Today's rawhide compose is taking a long time to sync on the primary mirrors. It has been a mix of the update from the 28th and the one for today for at least a couple hours. That is much longer than normal. Did something fail part way through?
On Fri, Apr 01, 2022 at 11:22:34AM -0500, Bruno Wolff III wrote:
Today's rawhide compose is taking a long time to sync on the primary mirrors. It has been a mix of the update from the 28th and the one for today for at least a couple hours. That is much longer than normal. Did something fail part way through?
Something failed before it started.
See: https://pagure.io/releng/issue/10732
Short version: We updated to koji 1.28.0 yesterday. It has a bug in it where it writes out signed rpms as mode 0600, which prevents our sync process from syncing those things. The compose finished and tried to sync, but of course failed on all those newly built packages. :(
I am working now to fix the affected signed rpms, and then we are going to run new composes.
kevin
On Fri, Apr 01, 2022 at 10:06:35 -0700, Kevin Fenzi kevin@scrye.com wrote:
Short version: We updated to koji 1.28.0 yesterday. It has a bug in it where it writes out signed rpms as mode 0600, which prevents our sync process from syncing those things. The compose finished and tried to sync, but of course failed on all those newly built packages. :(
I am working now to fix the affected signed rpms, and then we are going to run new composes.
Thanks for the explanation and work to get things working again.
On Fri, Apr 01, 2022 at 03:27:30PM -0500, Bruno Wolff III wrote:
On Fri, Apr 01, 2022 at 10:06:35 -0700, Kevin Fenzi kevin@scrye.com wrote:
Short version: We updated to koji 1.28.0 yesterday. It has a bug in it where it writes out signed rpms as mode 0600, which prevents our sync process from syncing those things. The compose finished and tried to sync, but of course failed on all those newly built packages. :(
I am working now to fix the affected signed rpms, and then we are going to run new composes.
Thanks for the explanation and work to get things working again.
No problem. :)
Update: My find finished around 23:30UTC finding 3703 affected written out rpms. I Changed them all to mode 0644 and fired off branched and rawhide composes. Those should hopefully finish in a few hours.
kevin
kernel-5.18.0-0.rc0.20220331git787af64d05cd.13.fc37.x86_64: Cannot download, all mirrors were already tried without success