Stephen John Smoogen wrote:
If a minute downtime is millions of USD, you surely can afford a few
thousands to set up several machines and failover, soy you /can/ do
the patching and rebooting without visible downtime.
And testing of your failover system, and practicing emergency drills,
and....
What process transformed mid/end life FC3 and FC6 into very stable,
reliable OS's very much like the subsequent RHEL cut? I can't be the
only person who sees the difference at those points from the previous
fedora versions.
Nothing did. EL-4 and EL-5 were cut from pre-releases of FC3 and
FC6... not mid/end-life versions. Fixes in final FC3/6 were pushed
upstream and vice versa but the 2 diverged in a yellow wood at least a
month before FC3 or 6 were released. So by the time FC3 and 6 were
released EL-4 and EL-5 was already set in stone. You can get FC3/6
stuff to work on an EL-4/5 box but only to an extent.
So there were no actual metrics or systematic processes used to
transform the buggy FC2 and FC4/5 releases into the fairly usable
FC3/FC6 versions? That's hard to believe, and especially that it just
happened to converge with the RHEL releases - unless it is the fixes
from the RHEL beta process simply overwhelming the normal fedora updates
and forcing stability. And as for differences between FC6 and RHEL5, I
thought that the Centos group reported that some binaries were not even
rebuilt from their FC6 versions. Did I miss something there?
--
Les Mikesell
lesmikesell@xxxxxxxxx
--
fedora-devel-list mailing list
fedora-devel-list@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/fedora-devel-list
