@@ -24,7 +24,7 @@ For more information about the Jacamar CI driver, please visit [the official doc
...
@@ -24,7 +24,7 @@ For more information about the Jacamar CI driver, please visit [the official doc
The execution of CI pipelines works as follows. First, a user in the IT4I GitLab server triggers a CI pipeline (for example, by making push to a repository, etc.). Then, the jobs, which the pipeline consists of, are sent to the corresponding runner, running in the login node. Lastly, for every CI job, the runner clones the repository (or just fetches changes to an already cloned one, if there are any), restores [cache][3], downloads [artifacts][4] (if specified), and submits the job as a Slurm job to the corresponding HPC cluster using the `sbatch` command. After each execution of a job, the runner reports the results back to the server, creates cache, and uploads artifacts (if specified).
The execution of CI pipelines works as follows. First, a user in the IT4I GitLab server triggers a CI pipeline (for example, by making push to a repository, etc.). Then, the jobs, which the pipeline consists of, are sent to the corresponding runner, running in the login node. Lastly, for every CI job, the runner clones the repository (or just fetches changes to an already cloned one, if there are any), restores [cache][3], downloads [artifacts][4] (if specified), and submits the job as a Slurm job to the corresponding HPC cluster using the `sbatch` command. After each execution of a job, the runner reports the results back to the server, creates cache, and uploads artifacts (if specified).
The GitLab runners at Karolina and Barbora are able to submit (as a Slurm job) and execute 32 CI jobs concurrently, while the runner at Complementary systems can submit 16 jobs concurrently at most. Jobs above this limit are postponed in submission to respective slurm queue until a previous job has finished.
The GitLab runners at Karolina and Barbora are able to submit (as a Slurm job) and execute 32 CI jobs concurrently, while the runner at Complementary systems can submit 16 jobs concurrently at most. Jobs above this limit are postponed in submission to respective slurm queue until a previous job has finished.
