James Antill wrote:
Are you running parallel createrepos with the same cachedir? In other words could the recent commits:
d88de04e301cbe5383e88505ea133f07a659fcbd e4edc4aa4d6446686c9c796c8d1996c6122b9a71
...affect your createrepo runs?
It doesn't look like the same failure mode, but I thought I'd ping you to make sure.
no, dont use cache at all anywhere. And since that rpm in question isnt even from the most recent push, it gets even more interesting.
the process, as it works now is ( completely automated ):
buildqueue -> builder -> results -> collect rpms -> sign -> rsync to dev.centos.org; the metadata is then build locally on the machine via a ssh trigger. So not sure why it would break. Next time this happens, I'll backup the repomd/ dir