Am 02.11.2013 22:29, schrieb Michael Scherer:
Ars technica summarize quite clearly the situation on this problem :
http://arstechnica.com/information-technology/2013/11/its-the-little-thin...
And I do not even speak of the users who reboot during a upgrade,
resulting into unbootable system due to issue like this
(
https://bugzilla.redhat.com/show_bug.cgi?id=1002891 ). Sure, people
shouldn't do it. Yet they do, that's purely a statistical problem. Maybe
you do not see it with your small set of 20 servers, but with ~ 40 RHEL
desktops in my office, I have seen it 4 times. I have spend ~ 2h to fix
each of them. Now, take a bigger fleer of laptop, and count how much
this is costing in time to a company. Time lost by users, time lost by
having someone looking at it instead of focusing on others issues
strange - and instead fix the reboot/shutdown to delay the shutdown in case
of a running rpm/yum/dnf we go the crappy way of install updates offline
to work around statistics?
sorry, but i can't see the improvement here
in that case even windows is better which is technically wrong but with
such behavior and conclusions we sadly make it true