On 08/21/17 19:24, Roger Heflin wrote:
I would mount
/var/cache/dnf via nfs from my main machine on the others I was
updating and then change /etc/yum.conf (keepcache to 1) and that way
generally the first node would fill the cache, and then the 2nd node
would reuse the data downloaded by the first machine (assuming both
are the same fedora version).  You just have to make sure to do one
and let it finish then after it is done move on to the second.

Roger
+

I have three computers set up to look and work the same, however this one is booted every morning, the others not so often. Yesterday's update was unusually large since it had not been done for a week. There's also a portable configured similarly but rarely used. With time they tend to squire different applications making each a little different so I'm not sure how well a local collection of updates would work. I thought about it at times but usually just drop the idea.

It's simple enough to just do the updates during the "free" time from midnight to five and that's what I usually do. When my usage runs low for the month I sometimes elect to run an update during the day. That's what happened yesterday, it said it was a 374 MB download, I thought well it will reduce to much less with drpm's, and was surprised when I saw none.

It's not a disaster, just a surprise and an annoyance. I figure I can use ~800MB a day and stay within limits. Had I realized dnf was not using the delta rpms I might have waited.

-- 
Bob Goodwin - Zuni, Virginia, USA
http://www.qrz.com/db/W2BOD
box10  FEDORA-26/64bit LINUX XFCE Fastmail POP3