strigazi: please discuss something else, i cant access the machine right now 10:47:11 <strigazi> @all anything else for the meeting? 10:47:15 <slunkad> has anyone seen this error before http://paste.openstack.org/show/719367/? 10:47:19 <strigazi> brtknr: np 10:48:05 <slunkad> I'm not sure which config file it is looking into for the auth_url 10:48:16 <slunkad> when running the k8s functional tests 10:48:35 <slunkad> (in pike) 10:49:48 <strigazi> slunkad: do you want to discuss it after the meeting? We don't have to log this 10:49:53 <slunkad> sure 10:50:17 <strigazi> Let's end it then, objections? 10:50:41 <strigazi> none :) See you 10:50:45 <strigazi> #endmeeting