Sandro Bonazzola kirjoitti 11.5.2018 klo 9.43:
Hi,
this morning yum update complains with:
One of the configured repositories failed (Unknown),
and yum doesn't have enough cached data to continue. At this point the only
safe thing yum can do is fail. There are a few ways to work "fix" this:
1. Contact the upstream for the repository and get them to fix the problem.
2. Reconfigure the baseurl/etc. for the repository, to point to a working
upstream. This is most often useful if you are using a newer
distribution release than is supported by the repository (and the
packages for the previous distribution release still work).
3. Run the command with the repository temporarily disabled
yum --disablerepo=<repoid> ...
4. Disable the repository permanently, so yum won't use it by default. Yum
will then just ignore the repository until you permanently enable it
again or use --enablerepo for temporary usage:
yum-config-manager --disable <repoid>
or
subscription-manager repos --disable=<repoid>
5. Configure the failing repository to be skipped, if it is unavailable.
Note that yum will try to contact the repo. when it runs most commands,
so will have to try and fail each time (and thus. yum will be be much
slower). If it is a very temporary problem though, this is often a nice
compromise:
yum-config-manager --save --setopt=<repoid>.skip_if_unavailable=true
file is encrypted or is not a database
Can you please regenerate metadata?
As far as I can see, this has now been fixed.
http://dl.fedoraproject.org/pub/epel/7/x86_64/repodata/ has a working
primary.sqlite.bz2 and it is currently being synced to mirrors. The
usual "yum clean all" or adjusting baseurl tricks may help in the meantime.