Jeff Spaleta wrote:
On 12/8/05, seth vidal skvidal@phy.duke.edu wrote:
oh - but a word of warning - I've had a couple of different users tell me that b/c of bandwidth constraints they have one machine that downloads the packages into its cache then they nfs export their cache dir out to their other linux boxes.
Yep I do this sort of thing on one network. a common share for the cache so all the fc4 machines that i admin cache packages to the same filesystem instead of having to regrab the same updates from the outside network. Some machines have updates-testing enabled.. some of them don't... but they all use the same cache tree.
-jef
Are your repos disabled on these server systems? i don't suspect you're running yum clean all on them either way unless you really want to re-get it all?
Also, IMO, this should be done as a mirror function (perhaps an idea for yum-utils) and stored elsewhere. as long as it's called a cache, it should be considered like that. Not something you want to count on is still there anyway.
I understand the bandwidth limitation issue, but those repos are likely to be enabled anyway so a yum clean would do just as much damage on those systems.
implement a noclean or protech option and we can all be happy. This would let yum know our intentions for disabling the repos.
/Thomas