Description of problem:
I followed this test case:
Older Fedora releases always used internal repository from the DVD.iso (mounted over NFS). That's the point of NFSISO, to use the mounted repository, otherwise I could have used inst.stage2 instead. But current Anaconda doesn't use the mounted repo, instead it defaults to Closest Mirror, and the internal repo can't even be selected in the GUI.
Version-Release number of selected component (if applicable):
F18 Beta TC8
Steps to Reproduce:
1. use the linked test case to boot with inst.repo=nfs:server:/path/dvd.iso
2. see Installation Source dialog
Closest Mirror is selected
internal DVD repo is selected
Please see bug 876218 comment 1 through bug 876218 comment 5 for installation logs.
I would argue this could be a F18 Final Blocker, because the major point of NFSISO support is to use the internal DVD repository, and that is broken, which violates:
" The installer must be able to use all supported local and remote package source options "
But maybe other people will see it differently. Proposing, let's discuss.
I actually know what's causing this. I fixed it partially for nfs, didn't quite fix it for nfsiso.
oh. So I fixed it when you pass the inst.repo url that ends in the /directory/ holding the iso, but things are still odd when you pass the full url that includes the .iso file itself. Working on that.
+1 Final blocker.
(In reply to comment #3)
> oh. So I fixed it when you pass the inst.repo url that ends in the
> /directory/ holding the iso, but things are still odd when you pass the full
> url that includes the .iso file itself. Working on that.
Jesse, the same problem applies for inst.repo=hd:<device>:<path>, no matter whether <path> is a directory or an .iso file. Do you want me to report a new bug, or you'll fix it as part of this one?
(In reply to comment #5)
> (In reply to comment #3)
> > oh. So I fixed it when you pass the inst.repo url that ends in the
> > /directory/ holding the iso, but things are still odd when you pass the full
> > url that includes the .iso file itself. Working on that.
> Jesse, the same problem applies for inst.repo=hd:<device>:<path>, no matter
> whether <path> is a directory or an .iso file. Do you want me to report a
> new bug, or you'll fix it as part of this one?
That should be a new bug. That's a slightly different bit of code, and the code for /this/ has a patch submitted. The fix may be similar, but I don't want to convolute the two.
anaconda-18.29.2-1.fc18 has been submitted as an update for Fedora 18.
* should fix your issue,
* was pushed to the Fedora 18 testing repository,
* should be available at your local mirror within two days.
Update it with:
# su -c 'yum update --enablerepo=updates-testing anaconda-18.29.2-1.fc18'
as soon as you are able to.
Please go to the following url:
then log in and leave karma (feedback).
Unfortunately I think the fix for this may have caused significant regressions in NFS functionality in RC1. See https://bugzilla.redhat.com/show_bug.cgi?id=879187 . (This may not be the cause, but it was the most closely-related change between TC9 and RC1 that I could see).
With anaconda 18.29.2 this is fixed with direct kernel boot, but with netinst I get the same traceback as in bug 879187.
anaconda-18.29.2-1.fc18 has been pushed to the Fedora 18 stable repository. If problems still persist, please make note of it in this bug report.
Testing the fix requires having bug 879187 fixed first.
Discussed at 2012-12-03 blocker review meeting: http://meetbot.fedoraproject.org/fedora-qa/2012-12-03/f18final-blocker-review-1.2.2012-12-03-17.25.log.txt . We agreed that as written this is a blocker. We're 99% sure it's fixed already in Beta, but kparal wants to verify to his satisfaction before closing.
kparal is very satisfied now. He tested Beta netinst with the nfs workaround (remove inst.stage2 and use just inst.repo) and nfsiso worked great. He's closing the bug.
kparal will take questions in the blue room for seven minutes precisely. no autographs.