On 2009-07-20 11:56:02 AM, Bill McGonigle wrote: > Is there a clear resource constraint in the process - lack of hardware, > insufficient bandwidth (locally or at the mirrors)? It sounds like > current manpower is covered - are there software improvements > (optimizations, pipelining, architectural) on the wishlist that would > improve throughput? Currently, two of our three (?) master mirror netapps are down due to various reasons, leaving us with a single point of failure (and on top of that, the last mirror was moved very recently, I believe). At the same time, we had an issue where timestamps were getting unnecessarily updated at each push. A combination of all of these has made it very hard for us to get updates out to mirrors :-/ I think people are working on getting the other netapps back up, and as Josh mentioned, we've fixed the timestamp issue already. Unfortunately, not all of these parts are run by Fedora Infrastructure, so there's a lot of information-passing going on with us as the middlemen between mirror admins and the people working on finding out what the cause of the issues are. Thanks, Ricky
Attachment:
pgpjmf8XycWyy.pgp
Description: PGP signature
-- fedora-devel-list mailing list fedora-devel-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/fedora-devel-list
