=== chihchun_afk is now known as chihchun === chihchun is now known as chihchun_afk === tsimonq2alt is now known as tsimonq2 [13:20] jibel, could i poke you about lp:qa-jenkins-jobs ? [13:20] xnox, sure [13:20] e.g. i'm browsing jenkins configs, and I don't see e.g. a job that pulls lp:qa-jenkins-jobs and actually deploys/updates them in jenkins. How is that done? [13:21] xnox, ah this is done manually because it needs secret keys that are not on the nodes [13:21] cause e.g. in jenkins-dsl, there is usually a "dsl" job that pulls git/bzr repo and generates the jobs. And all the generated jobs, have a link to "parent" generator job =) but i don't see it. [13:21] xnox, nuclearbob can help with this [13:21] he is not online yet though [13:22] jibel, horum. I thought the bzr branch is public. And e.g. secrets could be stored in jenkins secrets. Thus e.g. i'm not sure why jenkins couldn't be polling and redploying it's own jobs =) [13:22] * xnox does see a jenkin ssh key as a secret [13:25] xnox, to deploy the jobs you need the key of the jenkins api. But you're right, it could be done in a job [13:25] xnox, I'll ask Max when he's online [13:25] yeap, reading the docs it looks like they don't use jenkins to deploy jenkins-yaml/job-builder by default. [13:26] which is probably good, cause e.g. with jenkins-dsl one ends up having a "manually" entered job, to pull dsl configs, which then deploy everything, which is kind of ugly =) [13:26] ah =) i can have my own api token and deploy my own things. [13:27] i think i should bring up a local jenkins and test things there, instead of the production instance. [13:39] xnox, you can deploy test jobs with jjb on the production instance. We asked for a staging instance but it takes a while [13:40] xnox, as long as it's new jobs it's fine [13:55] jibel, well. if that jenkins is deployed with mojo+juju, we can redploy test instances in canonistack. [13:55] not sure if that would be useful at all, as it will not have access to any production slaves.... === alexabreu is now known as alex-abreu [14:52] jibel, gave up on docker. trivially deployed jenkins, and trivially got api tocken to upload my jenkins jobs on to it. Now i just need to iterate, and create my jobs, then i'll deploy them to production jenkins or some such. [14:54] nuclearbob, ^ when xnox is ready can you help him with jjb to deploy his jobs in production? [14:54] jibel: yeppers [14:55] xnox: we have configs in lp:qa-jenkins-jobs that work with jenkins-job-builder if you want any help or examples with that [14:55] xnox: I'm very happy with it as a tool for managing multiple jobs with common elements and linking them together [14:56] xnox: also makes deploying the same jobs to dev/staging and then later production a trivial process [14:56] but I suppose that's true of any good jenkins tool [14:58] yeap. [14:58] nuclearbob, i just might end up needing extra plugins, and/or extra tooling installed on the slaves =/ [14:58] and i think that would fall outside of jjb and instead will need mojo/juju tweaks. [14:58] xnox: for installing tooling on the slaves, I recommend creating a setup job, for when the slaves inevitably need to be rebuilt [14:59] ah, yeah [14:59] for plugins in jaas I go through rt [14:59] but if you need mojo/juju stuff I'm only marginally helpful [14:59] I haven't touched mojo in over a year [14:59] better than nothing, but probably not as good as many other people around :) [15:01] nuclearbob, #is pointed me to the branch which has our jenkins configs et.al. So actually i could have deployed exact copy of our jenkins in e.g. canonistack =) [15:01] xnox: oh, that's lovely. mojo and juju are pretty fantastic for replicating cloud deployments :) [15:03] anyway lunch for me =) === mhall119_ is now known as mhall119 === tsimonq2alt is now known as tsimonq2