On Fri, Apr 24, 2009 at 07:15:51PM -0500, Mike McGrath wrote:
On Sat, 25 Apr 2009, Matt Domsch wrote:
> I'm seeing the metalink problem and will investigate the cause.
I haven't actually sat down and looked at this yet, is it completely a
metalink problem or are there two different things going on?
metalinks are definitely sometimes wrong. For example, a repomd.xml
file for updates/10/x86_64 that updated at
-rw-r--r-- 1 263 mirrors 2391 Apr 24 19:35 repomd.xml
was not being reported in the metalink at
Date: Fri, 24 Apr 2009 23:15:02 GMT
and I'm not yet sure why. update-master-directory-list is running a
little more often than twice an hour on average, and the mirrorlist
cache is updating every hour as expected. So the data is either not
getting picked up during u-m-d-l runs, or is somehow not getting from
the DB into the mirrorlist cache.
If you see me monkey with u-m-d-l on bapp1, that's what I'm trying to
Linux Technology Strategist, Dell Office of the CTO