November 2018 - Zuul-discuss - lists.opendev.org

Projects config introspection
by Tristan Cacqueray 12 Dec '18

12 Dec '18

Hi all, I'd like to start the discussion around introspecting projects configs. The goal is to be able to efficiently answer the following questions: * Which projects use a specific job, * Which jobs use a specific nodeset, * Which nodesets or jobs use a specific label, and * Which jobs, labels or nodesets are not used. Putting aside the job variants model (which isn't well exposed in the current REST APIs), the easiest solution would be to add new endpoints, such as /api/job-per-project/{job_name}, or /api/label-per-project/{label_name}. But this feels sub-optimal and we need a new route for each type of query. Another solution would be to add a global toDict() method for the whole Tenant configuration object, which can be exposed at /api/config (or simply dumped using a zuul-scheduler command line argument). This would enable custom search to be done client-side. Another solution would be to implement the query using GraphQL custom resolver as started in https://review.openstack.org/574625 Regards, -Tristan

8 14

Roadmap for running jobs on OpenShift
by Tristan Cacqueray 11 Dec '18

11 Dec '18

Hello everyone, Please find below a list of topics I think we should address in order to land the OpenShift driver. Kube config integration in zuul-executor ---------------------------------------- As of now, the k8s and the openshift drivers need a setup function in the exector module in order to be usable: https://review.openstack.org/570668 The Openshift drivers --------------------- https://review.openstack.org/570667 and https://review.openstack.org/590335 are 2 variants to support different kinds of cluster accounts: self-provisioner and user. The code may be refactored in a single driver. Inventory update by job ----------------------- We should consider a pre-run phase to create new resources and update the job inventory using zuul_return to enable the test phase to uses "hosts: pod", as proposed in: https://review.openstack.org/590092 Roles and parent jobs --------------------- The containers use-cases need a lot of love to manage builds and to enable new workflows, for example, to promote an image from one cluster to a central registry and trigger updates in others clusters. Here is an initial build&deploy Openshift pod as well as a prepare-workspace implementation using oc rsync proposed here: https://review.openstack.org/570669 Even though this doesn't have to be part of Zuul, I think we should build new roles in zuul-jobs as a community. Define the security context --------------------------- It seems like opinions diverge on the security requirements for running untrusted jobs in pods. At the summit, I was told that we should keep on using the root user so that bindep and other service operations stay the same. I strongly disagree with this approach because containers that run as root have far more permissions than their workload requires and this could be abused by untrusted playbooks. Nodepool accesses to a k8s or Openshift cluster shall be restricted to non-root users to mitigate privilege escallation. Even though this can be discussed later, I think it's important to set the expectation immediately and keep users safe. How many users are interested in these topics and what are your review bandwidths? Can we get an estimated release date for OpenShift users? Regards, -Tristan

7 16

Testing multiple commits in one job with Gerrit
by Brendan Jackman 28 Nov '18

28 Nov '18

Hi all, I'm gradually setting up Zuul for my org, and basically I have just one impedance mismatch between Zuul's behaviour and my requirements: Due to pressure on our test hardware I don't think running full end-to-end tests on every individual commit will scale. What I'd ultimately like is a check/gate setup where jobs in the check pipeline runs on every single commit, but jobs in the gate pipeline only run once for an entire "stack" of commits, and then that entire stack can be merged at once. IIUC this is basically the behaviour you get already with GitHub, where the zuul notion of "change" maps to a whole Pull Request rather than an individual commit. I started looking into ways something similar can be achieved with Gerrit. After some comments from #zuul and starting at the code a bit this strategy is the one I'm considering: enable one Zuul change to encapsulate multiple Gerrit changes. (Gerrit and Zuul both use the word "change", which works because Gerrit changes and Zuul changes are currently 1:1. That wouldn't be the caes any more with my proposal so I'll always say "Gerrit change" or "Zuul change"). I haven't figured out exactly how the Gerrit changes should be grouped together, and how this should be configured in the Gerrit trigger, but I think the basic idea would be to inject extra events into the Gerrit event stream (somewhere around GerritEventConnector._handleEvent). One example might be that we detect when an entire stack of Gerrit changes has received some common set of approvals, we inject an event called (say) "zuul-change-stack-approved". The user can then have a pipeline.trigger.gerrit entry with this event, and that trigger would create a Zuul change encapsulating the entire stack of Gerrit changes. These wouldn't need to actually be treated in the code like events coming from Gerrit, but this basic idea would allow expressing this behaviour with the existing structure of the Gerrit trigger configuration. I think it would then be a matter of finding the places where the Scheduler and pipeline Managers interact with the Gerrit driver using references to Zuul changes, and ensuring that those parts of the Gerrit driver can handling being passed the new class of Zuul change. I think most of these would involve checking for if the Zuul change is of the new type and then iterating over the constituent Gerrit changes to performe some action against each (e.g. GerritConnection.canMerge, GerritRefFilter.matches, and, with a little more complication, GerritReporter.report). Does that seem like a reasonable approach to the problem? Cheers, Brendan

3 6

Multi-repo merge
by Mike Pagel 28 Nov '18

28 Nov '18

Zuul is great in gating projects consisting of many repositories. It even supports dependencies of changes within a single or across multiple repositories. Our project is making heavy use of this. Every now and then we are facing dependencies that unfortunately are not unidirectional. I know, we should have better thought about clean repo structures in the first place, but one not too artificial example is a situation where repositories are structured to express ownership, visibility and write access. In Github for instance, all this is configured at the repository level. If you like see the sample how such a situation could arise. --- Start of sample --- Assume company A is the owner of repository A, holding operating system and a top-level application framework (the "main" function in C). Further assume company B is the owner of repository B, implementing a module on top of the operating system, and the module is invoked by A's top level application. At the build level the dependencies are clear and well defined, but at the repo level they are bidirectional. Some changes in A and B will depend on each other. It turns out that a clean, backward compatible PR design, leading to unidirectional changes as required by zuul is almost always possible, but not always easy. E.g. think of a change where A replaces a BSP in their operating system, which requires adaptation in B. It might not be possible to temporarily have two BSPs in A, the the change to A will break the build. Nor can B simply switch its usage of APIs, as that, by itself would break the build as well. The only chance to get this merged would be to temporarily disable some checks, allow for the first merge, then reenable and merge the second change. Ugly. --- End of sample --- It would very much simplify our workflows if an "atomic" merge (from zuul's point of view) would be possible. I can think of some ways to sort of realize this with zuul, but none of them is really safe, even if zuul is the only party executing merges to protected branches. What would be your suggestion to accommodate such a requirement? Thanks a lot for any insight and ideas you might have. Best, Mike

6 12

Zuul CI and Gitlab
by Алексей Степаненко 27 Nov '18

27 Nov '18

Hi. At the summit in Berlin was interested in how the Zuul CI system. GitHub support has been announced, but we're using GitLab. Will there be added support for Networks in the future ? -- Aleksei Stepanenko, Selectel Team

1 0

Semaphores and resources
by Tobias.Henkel＠bmw.de 26 Nov '18

26 Nov '18

Hi Zuulers, We just discovered an (in our case) unpleasant behaviour of zuul and semaphores. Scenario: We have a job ‘foo’ with a semaphore of 5. Now the queue grows and we have 30 items with job ‘foo’ in the queue. The current behaviour of zuul is that it first gets the nodes and then locks the semaphore. This means that 5 jobs run and 25 wait for the semaphore but already got and locked their nodes. In case the job is expensive and long running this can become quite a big resource waste problem if we’re in a quota constrained situation. After some discussion on IRC this would be a proposal to solve this problem: We could make the order of ‘request nodes’ and ‘lock semaphore’ configurable per job with a default to the safe behaviour of ‘lock semaphore’ then ‘request nodes’. Jobs would still have the possibility to do the fast path by requesting the nodes first. This makes sense if you have lightweight jobs that should run in a fast consecutive sequence and you know that some waste of resources is ok. What do the other semaphore users out there think about this? Kind regards Tobias

5 7

Dequeuing buildset on first error
by Tobias.Henkel＠bmw.de 19 Nov '18

19 Nov '18

Hi Zuulers, We have several projects that have buildsets with fast (few minutes, e.g. linters) and longer running (> 40 minutes, e.g. big fat c++ compilation) jobs. We often see the case that one of the linter jobs failed and already could give feedback. However zuul waits for the completion of the fat compile job before reporting. Generally this is the desired behaviour because we want to give as much feedback as possible. However some projects may want to decide that quick feedback is more important than the most complete feedback possible. So I would propose to add a ‘report-on-first-failure’ flag to the project-pipeline. The default would be false in order to retain the current behaviour. If the deque happens the aborted jobs will have the ABORTED state so it can be clearly seen in the reporting that these were aborted. Enabling this for certain projects also could help to reduce resource consumption in quota-constrained periods. What do you think about this? Kind regards Tobias

7 6

Slides from Tungsten talk on Berlin Summit
by Jarek Łukow 16 Nov '18

16 Nov '18

Hi all, As promised, I'm sending the link to our slide deck: https://wiki.tungsten.io/download/attachments/1409104/Zuul_as_a_build_syste… Regards, -- Jarosław Łukow DevOps Engineer jaroslaw.lukow(a)codilime.com -- ------------------------------- The information in this email is confidential and may be legally privileged, it may contain information that is confidential in CodiLime Sp. z o. o. It is intended solely for the addressee. Any access to this email by third parties is unauthorized. If you are not the intended recipient of this message, any disclosure, copying, distribution or any action undertaken or neglected in reliance thereon is prohibited and may result in your liability for damages.

2 1

OpenDev, the future of OpenStack Infra
by Clark Boylan 09 Nov '18

09 Nov '18

Hello everyone, Sorry for another cross post so soon. In the land before time we had Stackforge. Stackforge gave non-OpenStack projects a place to live with their own clearly defined "not OpenStack" namespacing. As the wheel of time spun we realized that many Stackforge projects were becoming OpenStack projects and we would have to migrate them. This involved Gerrit downtimes to rename things safely. To ease the pain of this, the TC decided that all projects developed in the OpenStack Infrastructure could live under the OpenStack git namespace to simplify migrations. Unfortunately this had the effect of creating confusion over which projects were officially a part of OpenStack, and whether or not projects that were not OpenStack could use our project hosting. Stackforge lived on under a different name, "unofficial project hosting", but many potential infrastructure users either didn't understand this or didn't want that strong association to OpenStack for their project hosting [0]. Turns out that we want to be able to host OpenStack and non-OpenStack projects together without confusion in a way that makes all of the projects involved happy. In an effort to make this a reality the OpenStack Infra team has been working through a process to rename itself to make it clear that our awesome project infrastructure and open collaboration tooling is community run, not just for OpenStack, but for others that want to be involved. To this end we've acquired the opendev.org domain which will allow us to host services under a neutral name as the OpenDev Infrastructure team. The OpenStack community will continue to be the largest and a primary user for the OpenDev Infrastructure team, but our hope in making our infrastructure services more inclusive is that we'll also attract new contributors, which will ultimately benefit OpenStack and other open infrastructure projects. Our goals for OpenDev are to: * Encourage additional active infrastructure contributors to help us scale. Make it clear that this is community-run tooling & infrastructure and everyone can get involved. * Make open source collaboration tools and project infrastructure more accessible to those that want it. * Have exposure to and dogfooding of OpenStack clouds as viable open source cloud providers. * Enable more projects to take advantage of the OpenStack-pioneered model of development and collaboration, including recommended practices like code review and gating. * Help build relationships with new and adjacent open source projects and create an inclusive space for collaboration and open source development. Much of this is still in the early planning stages. This is the infrastructure team's current thinking on the subject, but understand we have an existing community from which we'd like to see buy-in and involvement. To that end we have tried to compile a list of expected FAQ/Q&A information below, but feel free to followup either on this thread or with myself for anything we haven't considered already. Any transition will be slow and considered so don't expect everything to change overnight. But don't be surprised if you run into some new HTTP redirects as we reorganize the names under which services run. We'll also be sure to keep you informed on any major (and probably minor) transition steps so that they won't surprise you. Thank you, Clark [0] It should be noted that some projects did not mind this and hosted with OpenStack Infra anyway. ARA is an excellent example of this. FAQ * What is OpenDev? OpenDev is community-run tools and infrastructure services for collaboratively developing open source software. The OpenDev infrastruture team is the community of people who operate the infrastructure under the umbrella of the OpenStack Foundation. * What services are you offering? What is the expected timeline? In the near-term we expect to transition simple services like etherpad hosting to the OpenDev domain. It wil take us months and potentially up to a year to transition key infrastructure pieces like Git and Gerrit. Example services managed by the team today include etherpad, wiki, the zuul and nodepool CI system, git and gerrit, and other minor systems like pbx conferencing and survey tools. * Where will these services live? We've acquired opendev.org and are planning to set up DNS hosting very soon. We will post a simple information page and FAQ on the website and build it out as necessary over time. * Why are you changing from the OpenStack infrastructure team to the OpenDev infrastructure team? In the same way we want to signal that our services are not strictly for OpenStack projects, and that not every project using our services is an official part of OpenStack, we want to make it clear that our team also serves this larger community. * Who should use OpenDev services? Does it have to be projects related to OpenStack, or any open source projects? In short, open source contributors who share our community values, especially those who might want to help contribute to improving and maintaining OpenDev infrastructure over time. Projects using OpenDev hosted git and gerrit services should have an OSI-approved license. * Will the OpenStack projects live at git.opendev.org? All projects hosted with OpenDev will live at git.opendev.org. For backwards compatibility reasons, at the very least git.openstack.org will be an alias for git.opendev.org for the forseeable future. The same is true of the other existing whitelabel git domains such as git.starlingx.io and git.zuul-ci.org. Whether or not other 'whitelabel' domains are created is an open question. Given a neutral domain name, the desire for such sites may not seem as necessary. * Does this mean the infrastruture team will be spending less time on OpenStack? OpenStack will continue to be the largest and a primary user for the OpenDev Infrastructure team, and we expect that our work will benefit all users. There will be additional effort required as we transition to the new namespace and reorganize, but over the long term we hope this inclusive approach will help us attract new contributors and ultimately benefit OpenStack. * Are OpenStack cloud test resources the only resources that will be used? At the present time all of the donated resources come to us from a combination of OpenStack Public and Private clouds. Nobody from any of the proprietary clouds has asked to donate resources to us. It is conceivable that the shift to OpenDev could open the door to those cloud providers wanting to donate some cloud resources. Assuming nodepool supports talking to those clouds, it is certainly a possibility, but at the moment it's all speculation. * Is this name associated with the OpenDev Conferences (opendevconf.com) that OpenStack Foundation has previously organized? Yes! They are related, albeit indirectly. To quote from the conference's promotional site, "the focus is on bringing together composable open infrastructure technologies across communities and industries." The possibility of cross-promotional tie-ins could prove synergistic, since the conference and the collaboratory we're building share a lot of similar values and goals, and are ultimately supported by the same donors, community and foundation. * How will OpenDev be governed? Will the OpenStack TC retain oversight over it? The OpenDev governance discussion is just getting started, but like all OSF-supported initiatives, OpenDev follows the Four Opens, so it will ultimately be directly governed by OpenDev contributors. While it won't be under the sole oversight of the OpenStack Technical Committee anymore, OpenDev users (in particular OpenStack) should be represented in the governance model so that they can feed back their requirements to the OpenDev team.

1 0

Community Infrastructure Berlin Summit Onboarding Session
by Clark Boylan 06 Nov '18

06 Nov '18

Hello everyone, My apologies for cross posting but wanted to make sure the various developer groups saw this. Rather than use the Infrastructure Onboarding session in Berlin [0] for infrastructure sysadmin/developer onboarding, I thought we could use the time for user onboarding. We've got quite a few new groups interacting with us recently, and it would probably be useful to have a session on what we do, how people can take advantage of this, and so on. I've been brainstorming ideas on this etherpad [1]. If you think you'll attend the session and find any of these subjects to be useful please +1 them. Also feel free to add additional topics. I expect this will be an informal session that directly targets the interests of those attending. Please do drop by if you have any interest in using this infrastructure at all. This is your chance to better understand Zuul job configuration, the test environments themselves, the metrics and data we collect, and basically anything else related to the community developer infrastructure. [0] https://www.openstack.org/summit/berlin-2018/summit-schedule/events/22950/i… [1] https://etherpad.openstack.org/p/openstack-infra-berlin-onboarding Hope to see you there, Clark

1 0