Redhat released Redhat Enterprise Linux 4 last week. It was only two weeks ago that I completely slammed Redhat, and for that point in time, I think that was perfectly valid. Now with the release of RHEL4, there are very extremely large changes, and Redhat’s intent going forward is solidified and made much more clear. Overall, things with RHEL4, to the extent I’ve tested in the past week, look pretty damn OK.
I guess there was some miscommunication, or lack of communication from the redhat camp with regard to their roadmap, because I certainly wasn’t expecting what I’m seeing in RHEL4. Redhat EL 4 looks quite a lot like Fedora, whereas RHEL 3 looks a lot like Redhat 9. That’s an enormous change. It also sends a strong message about Redhat’s expectations of the users of their commercial product. The expectation is that larger, slower shops will stay on RHEL 3.x until either RHEL4 is to their liking (and not painful to migrate their applications to), or for five years, at which time RHEL 3 will be end-of-lifed.
RHEL4 changes many things drastically, as opposed to the RHEL3 update releases, which changed fewer things incrimentally. The version of the base default kernel moves to 2.6, SELinux is now included as a default, and versions of most server daemons I’ve checked out are much newer. My pet project, OpenLDAP deployment, is actually doable without compiling all of the components from source. The base version of OpenLDAP moves from 2.0 to 2.2, which in OpenLDAP time represents a sea change.
So far, so good. I’ve played around with some things, and I haven’t found anything that appears disgustingly out of place or very badly broken yet, but I’m sure they’ll pop up, so I’ll follow up when I find them.