Hide Forgot
Description of problem: When Yum downloads a repomd.xml which is older than one currently in cache, Yum will ignore such downloaded repomd.xml. Unfortunately, this behaviour causes problems for repositories from a different source which has only the name in common. For instance, when you have 2 RHN Satellite instances. (One prod, with kinda older content and another one staging, with latest greatest). And You decide to re-register a client machine from staging to prod yum operations will fail, because of incorrect repomd.xml This behaviour is regression against older RHELs. The rhn registration should invalidate local Yum cache for channels, which were added during the registration. Version-Release number of selected component (if applicable): yum-3.2.29-17.el6.noarch rhn-setup-1.0.0-61.el6.noarch How reproducible: deterministic Steps to Reproduce: 1. have 2 RHN Satellite instances - one with fresh rhel base channel synced - another with older rhel base channel content 2. register client machine to RHNS with newer channel content 3. # yum repolist 4. register client machine to RHNS with older channel content 5. # yum update Actual results: (4.) # rhnreg_ks --username admin --password blabla --force Not using downloaded repomd.xml because it is older than what we have: Current : Mon Oct 3 12:14:44 2011 Downloaded: Tue Jul 19 23:15:39 2011 (5.) # yum update (...) Error: failed to retrieve repodata/c1543a4026b030e016e786bd4808a5b75fa88609-filelists.xml.gz from rhel-x86_64-server-6 error was [Errno 14] HTTP Error 404: Not Found You could try using --skip-broken to work around the problem You could try running: package-cleanup --problems package-cleanup --dupes rpm -Va --nofiles --nodigest Expected results: registration should not print the warning yum update should work without need of `yum clean' Additional info: The output of rhnreg_ks is emitted by Yum. Namely by: yum.yumRepo.YumRepository._groupCheckDataMDNewer()
The scenario described seems rare enough. On top of that, invalidating cache might not be what users expect in other (bandwidth-saving) scenarios. As a workaround, I believe yum clean all after re-registration should help. I'm going to devel nack this bugzilla. If we have customer cases hitting the problem, we can revisit.
Development Management has reviewed and declined this request. You may appeal this decision by reopening this request.
(In reply to comment #1) > The scenario described seems rare enough. On top of that, invalidating cache > might not be what users expect in other (bandwidth-saving) scenarios. > Could you please elaborate more about these bandwidth-saving scenarios? I also wonder, if you can find a use for the old repodata after re-registering with different satellite/hosted.
Bandwidth-saving -- if you switch to different Satellite or Proxy, you might *not* want to re-download new metadata. On the other Satellite, the *content* of the metadata might be the same, even if the timestamp of the metadata would be different. Think about master Satellite and slave ISS Satellite. The ISS slave Satellite might have exactly the same packages as the master, with newer repodata. If you switch from the slave to the master, the content which is available to you might be exactly the same, in spite of the timestamp differences.
This request was not resolved in time for the current release. Red Hat invites you to ask your support representative to propose this request, if still desired, for consideration in the next release of Red Hat Enterprise Linux.