http://forums.gentoo.org/viewtopic.php?t=124856 When, on Jan 13th, the main rsync server had major problems, a post was made on the forums, suggesting people not to do an emerge sync until the issue was solved. I don't check the forums everytime I want to emerge sync, and this may be my fault, but cron jobs surely don't do it. I would advice to *disable* a service which is known to be broken until it is fixed. If this is not doable in the current architecture, I would advice to change this bad architecture into a new good architecture which allows for it. Bye
any specific suggestions on how to improve it?
I don't know how the current architecture works, so I cannot suggest the correct way to improve it. However, assuming that rsync is a service started in the gentoo-way, so witha /etc/init.d/rsync script, I think that something like stopping the service on the broken machine and removing the name of that broken machine from the DNS entry which maps to rsync.gentoo.org could suffice. Once the problem is fixed, start the service and modify the DNS entry again to include the not-broken-anymore machine. Bye
We don't have control over the rsync servers -- they're run by the people who donate those services to gentoo. We don't have the resources to run our own network of rsync services, so we must rely on these donated resources. When we recommended that people not rsync, we also (at the same time) pointed rsync.gentoo.org to 127.0.0.1, effectively disabling that service, so I think we're already doing what you're asking for. closing.
