| *** jesusaur has quit IRC | 00:26 | |
| *** jesusaur has joined #zuul | 00:30 | |
| *** jesusaur has quit IRC | 00:31 | |
| *** jamielennox is now known as jamielennox|away | 01:01 | |
| *** jesusaur has joined #zuul | 01:15 | |
| *** jamielennox|away is now known as jamielennox | 01:19 | |
| *** jesusaur has quit IRC | 01:34 | |
| *** jesusaur has joined #zuul | 01:40 | |
| *** tflink has quit IRC | 03:21 | |
| *** tflink has joined #zuul | 03:55 | |
| *** jamielennox is now known as jamielennox|away | 04:14 | |
| *** jamielennox|away is now known as jamielennox | 04:22 | |
| *** tflink has quit IRC | 04:28 | |
| *** tflink has joined #zuul | 04:57 | |
| *** bhavik1 has joined #zuul | 05:33 | |
| *** bhavik1 has quit IRC | 05:36 | |
| *** isaacb has joined #zuul | 06:43 | |
| *** hashar has joined #zuul | 06:46 | |
| *** bhavik1 has joined #zuul | 06:49 | |
| *** bhavik1 has quit IRC | 06:53 | |
| *** rcarrillocruz has joined #zuul | 07:45 | |
| *** openstackgerrit has quit IRC | 08:33 | |
| *** rcarrillocruz has quit IRC | 09:57 | |
| *** openstackgerrit has joined #zuul | 11:39 | |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Silence command warnings https://review.openstack.org/448748 | 11:39 |
|---|---|---|
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Replace ssh-keyscan with host keys from node https://review.openstack.org/446785 | 11:43 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Add entry for ref-replication-scheduled event https://review.openstack.org/446251 | 11:43 |
| *** isaacb_ has joined #zuul | 12:01 | |
| *** isaacb has quit IRC | 12:04 | |
| openstackgerrit | David Shrewsbury proposed openstack-infra/nodepool feature/zuulv3: Add a test for multiple provider pools https://review.openstack.org/450183 | 12:07 |
| Shrews | jeblair: I took the liberty of creating a multi-pool test ^^^^ (mainly b/c I felt bad about asking for it), but it does point out an issue where you aren't setting the 'pool' attribute on new nodes. | 12:08 |
| *** isaacb_ has quit IRC | 12:33 | |
| *** isaacb_ has joined #zuul | 12:35 | |
| rbergeron | shrews: i am in the office today and tomorrow if you are bored and want to get in a car and say hello at some point :) | 14:35 |
| Shrews | rbergeron: neat! today is probably no good, especially with the late zuul meeting time. tomorrow maybe? | 14:36 |
| rbergeron | shrews: suuuuure, maybe in morningish + meeting + lunchtime? I get to sit in meetings tomorrow afternoon about open sourcing that one thing :) | 14:40 |
| rbergeron | i fly out at 6 or 7pm or something | 14:40 |
| openstackgerrit | Paul Belanger proposed openstack-infra/nodepool master: Add debian-jessie DIB for dsvm testing https://review.openstack.org/449938 | 15:01 |
| jeblair | Shrews: oh cool. i would have happily written one, but thanks! i'll go fix up the patch now. | 15:14 |
| openstackgerrit | Merged openstack-infra/nodepool feature/zuulv3: Remove SSH support from nodepool https://review.openstack.org/446683 | 15:22 |
| openstackgerrit | James E. Blair proposed openstack-infra/nodepool feature/zuulv3: Remove keypair from provider section https://review.openstack.org/447673 | 15:31 |
| openstackgerrit | James E. Blair proposed openstack-infra/nodepool feature/zuulv3: Remove legacy openstack settings from nodepool.yaml https://review.openstack.org/447647 | 15:31 |
| openstackgerrit | James E. Blair proposed openstack-infra/nodepool feature/zuulv3: Update nodepool config syntax https://review.openstack.org/448814 | 15:31 |
| openstackgerrit | James E. Blair proposed openstack-infra/nodepool feature/zuulv3: Update nodepool config syntax https://review.openstack.org/448814 | 15:32 |
| openstackgerrit | James E. Blair proposed openstack-infra/nodepool feature/zuulv3: Add a test for multiple provider pools https://review.openstack.org/450183 | 15:33 |
| openstackgerrit | James E. Blair proposed openstack-infra/nodepool feature/zuulv3: Remove api-timeout and provider.image-type https://review.openstack.org/449342 | 15:34 |
| openstackgerrit | James E. Blair proposed openstack-infra/nodepool feature/zuulv3: Remove ipv6-preferred and rely on interface_ip https://review.openstack.org/449705 | 15:34 |
| openstackgerrit | James E. Blair proposed openstack-infra/nodepool feature/zuulv3: Remove deprecated networks syntax https://review.openstack.org/449354 | 15:34 |
| openstackgerrit | James E. Blair proposed openstack-infra/nodepool feature/zuulv3: Update docs for configuration syntax change https://review.openstack.org/449339 | 15:34 |
| openstackgerrit | James E. Blair proposed openstack-infra/nodepool feature/zuulv3: Exercise statsd in tests and fix https://review.openstack.org/449737 | 15:34 |
| openstackgerrit | James E. Blair proposed openstack-infra/nodepool feature/zuulv3: Remove mention of non-clouds.yaml from docs https://review.openstack.org/449781 | 15:34 |
| jeblair | mordred: did you start on the shade get az list thing? | 15:35 |
| mordred | jeblair: no! I knew there was a thing I was suppposed to be doing in shade - I will do that as soon as I finish the patch I'm working on now | 15:41 |
| jeblair | mordred: cool, thx | 15:41 |
| *** isaacb_ has quit IRC | 15:43 | |
| *** hashar is now known as hasharAway | 15:57 | |
| *** tobiash has joined #zuul | 16:02 | |
| *** tobiash has quit IRC | 16:02 | |
| *** MapspaM is now known as SpamapS | 16:08 | |
| *** hashar has joined #zuul | 16:10 | |
| *** cinerama has joined #zuul | 16:12 | |
| openstackgerrit | Merged openstack-infra/nodepool feature/zuulv3: Refactor nodepool apps into base app https://review.openstack.org/448395 | 16:21 |
| openstackgerrit | Merged openstack-infra/nodepool feature/zuulv3: Add check for valid zk attribute before disconnect https://review.openstack.org/449110 | 16:22 |
| openstackgerrit | Merged openstack-infra/nodepool feature/zuulv3: Do not require secure file for nodepoold https://review.openstack.org/449118 | 16:23 |
| openstackgerrit | Merged openstack-infra/nodepool feature/zuulv3: Docs: Remove "jenkins" reference https://review.openstack.org/449129 | 16:23 |
| openstackgerrit | Merged openstack-infra/nodepool feature/zuulv3: Docs: Clarify secure file usage https://review.openstack.org/449130 | 16:23 |
| jeblair | harlowja: i responded to your question on https://github.com/python-zk/kazoo/pull/419 | 16:28 |
| jeblair | harlowja: i also pushed up a new commit with an improvement (detailed in my response) | 16:28 |
| openstackgerrit | James E. Blair proposed openstack-infra/nodepool feature/zuulv3: Remove api-timeout and provider.image-type https://review.openstack.org/449342 | 16:34 |
| openstackgerrit | James E. Blair proposed openstack-infra/nodepool feature/zuulv3: Remove ipv6-preferred and rely on interface_ip https://review.openstack.org/449705 | 16:34 |
| openstackgerrit | James E. Blair proposed openstack-infra/nodepool feature/zuulv3: Remove deprecated networks syntax https://review.openstack.org/449354 | 16:34 |
| openstackgerrit | James E. Blair proposed openstack-infra/nodepool feature/zuulv3: Update docs for configuration syntax change https://review.openstack.org/449339 | 16:34 |
| openstackgerrit | James E. Blair proposed openstack-infra/nodepool feature/zuulv3: Update nodepool config syntax https://review.openstack.org/448814 | 16:34 |
| openstackgerrit | James E. Blair proposed openstack-infra/nodepool feature/zuulv3: Exercise statsd in tests and fix https://review.openstack.org/449737 | 16:34 |
| openstackgerrit | James E. Blair proposed openstack-infra/nodepool feature/zuulv3: Remove mention of non-clouds.yaml from docs https://review.openstack.org/449781 | 16:34 |
| openstackgerrit | James E. Blair proposed openstack-infra/nodepool feature/zuulv3: Add a test for multiple provider pools https://review.openstack.org/450183 | 16:34 |
| openstackgerrit | James E. Blair proposed openstack-infra/nodepool feature/zuulv3: Add a test for multiple provider pools https://review.openstack.org/450183 | 16:37 |
| SpamapS | whoa | 16:37 |
| SpamapS | go jim go | 16:37 |
| jeblair | mostly rebases and minor fixes; it's a big stack :/ | 16:38 |
| openstackgerrit | James E. Blair proposed openstack-infra/nodepool feature/zuulv3: Remove api-timeout and provider.image-type https://review.openstack.org/449342 | 16:39 |
| openstackgerrit | James E. Blair proposed openstack-infra/nodepool feature/zuulv3: Remove ipv6-preferred and rely on interface_ip https://review.openstack.org/449705 | 16:39 |
| openstackgerrit | James E. Blair proposed openstack-infra/nodepool feature/zuulv3: Remove deprecated networks syntax https://review.openstack.org/449354 | 16:39 |
| openstackgerrit | James E. Blair proposed openstack-infra/nodepool feature/zuulv3: Update docs for configuration syntax change https://review.openstack.org/449339 | 16:39 |
| openstackgerrit | James E. Blair proposed openstack-infra/nodepool feature/zuulv3: Exercise statsd in tests and fix https://review.openstack.org/449737 | 16:39 |
| openstackgerrit | James E. Blair proposed openstack-infra/nodepool feature/zuulv3: Remove mention of non-clouds.yaml from docs https://review.openstack.org/449781 | 16:39 |
| *** hashar has quit IRC | 16:54 | |
| jeblair | mordred: http://git.openstack.org/cgit/openstack-infra/nodepool/tree/nodepool/nodepool.py#n1325 | 16:55 |
| jeblair | mordred: essentially, if the user doesn't specify a list of azs, instead of setting None there to let nova decide, we want to pick one from the full list | 16:57 |
| jeblair | mordred: (i know that doesn't sound like much of a difference, but if we choose the az instead of nova, it means we know the az early enough to request the same az for the rest of the nodes which are launched) | 16:58 |
| jeblair | mordred: (if we wait for nova to choose, it's too late for us to pick the same az for the other nodes) | 16:58 |
| mordred | jeblair: ++ | 16:58 |
| jeblair | mordred, Shrews: my guess is we want to put something on the providermanager to cache that (like flavors, networks, etc) | 17:00 |
| mordred | jeblair: hrm. I guess so | 17:06 |
| mordred | I'm sad we're not consuming shade caching yet, as we _also_ have caching for it there - but such is life :) | 17:06 |
| openstackgerrit | Monty Taylor proposed openstack-infra/nodepool feature/zuulv3: Fetch list of AZs from nova if it's not configured https://review.openstack.org/450345 | 17:12 |
| SpamapS | is there some big reason we aren't just using shade's caching? | 17:12 |
| mordred | jeblair: that won't pass tests right now, because I bumped the requirement for shade in the patch | 17:12 |
| mordred | SpamapS: we haven't dug in to it recently - last time we tried there was an issue, and it hasn't come back around to figuring out what was up with that | 17:13 |
| mordred | SpamapS: other things have been at the top of the list | 17:14 |
| mordred | :) | 17:14 |
| SpamapS | somanythings | 17:16 |
| openstackgerrit | James E. Blair proposed openstack-infra/nodepool feature/zuulv3: Remove deprecated networks syntax https://review.openstack.org/449354 | 17:18 |
| *** harlowja has quit IRC | 17:18 | |
| openstackgerrit | James E. Blair proposed openstack-infra/nodepool feature/zuulv3: Remove ipv6-preferred and rely on interface_ip https://review.openstack.org/449705 | 17:19 |
| openstackgerrit | James E. Blair proposed openstack-infra/nodepool feature/zuulv3: Exercise statsd in tests and fix https://review.openstack.org/449737 | 17:19 |
| openstackgerrit | James E. Blair proposed openstack-infra/nodepool feature/zuulv3: Remove mention of non-clouds.yaml from docs https://review.openstack.org/449781 | 17:19 |
| jeblair | mordred: jhesketh had a question on https://review.openstack.org/448750 | 17:20 |
| Shrews | mordred: thanks for that AZ patch | 17:23 |
| *** harlowja has joined #zuul | 17:24 | |
| mordred | jeblair: that's a good question. I believe the answer is "no" | 17:25 |
| mordred | although if we decided that was important, we could likely write in a special case to our callback plugins | 17:25 |
| SpamapS | jeblair: After seeing pabelanger's comments on the security spec about nodepool-builder, I had a thought. What if we just do this bubblewrap work in nodepool as "container enablement" ? | 17:26 |
| jeblair | mordred: yeah, i think i'm okay with the "let's see what the errors look like" approach. | 17:26 |
| SpamapS | We still end up with 'ansible running ansible', but I wonder if it's a shorter path since we git rid of the duplication efforts around managing images and contexts. | 17:27 |
| mordred | SpamapS: I don't fully understand what you mean by "container enablement" | 17:27 |
| jeblair | SpamapS: you mean your earlier suggestion of having zuul request a container from nodepool from which to run ansible? | 17:27 |
| SpamapS | jeblair: Right, instead of requesting a VM based node, Zuul would request a container. | 17:28 |
| SpamapS | And then do the Ansible execution there. | 17:28 |
| SpamapS | So we don't get the added cost of extra vms for every concurrent job. | 17:28 |
| jeblair | SpamapS: yeah, i think that's a good idea that we should look into in the future, but i think that adding container support to nodepool is a major endeavor. i'd guess add 1-2 months onto our effort. i think bubblewrap on the launcher is a much shorter path. | 17:29 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Create zuul.executor.hostname ansible variable https://review.openstack.org/448820 | 17:29 |
| SpamapS | jeblair: I hope you're right. I don't have a good estimate on how long the work will take, but I do believe it's not something to continue investing in once in place. | 17:32 |
| Shrews | jeblair: i had a -1 on 449339. not sure if you missed the comments in the rebase bomb | 17:35 |
| openstackgerrit | Merged openstack-infra/nodepool feature/zuulv3: Remove legacy openstack settings from nodepool.yaml https://review.openstack.org/447647 | 17:35 |
| openstackgerrit | Merged openstack-infra/nodepool feature/zuulv3: Remove keypair from provider section https://review.openstack.org/447673 | 17:35 |
| jeblair | SpamapS: well, investment as far as features go -- sure -- it's not really a feature of the system, it should really be as invisible as possible to operators and users. investment in maintenance? i think we will need to keep it working for some time, possibly indefinitely. growing an explicit dependency on a container technology in nodepool is a big change. we've agreed having it as an option in nodepool would be good. but requiring it ... | 17:37 |
| jeblair | ... would place a larger burden on operators. | 17:37 |
| jeblair | Shrews: thanks | 17:38 |
| SpamapS | jeblair: Investment in features is definitely what I mean. :) | 17:39 |
| SpamapS | These things tend to grow legs. | 17:39 |
| pabelanger | look at all the new merges :) | 17:39 |
| jeblair | SpamapS: then yeah, i agree this is basically an "implementation detail" as far as that goes. :) | 17:39 |
| pabelanger | need to catch up and see where we are | 17:39 |
| SpamapS | Like if we initially run w/o a MAC.. and we're comfortable with the security implications of that, we probably just need to live with that until we get container support in nodepool. | 17:39 |
| SpamapS | (hypothetical.. Rob Clark suggests we'll want a MAC) | 17:40 |
| * clarkb wonders why zuul security eg MAC is related to nodepool supporting containers (whcih it does today via nova flavor choice fwiw)) | 17:40 | |
| SpamapS | clarkb: that's assuming nova has a container flavor. ;) | 17:40 |
| SpamapS | some do I'm sure | 17:41 |
| SpamapS | clarkb: perhaps read the spec? there's a vibrant discussion in patchset 8 (I'll be pushing #9 shortly) 444495 | 17:41 |
| SpamapS | we're all wrestling with our overactive engineer desire to practice DRY while also needing to get things done. :) | 17:42 |
| SpamapS | or at least.. I am. :) | 17:44 |
| clarkb | I think the trap to avoid here is conflating orthogonal needs/features even if there is some overlap because doing so greatly increases the time to getting anything working (just look at how long it took us to get off of snapshot images) | 17:44 |
| pabelanger | The big thing for me, was having nb01 / nb02 build the container. So we didn't have to install and manage disk-image-builder on zuul-executor systems | 17:46 |
| SpamapS | If we can spin up cheap throw-away containers for running untrusted playbooks, it negates the need for the executor security enhancements. | 17:46 |
| SpamapS | And assuming people start using Zuul v3, they're all going to want containers. :) | 17:47 |
| clarkb | SpamapS: but only if you run them in non shared kernels that are also throwaway right? (eg just run a VM?) | 17:47 |
| SpamapS | s/people/more people/ | 17:47 |
| Shrews | pabelanger: yeah, lots of merging (and conflicting). which worries me because nodepool-launcher has been running like a champ the last few days :) | 17:47 |
| pabelanger | Shrews: :) I'm cleaning up our nodepoolv3.yaml file right now | 17:47 |
| SpamapS | clarkb: They still have to be secured regardless of where they run. But that security work is already happening in the container world and is more intense than what we're suggesting. | 17:47 |
| clarkb | SpamapS: or you are back to run a MAC | 17:47 |
| SpamapS | There's no other security model for containers. | 17:48 |
| clarkb | I mean proper selinux | 17:48 |
| clarkb | which aiui docker and friends don't do | 17:48 |
| SpamapS | docker does | 17:48 |
| SpamapS | very well | 17:48 |
| SpamapS | every docker container gets its own special SELinux context and you can basically throw that context into the "you get NOTHING" | 17:49 |
| SpamapS | bin | 17:49 |
| SpamapS | http://www.projectatomic.io/docs/docker-and-selinux/ <-- easy read on the subject | 17:49 |
| pabelanger | I still think zuul-executor should call bubblewrap directly for now, otherwise, that seem like lot of work to build out a container thing (ks8?) on the control plane | 17:49 |
| SpamapS | pabelanger: me too | 17:49 |
| SpamapS | I just keep having to shake off the "but nodepool does that too.." twitch | 17:50 |
| SpamapS | I think the eventual answer _is_ going to be a k8s of executors. | 17:50 |
| rbergeron | spamaps: is that *all* distros or just the red hat family which requires a special / forked docker versiom? | 17:50 |
| rbergeron | because $yaydrama | 17:50 |
| SpamapS | rbergeron: I thought it was any docker built with --enable-selinux .. but maybe I'm being naive. | 17:51 |
| SpamapS | I'm not focusing on that. :) | 17:52 |
| SpamapS | But, when we can circle back and add container support to nodepool, I think the way it happens is we teach nodepool to talk to k8s (and scale it with nodes from clouds?) and when somebody asks for container node types we give them containers and run playbooks inside said containers. | 17:54 |
| rbergeron | yeah -- i just wonder if the context ofvthat article makes assumptions (atomic amd all) which may not be universally truthy | 17:54 |
| clarkb | SpamapS: just for the record you can totally do that today if you need ti :) | 17:54 |
| SpamapS | rbergeron: I sense you have some experience in this realm and I defer to your pain. | 17:55 |
| SpamapS | clarkb: where would the containers run? | 17:55 |
| clarkb | SpamapS: wherever the nova scheduler puts them | 17:55 |
| SpamapS | what nova? | 17:55 |
| SpamapS | you have access to a nova with container flavors? | 17:55 |
| clarkb | the one your nodepool is talking to | 17:55 |
| SpamapS | I don't | 17:56 |
| clarkb | SpamapS: I;m not actually sure (I haven't listed all the flavors in the clouds to check) | 17:56 |
| SpamapS | I believe there's an IBM BlueMix container service that is a nova-docker | 17:56 |
| rbergeron | spamaps: all i know is i tried installing opebshift once on top of a standard opebstack deployment and woubd up discovering i needed thecred hat version of docker, which was not the one ibstalled, and i learned many things, lol | 17:56 |
| SpamapS | that one can sign up for as a beta. I know it's possible. But it's not actually useful for Infra's clouds. | 17:57 |
| jeblair | while it's true that we want zuul v3 to be friendly to users who are not openstack-infra, i also think that if we construct something that openstack-infra can't run, we have probably failed. | 17:57 |
| SpamapS | jeblair: s/probably/mightily/ :) | 17:57 |
| jeblair | SpamapS: it's possible i have a fondness for understatement. ;) | 17:58 |
| clarkb | SpamapS: I just want to clarify that because there is and has bee na lot of fud especially in third party ci land of "nodepool can't do x" | 17:58 |
| jeblair | maybe | 17:58 |
| SpamapS | Like flexing after you break the door handle to your car. It's not.. really helping. ;) | 17:58 |
| rbergeron | spamaps: but -- ppl are working on improved kube modules right now (https://github.com/ansible/ansible-kubernetes-modules ) and i think things will be fine when we get to that point in time. | 17:58 |
| clarkb | and it absolutely can... you just have to work with the api that is currently supported, nova | 17:58 |
| clarkb | (this was a big deal for third party ci and baremetal for a while) | 17:58 |
| SpamapS | rbergeron: Oh neat! | 17:58 |
| SpamapS | that might come in handy around July-ish :) | 17:59 |
| rbergeron | i suspect the shade of kubernetes will come to help with the variations btwn kube implementations / providers (since everyone now has their own e terprise k8s versions) | 17:59 |
| SpamapS | Just guessing that's around the time we get good at running VM based nodepool and try to scale BonnyCI up and realize we need containers nao. | 17:59 |
| rbergeron | (something something, why bother learning from other projects...) | 17:59 |
| SpamapS | clarkb: respect, very true. However, if I, the user of Zuul, can't add the pieces I need to all computer-providing-clouds, from outside, as a user, it's a much less useful thing. | 18:00 |
| SpamapS | are kubernetes's diverging? | 18:01 |
| SpamapS | that would be... ironic | 18:01 |
| rbergeron | im sure some are opinionated ;) | 18:02 |
| rbergeron | and god help us when kubevirt is a thing (launch vms from a kubernetes pod / cluster / whatever it is called) | 18:03 |
| rbergeron | *old lady shakes fist at cloud* | 18:03 |
| SpamapS | It's the Go rewrite of Nova we've all been resisting writing ourselves. | 18:03 |
| jeblair | SpamapS: regarding bubblewrap -- why create a chroot image vs bind-mount /usr ? | 18:04 |
| Shrews | rbergeron: i think you just broke my brain with that foreshadowing | 18:04 |
| SpamapS | jeblair: to take sharp objects away from attackers. | 18:04 |
| clarkb | rbergeron: k8s already has a vm option | 18:04 |
| SpamapS | probably worth noting in spec | 18:04 |
| rbergeron | clarkb: ah, i thought it was still in wip | 18:05 |
| clarkb | rbergeron: I think it only supports kvm but ya you can tell it to make "containers" that are actually VMs | 18:05 |
| SpamapS | Not sure how much mileage we get out of that since we still have to give them python though. | 18:05 |
| * SpamapS furiously battles his internal desire to rewrite Ansible in rust again. | 18:05 | |
| SpamapS | battles again, not rewrites again. | 18:05 |
| jeblair | SpamapS: ok. just wondering whether it's worth the extra complexity. or even whether bind mount could be an interim step with chroot image as a followup. | 18:05 |
| SpamapS | jeblair: yeah now that you ask.. it's entirely possible it's not that much of a win | 18:06 |
| SpamapS | jeblair: because we still have to give them python.. rsync.. internet access.. so.. yeah. | 18:06 |
| rbergeron | clarkb: ah. not sure if thats same or diff from https://github.com/kubevirt/kubevirt/blob/master/README.md | 18:06 |
| jeblair | SpamapS: like: "python -c 'download binary from internet and run it'" | 18:07 |
| SpamapS | or just python -c "exploit_stuff()" | 18:09 |
| SpamapS | Not much you can't do w/ python | 18:09 |
| clarkb | rbergeron: I'm trying to find it I thought it was in kubernets itself as a container deployment option but my googling is failing me | 18:09 |
| SpamapS | We could mount noexec and nosuid and probably get some protection from 'download binary' but we still have to let the python binary run python code. | 18:10 |
| SpamapS | I guess we could write an ansible-playbook that just shoves its CLI args into a local socket, teach ansible-playbook to read its args from said socket, and rm python once that has spawned. But zomg is this realy helping? | 18:11 |
| SpamapS | jeblair: yeah bind mount /usr is sounding more practical. | 18:11 |
| jeblair | ya | 18:11 |
| SpamapS | going to revise spec | 18:12 |
| * SpamapS once again resists urge to add "Rewrite Ansible in Rust" to Alternatives section | 18:12 | |
| * jeblair imagines a truly infinite array of alternatives | 18:13 | |
| jeblair | "Go to the movies" "Take a nap" | 18:13 |
| Shrews | pabelanger: Any chance you take a peek at the remaining np doc changes today? https://review.openstack.org/#/q/status:open+project:openstack-infra/nodepool+branch:feature/zuulv3+topic:docs | 18:16 |
| rbergeron | spamaps: im sure the rust folks would like that, since they use ansible to manage rust's infra ;) | 18:17 |
| SpamapS | jeblair: "Play Peter Molyneux's landmark 1989 game, Populous, in an Amiga emulator", is my favorite. | 18:18 |
| rbergeron | i generally recommend "think different / moar future" rather than think parallel ;) | 18:18 |
| pabelanger | Shrews: yes, looking at the stack now | 18:18 |
| SpamapS | rbergeron: I like the idea of compiling modules into little native binaries locally, but... python does make a wee bit more sense. :) | 18:19 |
| * mordred drives friend to airport and returns having missed all sorts of fun conversation! | 18:25 | |
| Shrews | pabelanger: looking at it, i suspect that my docs changes may conflict with jeblair's stack of changes. it would be easier for me to rebase, so let's get his changes merged first. | 18:25 |
| pabelanger | Shrews: sure | 18:25 |
| pabelanger | I think jeblair stack is ready, for the most part | 18:26 |
| Shrews | jeblair: do you want anyone else to review https://review.openstack.org/448814 ? | 18:26 |
| *** hasharAway is now known as hashar | 18:27 | |
| jeblair | Shrews: i think since it didn't really diverge from the email, we're probably good. | 18:28 |
| openstackgerrit | Merged openstack-infra/nodepool feature/zuulv3: Update nodepool config syntax https://review.openstack.org/448814 | 18:33 |
| openstackgerrit | Merged openstack-infra/nodepool feature/zuulv3: Add a test for multiple provider pools https://review.openstack.org/450183 | 18:33 |
| openstackgerrit | James E. Blair proposed openstack-infra/nodepool feature/zuulv3: Remove api-timeout and provider.image-type https://review.openstack.org/449342 | 18:35 |
| openstackgerrit | James E. Blair proposed openstack-infra/nodepool feature/zuulv3: Remove ipv6-preferred and rely on interface_ip https://review.openstack.org/449705 | 18:35 |
| openstackgerrit | James E. Blair proposed openstack-infra/nodepool feature/zuulv3: Remove deprecated networks syntax https://review.openstack.org/449354 | 18:35 |
| openstackgerrit | James E. Blair proposed openstack-infra/nodepool feature/zuulv3: Update docs for configuration syntax change https://review.openstack.org/449339 | 18:36 |
| openstackgerrit | James E. Blair proposed openstack-infra/nodepool feature/zuulv3: Exercise statsd in tests and fix https://review.openstack.org/449737 | 18:36 |
| openstackgerrit | James E. Blair proposed openstack-infra/nodepool feature/zuulv3: Remove mention of non-clouds.yaml from docs https://review.openstack.org/449781 | 18:36 |
| jeblair | Shrews: addressed your comments | 18:36 |
| *** eventingmonkey has quit IRC | 18:36 | |
| *** eggshell has quit IRC | 18:36 | |
| pabelanger | working on updating nodepoolv3.yaml now | 18:37 |
| Shrews | jeblair: k. putting that one through since it's a doc change that was already +2'd | 18:38 |
| *** eventingmonkey has joined #zuul | 18:42 | |
| *** eggshell has joined #zuul | 18:43 | |
| mordred | Shrews, jeblair: if the three heatclient-ectomy patches I just pushed up pass the dsvm tests, I'll land them and then cut a new shade release so that we can consume that az's method | 19:00 |
| mordred | if they don't pass dsvm we can cut a release without them | 19:00 |
| openstackgerrit | Merged openstack-infra/nodepool feature/zuulv3: Update docs for configuration syntax change https://review.openstack.org/449339 | 19:21 |
| jlk | jesusaur: I took at look at your patch. Other than the pep8 thing, it looks right. Not sure why the tests are failing. | 19:39 |
| jlk | Given that this is all one big series, I'm not sure what to do. I can't make my next patch on top of yours, because tests will fail. But I can't skip yours as it changes some things, so not sure what the next step is. | 19:39 |
| jesusaur | jlk: I'm looking at one of the failures now, and I think it's related to not ignoring formatting exceptions anymore | 19:41 |
| jesusaur | specifically, I think test_json_status has a working report_url now, so I need to change the test to build a correct url to check | 19:42 |
| jlk | ah | 19:45 |
| jlk | It's pretty annoying that a subtle error can generate huge amounts of test output to examine. | 19:46 |
| jesusaur | jeblair: when you replaced the use of build number with the build url, you failed to notice that the build number was a component of the build url | 19:46 |
| jesusaur | jeblair: I'm replacing the number with the uuid in the url, but I can re-add the number if you would prefer shorter log urls | 19:47 |
| jeblair | jesusaur: oh hi | 19:47 |
| jeblair | jesusaur: which build url? | 19:47 |
| jeblair | jesusaur: the one that the executor reports for log streaming while a job is running? | 19:48 |
| jesusaur | jeblair: url_pattern in the test configs | 19:48 |
| jesusaur | yeah | 19:48 |
| jesusaur | er, no, not the streaming, but the url that gets reported along with the success or failure message | 19:50 |
| jeblair | jesusaur: i think that is still somewhat in flux -- we're planning on having a websocket-based web stream of the ansible log; that doesn't exist yet, but that is where the build url will eventually point. i'm not sure of the final format for that, but it will probably be based on the uuid, so if that's what we're doing now, it's probably close enough in spirit for tests, i'd think... | 19:50 |
| jeblair | jesusaur: (that was for the streaming thing ^) | 19:50 |
| jeblair | jesusaur: as for success and failure urls, i think we're just about able to remove the url_pattern from the zuul server config and instead replace it with success-url and failure-url in the job config | 19:51 |
| jeblair | (the idea being that becomes something a site admin sets on the "base" job) | 19:52 |
| jesusaur | oh, ok, I didn't realize that was getting shuffled to another config location | 19:52 |
| jesusaur | but that makes a ton of sense (setting it in a base job as a site admin) | 19:52 |
| jeblair | jesusaur: yeah, it was sort-of already in the job config since you could override it, so it seemed to make sense to just make that the standard way of doing it | 19:53 |
| openstackgerrit | David Shrewsbury proposed openstack-infra/nodepool feature/zuulv3: Docs: Correct availability-zones documentation. https://review.openstack.org/449147 | 20:13 |
| openstackgerrit | David Shrewsbury proposed openstack-infra/nodepool feature/zuulv3: Docs: Remove refs to removed nodepool commands https://review.openstack.org/449152 | 20:13 |
| openstackgerrit | David Shrewsbury proposed openstack-infra/nodepool feature/zuulv3: Docs: Remove cron references https://review.openstack.org/449140 | 20:13 |
| *** jamielennox is now known as jamielennox|away | 20:13 | |
| jeblair | mordred: https://review.openstack.org/448913 may be of interest to you | 20:18 |
| jeblair | mordred: (and child) | 20:19 |
| mordred | jeblair: why yes it is - I was looking at that afew days ago in fact | 20:19 |
| mordred | jeblair: it's a remarkably small patch | 20:20 |
| pabelanger | jeblair: if you want to add 450416 to your review queue, I believe that is the correct syntax for nodepool now | 20:20 |
| jeblair | mordred: yeah, it's good. :) | 20:21 |
| jeblair | pabelanger: lgtm thanks! | 20:22 |
| *** jamielennox|away is now known as jamielennox | 20:24 | |
| jhesketh | Morning | 20:24 |
| pabelanger | Shrews: we'll need 450416 before we restart nl01.o.o | 20:25 |
| Shrews | pabelanger: now, if i +A that, it will be copied over via puppet, right? | 20:27 |
| Shrews | pabelanger: because the problem is, nodepool will notice it changed, then fail b/c it's still running the old code | 20:27 |
| Shrews | so i think we need to stop n-l, approve the change, wait for puppet, then restart | 20:28 |
| jeblair | ++ | 20:28 |
| jeblair | (though i think it'll just spam the logs for a while and otherwise not do anything, so it's not super critical) | 20:28 |
| pabelanger | Shrews: right. and agree with your statement | 20:29 |
| Shrews | pabelanger: going to stop n-l, then +A | 20:29 |
| openstackgerrit | Merged openstack-infra/nodepool feature/zuulv3: Remove api-timeout and provider.image-type https://review.openstack.org/449342 | 20:30 |
| openstackgerrit | Merged openstack-infra/nodepool feature/zuulv3: Remove deprecated networks syntax https://review.openstack.org/449354 | 20:30 |
| Shrews | oh, those are convenient merges | 20:30 |
| jeblair | jhesketh: i copy/pasted some earlier irc discussion to answer your question in https://review.openstack.org/448750 | 20:32 |
| jhesketh | jeblair: thanks (looking) | 20:32 |
| jhesketh | jeblair: sure, that's fair enough.. Shall I +w? | 20:33 |
| jeblair | jhesketh: ++ | 20:34 |
| * Shrews shakes fist at too many conflicting changes | 20:36 | |
| jeblair | jesusaur: https://review.openstack.org/438028 and child are the changes i was thinking of | 20:36 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Refactor out Changeish https://review.openstack.org/448913 | 20:37 |
| jeblair | i'm about to rebase that on the nullchange removal | 20:37 |
| jeblair | which is imminent :) | 20:37 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Remove NullChange https://review.openstack.org/449935 | 20:38 |
| jlk | hopefully that won't throw all my patch set into whack :) | 20:39 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Stop logging git repo rsync output https://review.openstack.org/448750 | 20:43 |
| openstackgerrit | David Shrewsbury proposed openstack-infra/nodepool feature/zuulv3: Docs: Correct availability-zones documentation. https://review.openstack.org/449147 | 20:43 |
| openstackgerrit | David Shrewsbury proposed openstack-infra/nodepool feature/zuulv3: Docs: Remove refs to removed nodepool commands https://review.openstack.org/449152 | 20:44 |
| openstackgerrit | David Shrewsbury proposed openstack-infra/nodepool feature/zuulv3: Docs: Remove cron references https://review.openstack.org/449140 | 20:44 |
| Shrews | if those ^^^ conflict again, i give up | 20:44 |
| mordred | Shrews: I knoew the feeling | 20:45 |
| openstackgerrit | Merged openstack-infra/nodepool feature/zuulv3: Exercise statsd in tests and fix https://review.openstack.org/449737 | 20:45 |
| openstackgerrit | Merged openstack-infra/nodepool feature/zuulv3: Remove ipv6-preferred and rely on interface_ip https://review.openstack.org/449705 | 20:45 |
| openstackgerrit | Merged openstack-infra/nodepool feature/zuulv3: Remove mention of non-clouds.yaml from docs https://review.openstack.org/449781 | 20:45 |
| *** hashar has quit IRC | 20:46 | |
| jeblair | jhesketh: what's the intent for sqlalchemy reporter when 'score' is not specified? store a 0 or NULL in the db? | 20:47 |
| jhesketh | jeblair: currently it stores a 0.. Null might make more sense though | 20:48 |
| jhesketh | jeblair: want me to update that? | 20:48 |
| jeblair | jhesketh: i was just updating it for other reasons, so i'll throw that in. | 20:49 |
| jhesketh | cool, thanks | 20:49 |
| jhesketh | jeblair: probably want to delete line 31 | 20:49 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Remove url_pattern config parameter https://review.openstack.org/447165 | 20:50 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Simplify the log url https://review.openstack.org/438028 | 20:50 |
| jeblair | jhesketh: i went the other way actually -- kept 31 and dropped 41... (see comments i just posted) | 20:50 |
| jeblair | jhesketh: is there a reason we should do that in the report method instead of the initializer? | 20:50 |
| jhesketh | jeblair: nope, just pointing out the useless line | 20:51 |
| jeblair | jhesketh: cool, hopefully that's gtg then. :) | 20:51 |
| jhesketh | yep, looking | 20:51 |
| jhesketh | jeblair: I wonder if we should do a getSafeAttirbutes on Pipeline, Job and Build objects (like Changes) to be consistent? | 20:53 |
| jeblair | jhesketh: sounds like a good idea -- followup patch? | 20:54 |
| Shrews | pabelanger: i have to afk for a few minutes before the meeting. can i leave it to you to restart n-l when it's ready? | 20:55 |
| pabelanger | Shrews: sure | 20:55 |
| Shrews | cool | 20:55 |
| jhesketh | jeblair: sure, I'll do that now | 20:55 |
| jhesketh | jeblair: we'll need to update the sa model to make the score nullable... I'll do that too if you like | 20:56 |
| jhesketh | jeblair: I don't think we need a new migration though unless we're back porting this to zuulv2? | 20:57 |
| jhesketh | or expecting to support an upgrade path | 20:57 |
| jeblair | jhesketh: it's in zuulv2, so we should do a schema upgrade (in both branches) or decide to keep it as 0 | 21:00 |
| jeblair | i'm happy with 0 for now if that's what's in the schema :) | 21:01 |
| jhesketh | jeblair: okay, I think we should do it as a follow up change | 21:01 |
| jhesketh | jeblair: it is, but I can fix it up | 21:01 |
| jeblair | ok | 21:01 |
| jeblair | jhesketh: how about we switch https://review.openstack.org/447165 to 0, then change it to null along with the schema migration? | 21:02 |
| jhesketh | jeblair: yep, just did that | 21:02 |
| jeblair | ok cool | 21:02 |
| jeblair | jhesketh: were you pushing up a new ps to 447165? | 21:11 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Re-enable test_stuck_job_cleanup https://review.openstack.org/446783 | 21:13 |
| openstackgerrit | Joshua Hesketh proposed openstack-infra/zuul feature/zuulv3: Remove url_pattern config parameter https://review.openstack.org/447165 | 21:14 |
| openstackgerrit | Joshua Hesketh proposed openstack-infra/zuul feature/zuulv3: Implement getSafeAttributes for other objects https://review.openstack.org/450459 | 21:14 |
| jhesketh | jeblair: ^ | 21:14 |
| openstackgerrit | Joshua Hesketh proposed openstack-infra/zuul feature/zuulv3: Make sql reporter score null when not set https://review.openstack.org/450460 | 21:14 |
| jeblair | w00t | 21:14 |
| openstackgerrit | Joshua Hesketh proposed openstack-infra/zuul feature/zuulv3: Make sql reporter score null when not set https://review.openstack.org/450460 | 21:15 |
| pabelanger | starting nl01.o.o | 21:15 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Consume interface_ip from nodepool node https://review.openstack.org/449706 | 21:17 |
| jeblair | mordred, jhesketh: ^ updated that to fix a merge conflict (and added a doc TODO) | 21:17 |
| jhesketh | jeblair: lgtm | 21:18 |
| jesusaur | jeblair: ah, thanks, 438028 is definitely going to change how I handle 449794 | 21:25 |
| jeblair | jhesketh: i'm fixing pep8 in 165 | 21:26 |
| jhesketh | jeblair: ack | 21:26 |
| jesusaur | jhesketh: would you be willing to copy my formatUrlPattern exception handling from 449794 into 438028? | 21:27 |
| jesusaur | (note: it exposes some bugs) | 21:27 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Make sql reporter score null when not set https://review.openstack.org/450460 | 21:28 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Implement getSafeAttributes for other objects https://review.openstack.org/450459 | 21:28 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Remove url_pattern config parameter https://review.openstack.org/447165 | 21:28 |
| pabelanger | Shrews: we look to be collecting Gathering host keys properly | 21:28 |
| jhesketh | jesusaur: I think we could do that as a follow up instead | 21:29 |
| jhesketh | jeblair: unless you disagree ^ | 21:29 |
| jhesketh | jesusaur: I'm happy to do that if you'd like though | 21:30 |
| jeblair | followup sounds good | 21:32 |
| jesusaur | jhesketh: yeah, it would be great if you could write that follow-up; it would let me go back to debugging 446275 | 21:34 |
| pabelanger | Shrews: and confirmed by using zk_shell | 21:34 |
| jhesketh | jesusaur: sure, will do | 21:34 |
| Shrews | pabelanger: we've been doing that for a few days | 21:35 |
| Shrews | at least, i thought we were | 21:35 |
| pabelanger | Shrews: Oh, neat. | 21:36 |
| openstackgerrit | Joshua Hesketh proposed openstack-infra/zuul feature/zuulv3: Make sql reporter score null when not set https://review.openstack.org/450460 | 21:41 |
| openstackgerrit | Joshua Hesketh proposed openstack-infra/zuul feature/zuulv3: Catch and log url pattern formatting errors https://review.openstack.org/450468 | 21:47 |
| jhesketh | jesusaur: ^ | 21:47 |
| jesusaur | jhesketh: awesome, thanks | 21:50 |
| jesusaur | jlk: I think we should rebase the github stack on top of 450468 | 21:50 |
| jesusaur | jlk: and I can help debug that stack if we start getting a ton of new failures | 21:55 |
| SpamapS | 'bout meeting time eh? | 21:59 |
| jeblair | yep! zuul meeting time in #openstack-meeting-alt | 22:00 |
| jlk | jesusaur: alrighty I will give it a shot. | 22:57 |
| jlk | jesusaur: looks like I have to rebase anyway since all my PRs just got a merge failure. | 23:06 |
| jlk | jesusaur: sadly 450468 is failing. | 23:08 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!