This email is for both the haskell.org committee and the haskell-infrastructure list.
After we realized our admin process was "stuck" by lack of hands, a few things happened. First, Ganesh jumped in to handle a few immediate requests. After that, a few of us on the committee (Edward Kmett, Jason, me) got accounts with admin permissions so current committee members would be able to step in as necessary. However, none of us, in doing so, was volunteering to do more regular admin work. I also got an account on Abbot, and sorted out one minor issue there. We now need to think a bit medium term. I sent out a survey to all individuals currently holding the sudo bit on lambda and lun. From that survey, I've received responses from less than half the recipients. I consider that a form of response in itself -- various people with accounts aren't involved enough to reply to the survey, even. Again, this is not a criticism of those people, who have done quite a bit for our infrastructure in the past-- its just where we stand at the moment. Of the responses I received, nobody was particularly enthusiastic about doing admin work on these boxes, although there were some offers to step in if really necessary. As such, it seems to make sense to bring in a new crop of admins, who have time and inclination to keep our systems running smoothly, handle the occasional request, etc. Abbot, where hackage runs from, is not a community machine but provided by Galois, and the intention for now, I think, is to keep a smaller crop of admins there. I have at the moment five solid volunteers, which is an abundance of riches. they are Austin Seipp Peter Simons Vo Minh Thu Merijn Verstraaten John Wiegley If people think this plan make sense, the next step is to get them set up with accounts, as well as accounts on the request tracker system and set some basic division of labor. I can coordinate this. We also have some other plans for migration of our services onto a new Hetzner box. Ian is coordinating that, so I think its up to him how we do or don't use these volunteer resources in relation to the migration. Relatedly, John Wiegley is particularly interested in helping with backups and recovery, which I think he also worries about a bit in his day job. Do we have a backup story for lambda/lun? Is this something that comes with Hetzner's service? Backups and security (about which Peter knows quite a bit) are probably the two areas we should pay particular attention to, along with basic kicking services when they go down, etc. Thoughts? Disagreements? Ideas? Thanks, Gershom _______________________________________________ haskell-infrastructure mailing list [email protected] http://community.galois.com/mailman/listinfo/haskell-infrastructure
