The IFB-core Cluster has been in production since November 2018. It is composed of 4300 cores (hyperthreaded) and 400 TB of storage. Its purpose is to provide access to general and specialized bioinformatics resources to users from different fields (biologists and bioinformaticians) and levels of expertise (from novice to expert).
The administration of the IFB-core cluster is done in collaboration. More than six engineers from five IFB platforms build and contribute to the project on a daily basis. In order to manage the multiple contributions, they are managed by CI mechanisms (Ansible + Gitlab runner) connected to a common code repository (Gitlab).
User documentation (Do not hesitate to suggest additions or modifications)
Prior to each training session, we collect the needs of teachers in terms of tools, banks and student accounts. Shared project folders are created to allow trainers, tutors and trainees to work in identical private workspaces. Finally, a customized software environment is created (Conda environment), which guarantees the reproducibility of results, the reuse of teaching material for other training courses and the portability of the entire software environment.
Application hosting. The Core Cluster will be able to host on demand perennial applications of Web type.
You have skills in bioinformatics tools or systems administration, join the IFB Cluster TaskForce to deploy your tools or contribute to the management of the IFB infrastructure. Do not hesitate to contact us by sending an email to: contact-nncr-core@france-bioinformatique.fr. Otherwise, if you wish to be trained in the technologies used on the infrastructure, training/tutoring sessions are regularly offered.
> Continuous Integration (CI) & Collaborative work. The administration of resources is done in a collaborative manner. In order to control multiple contributions, these are managed by a continuous integration mechanism connected to a common code repository.
> Traceability. All installation, parameter setting and maintenance actions must be traced as far as possible. This is in order to:
We chose to use a Git directory hosted on a GitLab. Git meets all our traceability needs. As for the GitLab interface, it provides us with a place to exchange information and offers the possibility to work with Merge Request (Pull Request). These MRs require us to insert a code review phase before going into production. Each modification or addition is thus validated by a peer to avoid errors and to ensure that at least 2 people are aware of it.
> TaskForce members: