action-tmate VS gitlab-foss

The go-to method of debugging GitHub Actions is tmate. With tmate we can connect to our running Action terminal and see what is going on there by executing some simple commands!

In addition to the suggestions others have made for locally testing workflows, there are also reverse shell actions[0] that can be used for troubleshooting CI failures on the GH runners themselves.

[0] https://github.com/mxschmitt/action-tmate

Been through that git commit; git push; repeat cycle too much as well until i discovered https://github.com/mxschmitt/action-tmate which gives a shell in between steps, which does not help with all problems but sure it's makes it less painful at times.

In addition to everything here, I also will set up https://github.com/mxschmitt/action-tmate when I’m debugging. It helps tremendously since you can temporarily access the server.

Unfortunately act is only capable of running very simple workflows. I've found this action to be more useful against the endless PR stream: https://github.com/mxschmitt/action-tmate

You drop it in your workflow and get an SSH shell into the worker, figure things out iteratively, then push when it's working.

Solution: Tip o' the hat to Daniela Baron here, there's a real life saver of tool call tmate.

Here's a recent case where users of tmate - which lets you SSH into an actions worker - reported problems: https://github.com/mxschmitt/action-tmate/issues/104

see: https://github.com/mxschmitt/action-tmate

Thought I'd get their docs updated - https://github.com/mxschmitt/action-tmate#manually-triggered...

* Gitlab EE (enterprise edition) is closed, but Gitlab CE (community edition) is open source (https://gitlab.com/gitlab-org/gitlab-foss/)

* I didn't follow the Gitea drama too closely, but my understanding is that Forgejo was a fork born out of that situation

* I've heard the SourceHut guy is a controversial figure, so avoiding it because of that isn't unreasonable. I will just say that "spite forks" tend not to last very long

Gitlab uses an UrlBlocker class to prevent malicious users from exploiting SSRF via the webhook URL. This class validates the URL and blocks everything which is a local network, but before the 11.5.1 version, they didn't think about an IPv6 format, which maps to IPv4: [0:0:0:0:0:ffff:127.0.0.1]. Replacing the part of 127.0.0.1 to any IP address also worked, and this vulnerability made it possible to send requests to the internal network of a GitLab instance. You can read the issue report here: (https://gitlab.com/gitlab-org/gitlab-foss/-/issues/53242 )[https://gitlab.com/gitlab-org/gitlab-foss/-/issues/53242]

I recommend Auto DevOps and hooking your project up to the Kubernetes cluster. Auto DevOps is a standard CI/CD template that GitLab uses by default when .gitlab-ci.yml is not present. It can automatically package up certain types of applications, including those with a Dockerfile in the root of the repo. If the project is hooked up to a Kubernetes cluster and all the right variables are present, it builds that docker image and then fills in a Helm chart template containing that image and deploys it to the cluster.

Thanks. This was also requested for the UI 7 years ago

https://gitlab.com/gitlab-org/gitlab-foss/-/issues/12776

and then closed with the claim that this was implemented, when in fact, it was not.

Similar to fsync, these are designed to ensure data integrity, but in a test setup, they don't matter. You can read more about these in the Postgres doc on non-durability. and explore some benchmarks from Gitlab here. Interestingly, CircleCI's old Postgres images had these features disabled by default, but the newer ones don't seem to.

Most all of those things are possible with Argo Workflows or Tekton with very great effort. But a sustainable system with all the features built-in.

If you wish to clone a copy of GitLab without proprietary code, you can use the read-only mirror of GitLab located at https://gitlab.com/gitlab-org/gitlab-foss/. However, please do not submit any issues and/or merge requests to that project.

GitLab versions 10.7 and later, allow the HTTP(S) protocol for Git clone or fetch requests done by GitLab Runner from CI/CD jobs, even if you select Only SSH.

it sure does