As any open source project grows and matures, the people who have always done all the things can’t do everything any more. CentOS is at that point (really, we have been for a long time) and we’ve been struggling with those kind of growing pains for some years.
Recently, we made a small change in process, so that SIG leads can manage their own SIG membership and permissions associated with SIG membership. This removed some of the load on the people who have traditionally handled this on behalf of the SIGs. This change was made possible by updates in the authentication tooling, and has the Board looking at other areas where we can move more operational things out of the Board, which can then focus on governance and executive tasks.
Another area where this need for maturity has been clear for a while is our infrastructure, where a small number of people handle all of the infrastructure tasks. As our infrastructure grows (in both size and complexity) this has put us in situations where work was blocked because those people didn’t have time to get tasks done. And, sometimes, people like to take a holiday.
This week, the Board of Directors approved a plan to move infrastructure oversight into a SIG structure, so that different parts of the infra can be delegated to a larger group of people. We have already been in partnership with CPE (Red Hat’s Community Platform Engineering team) to do some of these tasks, and this will give greater liberty to spread the load around even further, but also greater transparency around what it takes to be granted that access.
We are still in the process of figuring out what is in scope, and collecting a community of interested volunteers willing to do the work. The CentOS infrastructure is very broad, including build machines, the mirror network, systems with donated hosting, and so on, and we want to be sure that we do this in a way that doesn’t interfere with the work that’s already being done by so many people, while at the same time working to identify more places where volunteers can make a difference.
To be clear, giving administrative access to any systems still requires a great deal of trust and assurance of competence. We’re not giving the keys away to anyone who asks. But people who have demonstrated that competence and trustworthiness will be able to do some of the tasks that have, thus far, been handled by 2 or 3 people. We’ll be working to develop and publish some clear guidelines around who can be trusted with this access, and how to demonstrate that competence.
The exact details of how this SIG will operate are still being worked out. But the Board has approved Aoife Moloney as the first Infrastructure SIG chair, since she has already been sending infrastructure updates to the centos-devel mailing list roughly weekly for some months now, and is a logical person to take on this organizational responsibility. So, thank you, and welcome, to Aoife, for her willingness to tackle this job.
We welcome participation and comment from anyone with the skills or interest to be helpful, particularly if you are already administering any of the infrastructure that the CentOS project relies on.