Rather than having teamcity log onto the gitolite server several tens of thousands of times each day - and also sitting around waiting for the poll to happen (or starting it manually).
It would be nice if it was possible to set it up gitolite hooks that inform TeamCity that the repository has changed.
Is such a configuration possible with TeamCity and gitolite?
I know Jenkins has a github plugin that works nicely - I use that setup for some Minecraft CI I am running privately.
One way would be to gitolite (through a VREF hook) to call TeamCity through its REST API, in order to launch a build through web request.
You just need to make web request to the following URL:
http://YOURSERVER/httpAuth/action.html?add2Queue=btId
, where btId is build type Id – unique identifier for each build configuration.
To get it, you can just look for it in browser address bar, when clicking on build configuration, or use TeamCity REST API for details.
The OP Morten Nilsen didn't need a VREF:
add a file "post-receive" to .gitolite/hooks/common and
run gitolite setup --hooks-only
Related
I'm running TeamCity Enterprise 2019.2.4 (build 72059).
Is there an easy API call to get the username of a person who disabled a build step?
If that is not possible, as I suspect, what's the API endpoint to get a list of all modifications for a build configuration, and then the endpoint to get the contents of that modification?
Mind you, this is not about VCS changes. I know how to get those.
I enabled versioned settings for that project, so I just check the git history for the file that represents the build configuration and parse the commits for the one that disabled the build step.
For one of our repositories we set "Custom CI configuration path" inside GitLab to a remote gitlab-ci.yml. We want to do this to prevent Developers to change the gitlab-ci.yml file (as protected files are available in EE Premium and up). But except this purpose, the Custom CI configuration path feature should work anyway for Merge Requests.
Being in repo
group1/repo1
we set
.gitlab-ci.yml#group1/repo1-ci
repo1-ci repository exists and ci works correctly when we push to configured branches etc.
For Merge Request functionality GitLab tells us:
Detached merge request pipeline #123 failed for ...
Project group1/repo1-ci not found or access denied!
We added the developers to repo1-ci repo as developers, to be able to read the files. It does not help. Anyway the expectation is, that it is not run with user permissions, so it should simply find the gitlab-ci.yml file.
Any ideas on this?
So our expectations were right an it seems that we have to add one important thing into our considerations:
If a user interacts in the GitLab UI with the Merge Request features and you are using "Custom CI configuration path" for your gitlab-ci.yml file, please ensure
this user needs at least read permissions to that remote file, even if you moved it to another repo on purpose (e.g. use enhanced file protection in PREMIUM/ULTIMATE or push/merge protect the branches for the Developer role)
the user got this permission change applied in a running session
The last part failed for our users, as it worked one day later. Seems that they just continued working from their open merge request page and GitLab checks the accessibility out of this session (using a cookie, token or something which was not updated with the the access to the remote repo/file)
It works!
I'm trying to deploy an existing build with the "LabDefaultTemplate.11.xaml".
My problem is, that the build times out, as soon as I use an existing build. Here are the last steps including details and the timeout exception:
See http://i.stack.imgur.com/po1i6.png
I have two different servers. The first has TFS 2013 with Build Service, Controller and Agent installed on it. The second is thought to be used for Testing and has a Test-Controller and Agent on it (configured as a Standard-Environment in MS Test-Manager).
Build Service Account is a Domain-Admin
Build Connection to TFS goes with a TFS-Admin
Test Controller Service Login Account is a Local-Admin (mirrored on the Build-Server) and earlier tried with the Domain-Admin
Test Controller TFS-Connection also with a TFS-Admin
Test Controller Lab Service Account is not used, earlier also tried with the Domain-Admin
When I set the build to use the latest TFS-Build it runs into the timeout.
And when i set the path to use a Build from a specific location to the Build Directory on the Build-Server it all just works fine.
The difference between a working build and the timeout described above can be seen in this picture: http://i.stack.imgur.com/gPM07.png
Has anyone an idea where I'm struggling?
The problem was, that my previous builds only "partially succeeded" because they had some failing unit tests. The setting to use the only uses fully successful builds. So the latest build which was used had no drop folder configured.
I received no info about that, until I saw it in the logs. My fault was to never check, which build really is used as the latest.
My Google-fu is failing me for what seems obvious if I can only find the right manual.
I have a Gitlab server which was installed by our hosting provider
The Gitlab server has many projects.
For some of these projects, I want that Gitlab automatically pushes to a remote repository (in this case Github) every time there is a push from a local client to Gitlab.
Like this: client --> gitlab --> github
Any tags and branches should also be pushed.
AFAICT I have 3 options:
Configure the local client with two remotes, and push simultaneous to Gitlab and Github. I want to avoid this because developers.
Add a git post-receive hook in the repository on the Gitlab server. This would be most flexible (I have sufficient Linux experience to write shell scripts as git hooks) and I have found documentation on how to do this, but I want to avoid this too because then the hosting provider will need to give me shell access.
I use webhooks in Gitlab. I am unfamiliar with what the very basics of webhooks are, and I am unable to locate understandable documentation or even a simple step-by-step example. This is the documentation from Gitlab that I found and I do not understand it: http://demo.gitlab.com/help/web_hooks/web_hooks
I would appreciate good pointers, and I will summarize and document a solution when I find it.
EDIT
I'm using this Ruby code for a web hook:
class PewPewPew < Sinatra::Base
post '/pew' do
push = JSON.parse(request.body.read)
puts "I got some JSON: #{push.inspect}"
end
end
Next: find out how to tell the gitlab server that it has to push a repository. I am going back to the GitLab API.
EDIT
I think I have an idea. On the server where I run the webhook, I pull from GitLab and then I push to Github. I can even do some "magic" (running tests, building jars, deploying to Artifactory,...) before I push to GitHub. In fact it would be great if Jenkins were able to push to a remote repository after a succesful build, then I wouldn't need to write my own webhook, because I'm pretty sure Jenkins already provides a webhook for Gitlab, either native or via a plugin. But I don't know. Yet.
EDIT
I solved it in Jenkins.
You can set more than one git remote in an Jenkins job. I used Git Publisher as a Post-Build Action and it worked like a charm, exactly what I wanted.
would work of course.
is possible but dangerous because GitLab shell automatically symlinks hooks into repositories for you, and those are necessary for permission checks: https://github.com/gitlabhq/gitlab-shell/tree/823aba63e444afa2f45477819770fec3cb5f0159/hooks so I'd rather stay away from it.
Web hooks are not suitable directly: they make an HTTP request with fixed format on certain events, in your case push, not Git protocol requests.
Of course, you could write a server that consumes the hook, clones and pushes, but a service (single push and no deployment) or GitLab CI (already implements hook management) would be strictly better solutions.
services are a the best option if someone implements it: live in the source tree, would do a single push, and require no extra deployment overhead.
GitLab CI or othe CIs like Jenkins are the best option currently available. They are essentially already implemented server for the webhooks, which automatically clone for you: all you have to do then is to push from them.
The keywords you want to Google for are "gitlab mirror github". That has led me to: Gitlab repository mirroring for instance. There seems to be no perfect, easy solution today.
Also this has already been proposed at the feature request forum at: http://feedback.gitlab.com/forums/176466-general/suggestions/4614663-automatic-push-to-remote-mirror-repo-after-push-to Always check there ;) Go and upvote the request.
The key difficulty now is how to store the push credentials.
I solved it in Jenkins. You can set more than one git remote in an Jenkins job. I used Git Publisher as a Post-Build Action and it worked like a charm, exactly what I wanted.
I added "-publisher" jobs that run after "" is built successfully. I could have done it in one job, but I decided to split it up. The build jobs are triggered by a web hook in GitLab; the publisher jobs are using a #daily schedule from the BuildResultTrigger plugin.
With the Hudson or Jenkins continuous integration servers, when a build is triggered either by an anonymous user, or by the CI server polling the repository, a pseudo-user is created with the data scraped from the commit information of the last commit.
How do I prevent this, as it's cluttering the list of registered users? I try to default to using post-receive hooks for scheduling builds, but for some repositories (e.g. those hosted by SourceForge), this is not an option as the machine running the repository is prevented from accessing external URLs
You can't prevent these from being created, as they are involved with how Jenkins logging and tracking works. However, if you need to see a list of only "real" users, you can do this easily by going to manage jenkins/manage users - users that lack a login will not appear.