1.3 is getting pretty big, and I need to add the "Illustrated
Changelogs" info to better describe all the changes, though see here:
- (FEAT) ACLs to extend authz (see Wiki)
- (FEAT) puppet integration with --mgmt-classes and external nodes URL
- (FEAT) added puppet external nodes script, cobbler-ext-nodes
see
https://fedorahosted.org/cobbler/wiki/UsingCobblerWithConfigManagementSystem
- (FEAT) ability to use --enable-menu=0/1 to hide profiles from the PXE
menu, and config setting to change default value for --enable-menu
- (FEAT) added livecd based physical machine cloner script to "contrib"
- (FEAT) enable import for debian ISOs and mirrors (1 distro at a time
for now)
- (FEAT) auto-create rescue profile objects
- (FEAT) included network_config snippet and added --static=0/1 to
system objects
- (FEAT) added /usr/bin/cobbler-setup tool
- (FEAT) cobbler report gains additional options for Wiki formatting,
csv, and showing only certain fields
- (FEAT) changed default kernel options to include ksdevice=bootif (not
ksdevice=eth0) and added ipappend 2 to PXE
- (FEAT) distro edits now no longer require a sync to rebuild the PXE menu
- (BUGF) minor tweak to the blender function to remove a certain class
of typing errors where a string is being blended with a list, should not
have any noticable effect on existing installs
- (BUGF) add missing import of "_" in remote.py
- (FEAT) upgraded webui editing for multiple NICs
- (FEAT) "template_universe" variable created for snake's usage,
variable contains all template variables and is also passed to the template.
- (FEAT) refactored import with better Debian/Ubuntu support
- (FEAT) Func integration snippets and new settings
- (FEAT) settings file and modules.conf now generated by setup.py using
templates
- (FEAT) --template-files makes cobbler more of a full config management
system!
- (FEAT) cobbler reposync now supports --tries=N and --no-fail
- (FEAT) duplicate hostname prevention, on by default
- (FEAT) make import work for Scientific Linux
- (FEAT) distro remove will remove mirrored content when it's safe to do so
- (FEAT) repo remove will remove mirrored repo content
- (FEAT) added snippet for better post-install network configuration
The question is do we want to look at trying to get 1.3 devel finished
up 2 weeks or so (not including testing), and limit further features
given size?
I want the networking config stuff in pretty bad and want to resolve any
defects in Trac that are already there of course. This (winding up the
release) seems like a good idea to me and the release names are just
numbers anyway.
As usual we'd do a test release for a couple of weeks to keep the bugs
down to a reasonable limit. I think we also need to test this a good deal
harder next time, so hardening up the automated tests seems to be in
order, and some extra user help would be great too.
Then we can look at larger issues we had planned for the future in a 1.5
release.
We can hold this a bit if folks are planning things that largely
"complete" some of the other features, for instance, Debian support is
coming along rather nicely. If we were trying for fully automated
deployments from import and linking up profiles with repos and such,
that might warrant waiting a bit longer as I wouldn't want to hold that
too long from seeing a stable release.
This would push advanced web features towards 1.5 though probably
doesn't leave off too many other roadmap features. We've definitely
been busy with user patches for all sorts of unexpected features and I'm
extremely happy to see that.
Thoughts?
--Michael