| openstackgerrit | Merged zuul/zuul master: Remove fix-tox workaround for python3.5 https://review.opendev.org/712495 | 00:01 |
|---|---|---|
| *** jamesmcarthur has joined #zuul | 00:15 | |
| *** erbarr has quit IRC | 00:30 | |
| *** dangtrinhnt has joined #zuul | 00:38 | |
| *** dangtrinhnt has quit IRC | 00:41 | |
| *** dangtrinhnt_ has joined #zuul | 00:42 | |
| *** jamesmcarthur has quit IRC | 00:57 | |
| *** jamesmcarthur has joined #zuul | 00:57 | |
| *** dangtrinhnt_ has quit IRC | 01:48 | |
| *** rlandy has quit IRC | 01:58 | |
| *** jamesmcarthur has quit IRC | 02:02 | |
| *** jamesmcarthur has joined #zuul | 02:02 | |
| *** swest has quit IRC | 02:05 | |
| *** bhavikdbavishi has quit IRC | 02:20 | |
| *** zenkuro has quit IRC | 02:20 | |
| *** swest has joined #zuul | 02:22 | |
| *** jamesmcarthur has quit IRC | 02:54 | |
| *** jamesmcarthur has joined #zuul | 02:57 | |
| *** bhavikdbavishi has joined #zuul | 03:04 | |
| *** jamesmcarthur has quit IRC | 03:16 | |
| *** bhavikdbavishi has quit IRC | 03:22 | |
| *** bhavikdbavishi has joined #zuul | 03:27 | |
| *** jamesmcarthur has joined #zuul | 03:31 | |
| *** saneax has joined #zuul | 03:43 | |
| fungi | going to send this slide deck to the cdf interoperability sig ml tomorrow in preparation for presenting it on thursday's biweekly meeting call: http://fungi.yuggoth.org/presentations/2020-cdf/ | 04:10 |
| fungi | if anybody has last-minute suggestions, i'm happy to try to work them in | 04:10 |
| *** raukadah is now known as chandankumar | 04:37 | |
| *** bolg_ has joined #zuul | 04:50 | |
| *** bolg has quit IRC | 04:51 | |
| *** jamesmcarthur has quit IRC | 05:11 | |
| *** kmalloc has joined #zuul | 05:33 | |
| *** evrardjp has quit IRC | 05:35 | |
| *** evrardjp has joined #zuul | 05:36 | |
| *** freefood has joined #zuul | 06:17 | |
| freefood | Hello! :) :wave: | 06:17 |
| freefood | How long does it typically take for Zuul to trigger on new reply? I replied "recheck" on my patch 30 minutes ago but Zuul hasn't re-run the test jobs yet :( | 06:21 |
| freefood | FWIW: My patch is at: https://review.opendev.org/#/c/713076/ | 06:21 |
| AJaeger | freefood: it depends on the number of available nodes, let me check... | 06:49 |
| AJaeger | freefood: http://grafana.openstack.org/d/T6vSHcSik/zuul-status?orgId=1 shows the "node requests", so there's a backlog | 06:50 |
| AJaeger | freefood: and http://zuul.opendev.org/t/openstack/status shows all changes | 06:50 |
| AJaeger | so, expected time to finish is 12 mins - and if you look at previous run, the longest job run for 71 minutes, so that means all is fine and jobs must have been running already when you asked | 06:53 |
| *** Defolos has joined #zuul | 07:03 | |
| freefood | Thanks AJaeger! \o/ I'm new here :) | 07:04 |
| *** dpawlik has joined #zuul | 07:10 | |
| *** dpawlik has quit IRC | 07:26 | |
| AJaeger | welcome, freefood ! | 07:38 |
| *** kmalloc has quit IRC | 07:42 | |
| *** dmsimard|off7 has joined #zuul | 07:48 | |
| *** dmsimard|off has quit IRC | 07:48 | |
| *** dmsimard|off7 is now known as dmsimard|off | 07:48 | |
| *** saneax has quit IRC | 07:52 | |
| *** avass has joined #zuul | 07:52 | |
| *** tosky has joined #zuul | 07:57 | |
| *** dpawlik has joined #zuul | 07:58 | |
| openstackgerrit | Daniel Pawlik proposed zuul/zuul-jobs master: Add phoronix-test-suite job https://review.opendev.org/679082 | 08:05 |
| *** yoctozepto9 has joined #zuul | 08:15 | |
| *** yoctozepto has quit IRC | 08:15 | |
| *** yoctozepto9 is now known as yoctozepto | 08:15 | |
| *** hashar has joined #zuul | 08:16 | |
| *** saneax has joined #zuul | 08:22 | |
| *** jcapitao has joined #zuul | 08:23 | |
| *** sshnaidm has quit IRC | 08:29 | |
| *** saneax has quit IRC | 08:31 | |
| *** yoctozepto has quit IRC | 08:49 | |
| *** jpena|off is now known as jpena | 08:55 | |
| *** hashar is now known as hasharSchool | 08:55 | |
| *** harrymichal has joined #zuul | 08:58 | |
| *** yoctozepto has joined #zuul | 08:59 | |
| *** dpawlik has quit IRC | 09:03 | |
| *** dpawlik has joined #zuul | 09:03 | |
| *** dpawlik has quit IRC | 09:08 | |
| *** bhavikdbavishi has quit IRC | 09:17 | |
| openstackgerrit | Alfredo Moralejo proposed zuul/zuul-jobs master: Use centos8 deps repo for RHEL8 nodes https://review.opendev.org/708430 | 09:29 |
| *** bhavikdbavishi has joined #zuul | 09:50 | |
| *** yoctozepto has quit IRC | 09:59 | |
| *** mhu has quit IRC | 09:59 | |
| *** corvus_ has joined #zuul | 10:01 | |
| *** yoctozepto has joined #zuul | 10:03 | |
| *** jpena has quit IRC | 10:04 | |
| *** corvus has quit IRC | 10:04 | |
| *** jpena has joined #zuul | 10:05 | |
| *** saneax has joined #zuul | 10:10 | |
| *** yoctozepto has quit IRC | 10:13 | |
| *** yoctozepto has joined #zuul | 10:14 | |
| *** sshnaidm has joined #zuul | 10:20 | |
| zbr | there was a feature i wanted to implement on tox, when tox_envlist is a list to run each task separated, so we get much easier to read output. anyone against? | 10:46 |
| zbr | so instead of running "tox -e foo,bar", it will run run each of them. | 10:46 |
| zbr | if one fails, we will be able read the error, without spam from the other one. | 10:47 |
| *** sshnaidm has quit IRC | 10:54 | |
| *** sshnaidm has joined #zuul | 10:55 | |
| *** harrymichal has quit IRC | 10:57 | |
| *** okamis has joined #zuul | 11:19 | |
| *** harrymichal has joined #zuul | 11:28 | |
| *** rlandy has joined #zuul | 11:31 | |
| *** hasharSchool is now known as hashar | 11:39 | |
| *** harrymichal has quit IRC | 11:45 | |
| *** jcapitao is now known as jcapitao_lunch | 11:53 | |
| *** Goneri has joined #zuul | 12:16 | |
| *** hashar is now known as hasharLunch | 12:16 | |
| *** saneax has quit IRC | 12:25 | |
| *** avass has quit IRC | 12:32 | |
| *** jpena is now known as jpena|lunch | 12:39 | |
| *** harrymichal has joined #zuul | 12:40 | |
| *** saneax has joined #zuul | 12:46 | |
| *** hasharLunch is now known as hashar | 12:59 | |
| *** saneax has quit IRC | 13:07 | |
| *** jcapitao_lunch is now known as jcapitao | 13:08 | |
| openstackgerrit | Merged zuul/zuul-jobs master: Use centos8 deps repo for RHEL8 nodes https://review.opendev.org/708430 | 13:16 |
| mnaser | possible to have eyes on https://review.opendev.org/#/c/704266/ -- it'll help us so we can get cheroot tests going! | 13:19 |
| mnaser | and hopefully cherrypy after | 13:19 |
| *** harrymichal has quit IRC | 13:26 | |
| *** harrymichal has joined #zuul | 13:26 | |
| mnaser | corvus_: could we get some ML announcements when the ZK changes land? due to the fact there is not really tagged dockerhub images, "cd" users will get that breaking change | 13:27 |
| *** bhavikdbavishi has quit IRC | 13:37 | |
| *** TomConte has joined #zuul | 13:38 | |
| *** jpena|lunch is now known as jpena | 13:42 | |
| *** TomConte has quit IRC | 13:45 | |
| *** harrymichal has quit IRC | 13:53 | |
| tristanC | mnaser: is the ZK change breaking yet? | 13:55 |
| mnaser | tristanC: oh i dont think it is so far, but i assume there is a period of time when the zk enforced auth lands until zuul v4 (i think that was the agreed thing) is released | 13:57 |
| mnaser | which will make it borked | 13:58 |
| tristanC | mnaser: i think the plan is to first enable using tls, then announce that user needs to configure tls. I don't think we have change that enforce tls anytime soon | 14:00 |
| mnaser | tristanC: ok I see, it was just something that I was thinking about. | 14:02 |
| tristanC | mnaser: also I'm quite worried that no zookeeper deployment system support tls... i'm not sure we can enforce tls before such things is available for users | 14:02 |
| mnaser | tristanC: I’ve been so curious about this whole thing with how to do it on k8s with different cette | 14:03 |
| mnaser | certs per pod | 14:03 |
| tristanC | mnaser: it seems like that could be done easily using the same process that setup the myid file | 14:04 |
| tristanC | (e.g. copy the correct cert from a common secret based on hostname) | 14:04 |
| mnaser | tristanC: yeah but I guess then all of the zookeeper instances have access to each other secrets and I don’t know what that means from a security pov | 14:05 |
| tristanC | mnaser: then using a initContainer with elevated privilege could do that cert selection copy securely | 14:06 |
| tristanC | mnaser: i guess you are using the zookeeper incubator chart? are you going to add tls support there? | 14:06 |
| mnaser | tristanC: that’s true, we could do that and so in runtime zookeeper is nerve exposed to it | 14:06 |
| mnaser | tristanC: I plan to, I think the issue is around the fact that I think the repo doesn’t accept changes anymore at this point so we might have to fork and maintain elsewhere | 14:07 |
| tristanC | really? any reason why it doesn't accept changes? | 14:08 |
| tristanC | i've been looking at pravega/zookeeper-operator , though i couldn't get it working in dev mode yet | 14:10 |
| *** yoctozepto1 has joined #zuul | 14:12 | |
| *** yoctozepto has quit IRC | 14:13 | |
| *** yoctozepto1 is now known as yoctozepto | 14:13 | |
| tristanC | and golang seems overkill too. makes me wonder if we shouldn't maintain our own zk resources. Isn't what you are doing tobiash? | 14:15 |
| *** saneax has joined #zuul | 14:16 | |
| tobiash | tristanC: our zk is part of out zuul deployment and basically consists of a statefulset with service, nothing more nothing less | 14:23 |
| tobiash | and it just works without any notable maintenance effort (after tweaking to some infrastructure limitations) | 14:24 |
| tristanC | tobiash: that's what i think too, would you mind sharing your resources file? | 14:29 |
| tobiash | tristanC: of course, I just need to make them ready for sharing before and I'm deep in ops topics today, so cannot promise for today | 14:30 |
| tristanC | tobiash: no rush. that would be helpful to see what is really needed to run an acceptable zk service | 14:33 |
| tobiash | tristanC: basically it's a fork of this which puts data on tmpfs combined with backup to pvc in a sidecar and loading the backup in boot.sh: https://github.com/kubernetes-retired/contrib/tree/master/statefulsets/zookeeper | 14:41 |
| tristanC | arg, so you are also building your own image | 14:43 |
| tobiash | yes | 14:43 |
| tobiash | is that a blocker? | 14:44 |
| tristanC | well if we could do that in opendev, then that statefulsets recipe sounds easy to use | 14:44 |
| *** jamesmcarthur has joined #zuul | 14:44 | |
| tristanC | tobiash: i don't think so, but opendev would become a distributor of zookeeper image too | 14:45 |
| tobiash | yes, that would be the implication, no idea if this is acceptable from opendev and license perspective | 14:46 |
| *** bschanzel has joined #zuul | 14:49 | |
| corvus_ | the tmpfs/pvc backup sounds pretty specialized for tobiash's situation; is that something we think we should support for everyone? | 14:50 |
| tobiash | corvus_: I'm happy to share if there is interest | 14:50 |
| bschanzel | Hey Guys, I wanted to ask back if there are any plans on when I can get a workflow +1 on these https://review.opendev.org/#/c/711920/ https://review.opendev.org/#/c/712034/ | 14:51 |
| tobiash | corvus_: it's a (for us workable) tradeoff towards performance for either huge deployments or medium to large deployments that have no fast disks (aka network drive) available for zk | 14:51 |
| corvus_ | bschanzel: i'll take a look today, thanks! | 14:52 |
| bschanzel | corvus_ cool, thanks! | 14:52 |
| corvus_ | tobiash: right, it sounds important for you, but it doesn't necessarily sound like the right system for all zuul users | 14:53 |
| corvus_ | most users should be able to use the upstream image i would think | 14:55 |
| *** corvus_ is now known as jeblair | 14:56 | |
| clarkb | thinking outloud, do you need a special image to do what tobiash is doing? I guessif you want the container to make the tmpfs instead of the host? | 14:56 |
| *** jeblair is now known as corvus | 14:56 | |
| clarkb | the backup stuff could a sidecar right? | 14:56 |
| tobiash | clarkb: the backup stuff is a sidecar (however we use a custom boot.sh that copies over the backup, but that could be moved into an init container) | 14:57 |
| tobiash | the most important is the generated zk config during boot | 14:58 |
| tobiash | but no idea if there is an upstream image now that supports running in a statefulset | 14:58 |
| corvus | tobiash: what would the image require in order to support a statefulset? | 14:59 |
| tobiash | corvus: it's basically this what was required: https://github.com/kubernetes-retired/contrib/blob/master/statefulsets/zookeeper/zkGenConfig.sh | 15:00 |
| tobiash | (back then) | 15:00 |
| tobiash | more specifically MY_ID: https://github.com/kubernetes-retired/contrib/blob/master/statefulsets/zookeeper/zkGenConfig.sh#L64 | 15:01 |
| corvus | tobiash: the current zookeeper image has an entrypoint script which generates a zoo.cfg file from some env variables if it doesn't already exist; and also the myid file | 15:01 |
| corvus | but i'm not sure it supports arbitrary data in the zoo.cfg; so i don't think one could just add an env variable for the tls cert path | 15:02 |
| corvus | i think it only has some predefined variables it understands | 15:02 |
| tobiash | same with this config generate script | 15:02 |
| tobiash | so I guess that needs to be patched upstream or the config could be pre-generated by an init container | 15:03 |
| *** avass has joined #zuul | 15:03 | |
| *** bolg has joined #zuul | 15:05 | |
| *** bolg_ has quit IRC | 15:06 | |
| mnaser | tristanC: see https://github.com/helm/charts#status-of-the-project | 15:07 |
| corvus | here's the entrypoint script in the current image: https://github.com/31z4/zookeeper-docker/blob/6b7a8a55d9760f9737b03366e1b3b7edbe0bd895/3.5.7/docker-entrypoint.sh | 15:07 |
| corvus | maybe they would accept some SSL options? | 15:08 |
| *** saneax has quit IRC | 15:11 | |
| mordred | it does seem to be an active repo - so maybe that wouldn't be unwelcome | 15:12 |
| tobiash | corvus: looks like this entrypoint takes its id from an env var: https://github.com/31z4/zookeeper-docker/blob/6b7a8a55d9760f9737b03366e1b3b7edbe0bd895/3.5.7/docker-entrypoint.sh#L44 | 15:14 |
| *** okamis has quit IRC | 15:14 | |
| tobiash | corvus: I'm not sure if it's possible to have different env vars per pod within a statefulset | 15:14 |
| avass | I have a feeling that ensure-tox and tox roles are not compatible since ensure-tox installs tox to ~/.local/bin and that's not in PATH by default | 15:15 |
| avass | Oh wait, could it be the ansible varaible cache not working? | 15:18 |
| openstackgerrit | Graham Hayes proposed zuul/nodepool master: Implement an Azure driver https://review.opendev.org/554432 | 15:19 |
| mordred | avass: that could be a thing to look at - it's supposed to work via ansible variable cache | 15:20 |
| tobiash | corvus: there is an upstream kubernetes issue with exact this question: https://github.com/kubernetes/kubernetes/issues/40651 | 15:25 |
| tristanC | corvus: we can provides a custom zoo.cfg to the upstream zk image like so: https://review.opendev.org/#/c/712759/11/conf/zuul/resources.dhall@648 | 15:26 |
| tristanC | mnaser: i see, and i guess nobody started to maintain the zk chart? | 15:28 |
| mnaser | tristanC: havent checked that yet but i assume if it was incubator that may be the case | 15:29 |
| *** jcapitao is now known as jcapitao_afk | 15:31 | |
| mordred | fieldPath: metadata.annotations['spec.pod.beta.kubernetes.io/statefulset-index'] | 15:31 |
| mordred | rolls right off the tongue | 15:31 |
| tristanC | tobiash: corvus: so it seems relatively simple to work around the upstream zk image and integrate the statefulset tobiash linked within the zuul-operator. should i keep going down the cert-manager path and look into a initcontainer to setup the zk tls files? | 15:37 |
| tristanC | alternatively, we could look into moving the zk deployment system outside of the zuul-operator, though i don't know how and where would that happen | 15:38 |
| tristanC | or we could stay with the current implementation (single pod setup), and document that production system needs to provide an existing zk | 15:41 |
| corvus | tristanC: what issues are you having with the zk operator? | 15:42 |
| mugsie | I just rebased https://review.opendev.org/554432 and addressed the questions - but I was told there is a newer simpler interface for nodepool drivers - should I update this to the new interface? | 15:45 |
| corvus | mugsie: i took a quick look at azure when i made the "simple" driver interface, and my initial thought was that azure would be too complicated for it, but let me get you a link -- you'll know better | 15:46 |
| avass | mordred: it's probably something with the fact cache but looking at ansible.cfg it's enabled but it doesn't seem to work anyway. | 15:46 |
| mugsie | corvus: cool - thanks :) | 15:46 |
| corvus | mugsie: (basically the "simple" interface is a higher-level api abstraction that sits on top of the existing one, so there's no problem using the lower-level api if that's what's needed) | 15:47 |
| mugsie | its the one GCP driver used, right? | 15:47 |
| mugsie | one the GCP* | 15:47 |
| corvus | mugsie: yes; it's here: https://zuul-ci.org/docs/nodepool/devguide.html#simple-drivers | 15:48 |
| mugsie | cool, I will look at what you did there | 15:48 |
| corvus | mugsie: i'm pretty sure we can port aws to it. i'm certain that we can not port openstack to it. | 15:48 |
| corvus | tristanC: well, in general, i think if there's a suitable standalone zk operator that the zuul-operator could rely on (we can ask users to install the zk operator first, and have zuul manage the CRs for it), that would be great. if that's not possible, then i think managing a statefulset with init-containers inside of zuul-operator is the way to go. | 15:51 |
| corvus | avass: you may be able to interactively debug by asking the executor to hold the build directory. "zuul-executor keep" turns that on. then when the job is finished, you can cd into there and examine the fact cache. | 15:53 |
| tristanC | corvus: it's just that the zookeeper-operator build system assume you know golang, and it's not very helpful. it's now failing because of `dep: command not found` | 15:53 |
| mordred | tristanC, corvus: I agree - I thnik it would be nice if we can reuse existing operator logic - but I also don't think we should kill ourselves to do it - especially since zk itself doesn't need a _lot_ of complicated care and feeding | 15:56 |
| tristanC | of course it would be better to reuse an existing thing | 15:57 |
| tristanC | after installing the `dep` tool, the zk-operator built, but it doesn't seem to run locally using `operator-sdk run --local`, that fails with `build flag -mod=vendor only valid when using modules` | 15:59 |
| mordred | that doesn't inspire confidence | 16:00 |
| *** hashar is now known as hasharAway | 16:03 | |
| *** jcapitao_afk is now known as jcapitao | 16:03 | |
| *** sshnaidm has quit IRC | 16:28 | |
| *** sshnaidm has joined #zuul | 16:28 | |
| *** bolg has quit IRC | 16:36 | |
| *** harrymichal has joined #zuul | 16:44 | |
| *** johanssone has quit IRC | 16:51 | |
| *** bolg has joined #zuul | 16:52 | |
| *** bolg has quit IRC | 16:53 | |
| *** johanssone has joined #zuul | 16:56 | |
| *** johanssone has quit IRC | 16:57 | |
| *** bschanzel has left #zuul | 17:02 | |
| *** johanssone has joined #zuul | 17:03 | |
| *** johanssone has quit IRC | 17:03 | |
| *** johanssone has joined #zuul | 17:05 | |
| *** johanssone has quit IRC | 17:05 | |
| *** johanssone has joined #zuul | 17:06 | |
| *** johanssone has quit IRC | 17:08 | |
| openstackgerrit | Graham Hayes proposed zuul/nodepool master: Implement an Azure driver https://review.opendev.org/554432 | 17:11 |
| *** johanssone has joined #zuul | 17:14 | |
| *** johanssone has quit IRC | 17:15 | |
| *** johanssone has joined #zuul | 17:20 | |
| *** johanssone has quit IRC | 17:20 | |
| *** johanssone has joined #zuul | 17:23 | |
| noonedeadpunk | mordred: mnaser what do you think regarding this elements patch? https://review.opendev.org/#/c/713692/ | 17:23 |
| *** johanssone has quit IRC | 17:27 | |
| *** johanssone has joined #zuul | 17:30 | |
| *** johanssone has quit IRC | 17:32 | |
| *** evrardjp has quit IRC | 17:36 | |
| *** evrardjp has joined #zuul | 17:36 | |
| *** johanssone has joined #zuul | 17:46 | |
| mnaser | noonedeadpunk: it makes sense to me, it'd be nice to hear about what others think too | 17:49 |
| *** johanssone has quit IRC | 17:54 | |
| *** johanssone has joined #zuul | 18:02 | |
| *** johanssone has joined #zuul | 18:03 | |
| *** johanssone has quit IRC | 18:08 | |
| *** johanssone has joined #zuul | 18:13 | |
| *** jpena is now known as jpena|off | 18:16 | |
| *** jcapitao is now known as jcapitao_off | 18:19 | |
| *** johanssone has quit IRC | 18:19 | |
| *** johanssone has joined #zuul | 18:22 | |
| *** johanssone has quit IRC | 18:25 | |
| *** johanssone has joined #zuul | 18:34 | |
| mordred | noonedeadpunk: oh wow, that looks nice! | 18:37 |
| *** johanssone has quit IRC | 18:37 | |
| mordred | noonedeadpunk: that said - a further optimization you might want to make would be to remove the -dev packages and build tools and just install the normal libs after you've done the pyenv thing. I think you can tie different packages to different stages in package-installs? | 18:39 |
| mordred | so maybe you could mark some as remove in post-install? | 18:39 |
| mordred | (just so you don't have your images polluted with the build toolchains) | 18:40 |
| *** sshnaidm is now known as sshnaidm|afk | 18:41 | |
| *** hasharAway has quit IRC | 18:41 | |
| *** hasharAway has joined #zuul | 18:41 | |
| *** johanssone has joined #zuul | 18:43 | |
| *** hasharAway is now known as hashar | 18:44 | |
| mordred | hrm. it doesn't look like you can easily say "install this package in pre-install and then uninstall it in post-install" | 18:44 |
| mordred | might need to be a second element that just has the same list but with uninstall true | 18:44 |
| *** johanssone has quit IRC | 18:47 | |
| *** johanssone has joined #zuul | 18:50 | |
| noonedeadpunk | mordred: yeah, ok, that's actually fair | 18:50 |
| mordred | but that's an exciting element :) | 18:51 |
| *** johanssone has quit IRC | 18:57 | |
| noonedeadpunk | mordred: do you think packages removal should be accomplished with adjusting existing element or creating simmilar one? | 18:59 |
| mordred | noonedeadpunk: if it WORKS to do it in the existing one - I thnik that would be best - but I'm not sure package-installs.yaml is flexible enough for that sadly | 19:00 |
| noonedeadpunk | Yeah, that's why I've asked... | 19:01 |
| noonedeadpunk | will try to think what to do with that | 19:01 |
| mordred | so I think, assuming I'm right and it's not good enough - we could either improve it - probably more difficult - or just made a second cleanup element | 19:01 |
| mordred | it would be nice if you could say: "libbz2-dev: [{phase: pre-install, action: install}, {phase: post-install, action: uninstall}] | 19:02 |
| noonedeadpunk | yeah | 19:03 |
| mordred | but it wriiting that might suck- so you might want to just do the cleanup element and leave implementing that for next time you're bored :) | 19:03 |
| *** johanssone has joined #zuul | 19:04 | |
| *** johanssone has joined #zuul | 19:06 | |
| *** johanssone has quit IRC | 19:06 | |
| noonedeadpunk | mordred: are you sure it's not implemented? https://opendev.org/openstack/diskimage-builder/src/branch/master/diskimage_builder/elements/package-installs/install.d/99-package-uninstalls | 19:07 |
| noonedeadpunk | as it at first look seems like cleanup is done? | 19:08 |
| *** johanssone has joined #zuul | 19:11 | |
| noonedeadpunk | so the only thing I need to remove is python-build? | 19:12 |
| mordred | let me see ... I *think* it's mututally exclusive? | 19:12 |
| noonedeadpunk | hm, may be - not so big expert in elements tbh:) | 19:13 |
| *** johanssone has quit IRC | 19:14 | |
| noonedeadpunk | just thought that at 00 it installs it and at 95 - uninstalls? | 19:14 |
| mordred | yeah - but I think what it's doing is installing all packages marked for installation at 00 and all things marked for uninstallation at 95 | 19:18 |
| mordred | it's just - I think marking something for uninstall causes it to not be marked for indtall | 19:18 |
| mordred | install | 19:18 |
| * mordred is trying to prove this real quick | 19:18 | |
| *** johanssone has joined #zuul | 19:21 | |
| mordred | noonedeadpunk: yes - I have confirmed - if you mark a package as uninstall: true it is _not_ put into the list of things to install | 19:31 |
| *** Defolos has quit IRC | 19:32 | |
| noonedeadpunk | :( | 19:32 |
| mordred | so I think second element | 19:32 |
| mordred | yeah. it's sad-making | 19:33 |
| mordred | let me see if I can make a simple patch to support both | 19:33 |
| mordred | noonedeadpunk: got it | 19:37 |
| noonedeadpunk | mordred: oh, really? | 19:39 |
| mordred | noonedeadpunk: remote: https://review.opendev.org/713726 Add support for build-only packages | 19:40 |
| noonedeadpunk | oh, that's neat | 19:42 |
| avass | corvus: thanks for the tip earlier, I'll try to see if I can figure it out tomorrow | 19:50 |
| Shrews | ianw: i left some comments on your diskimages-globals changes without actually voting just yet | 19:55 |
| *** gmann is now known as gmann_afk | 19:57 | |
| *** hashar is now known as hasharAway | 20:00 | |
| *** jamesmcarthur has quit IRC | 20:09 | |
| *** Openk10s has joined #zuul | 20:14 | |
| openstackgerrit | Merged zuul/zuul-jobs master: Kubernetes Node Support for Mirroring Git Repos https://review.opendev.org/711920 | 20:18 |
| noonedeadpunk | mordred: rebased on top of it and added cleanup | 20:19 |
| openstackgerrit | Merged zuul/zuul-jobs master: Optimize repos for size when preparing for nodecache https://review.opendev.org/710720 | 20:23 |
| mordred | noonedeadpunk: sweet! I think there's one more thing missing - but I think that's gonna be really nice | 20:24 |
| mordred | noonedeadpunk: you also need to add the non-dev versions of a few of the libs - notably libbzip2 libreadline libssl sqlite and libzlib | 20:25 |
| mordred | noonedeadpunk: otherwise the pythons you installed won't have the runtime libs :) | 20:25 |
| noonedeadpunk | except adjusting zuul job to use stow? | 20:25 |
| mordred | and that | 20:25 |
| noonedeadpunk | hm, yeah, that's make sense... | 20:26 |
| *** jamesmcarthur has joined #zuul | 20:28 | |
| noonedeadpunk | but I think depends on do not work anyway? | 20:30 |
| noonedeadpunk | So zuul thing makes sense to place only once having element working... | 20:31 |
| mordred | oh - yeah - totally | 20:31 |
| mordred | well - we probably aren't going to have that element in place in opendev anyway - so I think the zuul-jobs patch can probably go ahead and be written | 20:32 |
| mordred | and we'll figure out a good way to test it upstream :) | 20:32 |
| mnaser | does nodepool use uncapped dib in its images? | 20:32 |
| mnaser | im just thinking how we can get access to noonedeadpunk's change when we use nodepool images | 20:32 |
| *** gmann_afk is now known as gmann | 20:33 | |
| mordred | mnaser: uhm... I think we use releases - so we'll need to cut a dib release then build a new nodepool-builder image | 20:34 |
| mordred | what a nice stack :) | 20:34 |
| mnaser | oh okay, i thought the nodepool images were just rolling latest * | 20:34 |
| mordred | they are of nodepool | 20:35 |
| mordred | but I think we're just installing dib from pip | 20:35 |
| openstackgerrit | Merged zuul/nodepool master: Kubernetes/OpenShift Provider: Don't Require Bash in Container Images https://review.opendev.org/712034 | 20:46 |
| noonedeadpunk | mordred: btw, I think these libs should be already present, as they're installed as dependecies to dev one, and moreover required for the base python.... | 20:49 |
| mordred | ah - ok | 20:50 |
| mordred | then let's not worry about that | 20:50 |
| *** jamesmcarthur has quit IRC | 20:52 | |
| *** jamesmcarthur has joined #zuul | 20:53 | |
| ianw | Shrews: thanks will go over soon | 20:55 |
| noonedeadpunk | maybe I should better clenup dev requirements though.... | 20:55 |
| noonedeadpunk | But that is smth I'll be able to continue only tomorrow... | 20:56 |
| *** jamesmcarthur has quit IRC | 20:58 | |
| *** jamesmcarthur has joined #zuul | 20:58 | |
| mordred | noonedeadpunk: I thnk it's looking great so far! | 20:58 |
| *** jamesmcarthur has quit IRC | 21:00 | |
| *** hasharAway has quit IRC | 21:00 | |
| *** hasharAway has joined #zuul | 21:01 | |
| *** jamesmcarthur has joined #zuul | 21:03 | |
| *** jamesmcarthur has quit IRC | 21:08 | |
| *** mattw4 has joined #zuul | 21:10 | |
| corvus | tristanC: the complete zk stack is green now; based on our earlier conversation, i will start rebasing the tls portions to not require sasl. i think that we may want to consider adding the acl support back in, because a site may potentially want to set x509 acls, but i don't think it's important we do that right now, and we can add it on after tls support. | 21:15 |
| tristanC | corvus: wfm, thanks! | 21:17 |
| *** openstackstatus has joined #zuul | 21:17 | |
| *** ChanServ sets mode: +v openstackstatus | 21:17 | |
| *** jamesmcarthur has joined #zuul | 21:18 | |
| openstackgerrit | Andreas Jaeger proposed zuul/zuul-jobs master: Update ansible-lint exception https://review.opendev.org/713742 | 21:18 |
| tristanC | corvus: though, if we consider supporting sasl, i'd rather we merge the change now and avoid future conflict/failure. otherwise i'd prefer abandon it now | 21:18 |
| *** hasharAway is now known as hashar | 21:19 | |
| corvus | tristanC: me too, but i feel like we got enough buy-in from folks saying tls-only is okay, and i think ultimately that's friendlier/easier than the sasl setup, so i'm inclined to agree with you and abandon it | 21:20 |
| *** kmalloc has joined #zuul | 21:25 | |
| openstackgerrit | James E. Blair proposed zuul/zuul master: Add TLS support for ZooKeeper https://review.opendev.org/712531 | 21:28 |
| openstackgerrit | James E. Blair proposed zuul/nodepool master: Add ZooKeeper TLS support https://review.opendev.org/712733 | 21:32 |
| openstackgerrit | James E. Blair proposed zuul/zuul master: Add TLS support for ZooKeeper https://review.opendev.org/712531 | 21:34 |
| openstackgerrit | James E. Blair proposed zuul/zuul master: Move zuul-quick-start requires to zuul-build-image https://review.opendev.org/713545 | 21:34 |
| openstackgerrit | James E. Blair proposed zuul/zuul master: Use ZK TLS in quickstart https://review.opendev.org/712817 | 21:34 |
| mordred | corvus: there's still sasl docs in that | 21:34 |
| corvus | mordred: which one? | 21:34 |
| mordred | https://review.opendev.org/#/c/712531/18/doc/source/howtos/zookeeper.rst | 21:35 |
| corvus | oh that, yeah, i missed that cause i added that in the tls change; | 21:35 |
| corvus | i'll pull that out now | 21:35 |
| corvus | and i'm rewriting it a little :) | 21:38 |
| *** jamesmcarthur has quit IRC | 21:40 | |
| openstackgerrit | James E. Blair proposed zuul/zuul master: Add TLS support for ZooKeeper https://review.opendev.org/712531 | 21:47 |
| openstackgerrit | James E. Blair proposed zuul/zuul master: Move zuul-quick-start requires to zuul-build-image https://review.opendev.org/713545 | 21:47 |
| openstackgerrit | James E. Blair proposed zuul/zuul master: Use ZK TLS in quickstart https://review.opendev.org/712817 | 21:47 |
| corvus | i put in a bit more info about the "shared CA" concern | 21:47 |
| *** jamesmcarthur has joined #zuul | 22:20 | |
| *** rlandy is now known as rlandy|bbl | 22:27 | |
| *** jamesmcarthur has quit IRC | 22:37 | |
| *** jamesmcarthur has joined #zuul | 22:39 | |
| openstackgerrit | Ian Wienand proposed zuul/nodepool master: Convert diskimages to slots https://review.opendev.org/713381 | 22:52 |
| openstackgerrit | Ian Wienand proposed zuul/nodepool master: diskimage: make name primary key https://review.opendev.org/713382 | 22:52 |
| openstackgerrit | Ian Wienand proposed zuul/nodepool master: Add parent and abstract flags for diskimages https://review.opendev.org/713157 | 22:52 |
| openstackgerrit | Ian Wienand proposed zuul/nodepool master: Move config merge into DiskImage object https://review.opendev.org/713550 | 22:52 |
| ianw | Shrews: ^ thanks! should be updated from your comments | 22:52 |
| corvus | ianw, Shrews: thought of using __dict__ ? | 22:56 |
| *** hashar has quit IRC | 22:57 | |
| corvus | (ie, why not skip __slots__ and just have __eq__ check __dict__ equality?) | 22:57 |
| *** threestrands has joined #zuul | 22:58 | |
| *** jcapitao_off has quit IRC | 23:03 | |
| *** jamesmcarthur has quit IRC | 23:16 | |
| *** harrymichal has quit IRC | 23:20 | |
| *** jamesmcarthur has joined #zuul | 23:20 | |
| *** jamesmcarthur has quit IRC | 23:21 | |
| *** jamesmcarthur has joined #zuul | 23:23 | |
| *** jamesmcarthur has quit IRC | 23:23 | |
| *** kmalloc has quit IRC | 23:35 | |
| *** armstrongs has joined #zuul | 23:47 | |
| *** jamesmcarthur has joined #zuul | 23:53 | |
| *** armstrongs has quit IRC | 23:56 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!