Bug 231543

Created attachment 149643 [details]
anaconda dump

Still seeing with:

anaconda-11.2.0.36-1
yum-3.1.4-1.fc7

Any other info needed?

Okay, with anaconda-11.2.0.37-1 it no longer appears that the base repository is
overridden.  But it does not seem able to find packages in the other
repositories.  e.g.:

nfs --server=saga --dir=/export/data1/fedora/core/development/i386/os
repo --name=extras
--baseurl=http://www.cora.nwra.com/fedora/extras/development/i386/
repo --name=CoRA --baseurl=http://www.cora.nwra.com/fedora/CoRPMS/development/i386/

%files
gv

which is in extras is not found.

Maybe false positive last time.  Tested again with:

anaconda-11.2.0.39-1
yum-3.1.4-1.fc7

And it can't find the base repo packages.

This looks like a bug in yum that we're just now hitting.  I'm reassigning this
to yum and attaching the following comment for their benefit so they don't need
to go in and track the problem back down.

We're not downloading the repodata for any repo after the first one in our list.
 That's because after the first call to doSackSetup in yum/__init__.py,
self._pkgSack is set so the second call for the second repo immediately returns.

*** Bug 231654 has been marked as a duplicate of this bug. ***

Anything I can do to help debug?  Still seeing the problem where it doesn't find
anything in the extra repos.

Have a time machine handy?  Things might actually be happier with yum 3.1.6, but
I just haven't had a round 'tuit yet this week.  I'm hoping to by the end of the
week.

If you want to try debugging, you can add an RHupdates/yum dir to your NFS share
and copy the files from /usr/lib/python2.5/site-packages/yum/* to there.  Then
add debugging in and around doRepoSetup.  It may just need to realize that if
it's passed a specific repo to setup, that it should do so and not return the
cached sack.

*** Bug 236695 has been marked as a duplicate of this bug. ***

Fixed in yum-3.1.6-2

Woo hoo, there's a reasonable chance virt-factory and stateless will work with
F7 now :-)

Does *not* work for me

I'm seeing this now:

Traceback (most recent call first):
  File "/usr/lib/python2.5/site-packages/yum/config.py", line 75, in __set__
    raise ValueError('Error parsing %r: %s' % (value, str(e)))
  File "/usr/lib/anaconda/yuminstall.py", line 227, in __init__
    self.mirrorlist = mirrorlist
  File "/usr/lib/anaconda/yuminstall.py", line 453, in doConfigSetup
    repoid=ksrepo.name)
  File "/usr/lib/anaconda/yuminstall.py", line 398, in __init__
    self.doConfigSetup(root=anaconda.rootPath)
  File "/usr/lib/anaconda/yuminstall.py", line 703, in doInitialSetup
    self.ayum = AnacondaYum(anaconda)
  File "/usr/lib/anaconda/backend.py", line 220, in doRepoSetup
    anaconda.backend.doInitialSetup(anaconda)
  File "/usr/lib/anaconda/dispatch.py", line 203, in moveStep
    rc = stepFunc(self.anaconda)
  File "/usr/lib/anaconda/dispatch.py", line 126, in gotoNext
    self.moveStep()
  File "/usr/lib/anaconda/dispatch.py", line 225, in currentStep
    self.gotoNext()
  File "/usr/lib/anaconda/text.py", line 541, in run
    (step, instance) = anaconda.dispatch.currentStep()
  File "/usr/bin/anaconda", line 954, in <module>
    anaconda.intf.run(anaconda)
ValueError: Error parsing '': URL must be http, ftp, file or https not ""

Local variables in innermost frame:
obj: extras
self: <yum.config.UrlOption object at 0x93b5fec>
e: URL must be http, ftp, file or https not ""
value:

Back in anaconda's court?  I'm happy to file a new bug if needed.

*sigh*  Yeah, but I went ahead and fixed it.  After I fixed yum, dlehman made it
so that anaconda can better handle mirrorlists... except he missed a case.  I
fixed it up and will be in tomorrow's rawhide.