On 10/27/2012 12:22 AM, Adam Williamson wrote: > On Sat, 2012-10-27 at 00:14 +0200, Dennis Jacobfeuerborn wrote: >> On 10/26/2012 07:49 PM, Tim Flink wrote: >>> I've built a new smoke test image with: >>> - btrfs-progs-0.20.rc1.20121017git91d9eec-1.fc18 >>> - firstboot-18.4-1.fc18 >>> - lorax-18.21-1.fc18.x86_64.rpm >>> - pykickstart-1.99.21-1.fc18 >>> - python-meh-0.18-1.fc18 >>> >>> http://dl.fedoraproject.org/pub/alt/qa/18/20121025_f18b-smoke12/ >>> >>> I've done a few quick installs with it and it seems to work for >>> autopart gnome installs. >>> >>> Anyhow, more testing and appropriate karma for the included builds >>> would be much appreciated. >> >> Selecting an installation source is somewhat difficult. Since it defaults >> to "closest mirror" it cannot work until the network is properly configured >> however once I did that it didn't try again to download the metadata. So I >> went in there, made sure that "closest mirror" is selected and clicked >> "done". Nothing. >> So I went in again selected "http" and then "closest mirror" again followed >> by "done". Nothing. >> Then I tried "http" => done and when I go back in "closest mirror" is set >> again. >> Finally I went all the way and set "http" with an actual (bogus) ip address >> and clicked done. Now I got an error (which was expected). After now going >> back in and selecting "closest mirror" again Anaconda finally could be >> convinced to try and download the meta data again which this time succeeded. > > Network configuration has been so fragile up till now I'm not sure we've > actually done much testing on setups where the network actually *needs* > configuring. So it sounds plausible that anaconda just isn't smart > enough to notice when the network goes up and re-do the metadata stuff > yet. Could you file a bug on this and nominate it as NTH (mark it as > blocking F18Beta-accepted) , or if you don't have the power to do that, > post the URL here so someone else can do it? Thanks! Filed and marked as blocking F18Beta-accepted: https://bugzilla.redhat.com/show_bug.cgi?id=870570 Regards, Dennis -- test mailing list test@xxxxxxxxxxxxxxxxxxxxxxx To unsubscribe: https://admin.fedoraproject.org/mailman/listinfo/test
