15:01:37 #startmeeting openstack-helm 15:01:38 Meeting started Tue Aug 13 15:01:37 2019 UTC and is due to finish in 60 minutes. The chair is portdirect. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:01:39 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:01:41 The meeting name has been set to 'openstack_helm' 15:01:59 hey everyone, lets give it a few mins for people to turn up 15:02:06 o/ 15:02:12 the agenda for today is here: https://etherpad.openstack.org/p/openstack-helm-meeting-2019-08-13 15:02:16 o/ 15:02:18 o/ 15:02:23 o/ 15:02:33 o/ 15:02:37 o/ 15:02:43 o/ 15:03:09 hello 15:06:18 ok - lets go 15:06:26 #topic Periodic jobs still broken 15:06:31 srwilkers: i think this is yours? 15:06:39 yep, thats me 15:07:33 so the multinode periodic jobs have been busted for awhile, and the culprit is the ceph provisioner helm tests executed when we activate the secrets for a namespace 15:07:39 specifically, the cephfs provisioner tests 15:08:19 any idea why? 15:08:32 we see occasional successes when the jobs are executed on rackspace vms, but i'm not sure why 15:09:07 after some digging, it seems the kernel versions are the same, so i don't think it's issues related to not using the hwe kernel, as some individuals have suggested 15:09:29 and i didnt think that'd be the issue now that we use bionic nodes from nodepool anyway 15:10:20 any delta between providers? 15:10:34 i'm at a loss here honestly, and the requests i've made for digging more into it haven't been successful. as a result, i've proposed disabling the cephfs provisioner in the multinode jobs until we can establish confidence in the tests associated with it 15:10:43 not that i've seen so far 15:10:59 ok - i think in the short term this may be the best path forward then 15:11:12 you have a ps? 15:11:51 we still deploy the cephfs provisioner in the single node logging job, so changes are still gated. this just unblocks the functionality tested in our multinode jobs, as potential regressions in that functionality are likely being masked by this issue 15:11:59 yeah: https://review.opendev.org/#/c/674816/ 15:12:53 that's all i've got here 15:13:42 ok - thanks steve 15:14:06 ok - lets move on: 15:14:14 #topic OpenInfra summit 15:14:21 georgk: i think you put this? 15:14:32 yes, but it is really just a question 15:14:49 sorry if that was already discussed 15:14:58 sure - at this point i think its still very much up in the air from the cores ive spoken to 15:15:04 but I am just curious who of you is planning to be there 15:15:33 i will not be there 15:15:58 I'd certainly like to be, but the dice is yet to be rolled on that 15:16:13 georgk: will you be attending? 15:16:41 ok, thanks. yes, given the general situation I was expecting that it nmight be more tricky this time 15:16:50 portdirect: not entirely decided yet :-) 15:17:37 i think once we get a bit closer to the time, then we should certainly try and have some presence 15:18:07 i think jp and some others will be attending, and he has offered to provide any assitance he can 15:18:14 which is really appreciated 15:18:27 ok, thanks for the info 15:19:08 I will be there, as it's right at ShangHai 15:19:25 nice - thats awesome cheng1 15:20:22 actually while you both are here 15:20:23 welcome to ShangHai :) 15:20:35 #topic ovs-dpdk update 15:20:56 i've been working on testing georgk, cheng1 and deepaks patches 15:21:01 and they look pretty solid 15:21:11 woohoo 15:21:22 im still awaiting feedback from deepak that everything is functional 15:21:34 although I just pushed an update to address the comments 15:21:47 and once we have that, i think will be in a much better place to get this over the finish line 15:22:40 good to know. thank you. let me know if I can help 15:23:12 will do, thanks for all the effort you've both been putting in here, its been massive effort 15:23:21 and making great progress 15:24:14 #topic reviews 15:24:31 we have a few patchsets out there that could do with some review love atm 15:24:52 the 1st two should be fairly simple: 15:24:55 https://review.opendev.org/#/c/675125/ - Remove unused images from openstack-helm-infra (these now exist in openstack-helm-images) 15:24:55 https://review.opendev.org/#/c/675172/ - Expose Tiller http port for metrics in Minikube deployments 15:25:16 and then a great ps from the folks over at tf: 15:25:16 https://review.opendev.org/#/c/663390/ - Add TungstenFabric compatibility to charts 15:25:38 some chainges to htk that allows us to configure job backoffs: 15:25:39 https://review.opendev.org/#/c/672738/ - Enhance HTK Job Manifests to be more flexible (cliffparsons) 15:26:05 and some enhancements to the postgres backup and restore process: 15:26:05 https://review.opendev.org/#/c/655231/ - Postgresql backup/restore enhancements (cliffparsons) 15:26:13 please get eyes on them 15:26:39 thats all i have for this week, anything else we should be aware of/looking out for? 15:28:37 if not we can wrap up, and give people 30 mins back 15:29:46 ok - thanks folks! 15:29:52 see you all in #openstack-helm 15:29:55 #endmeeting