summaryrefslogtreecommitdiffstats
path: root/TODO
blob: 458f45da66e5a0bbf5dca8adaa7ff207b08f94bf (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
cobbler TODO list
now with 1-5 priority ratings

always:
2  -  better documentation (particularly online articles)
2  -  add repo stuff in the manpages

more new thoughts:
1  -  support repo add taking http:// urls, which will use repotrack/reposync
3  -  possibly look at pungi, though maybe not if above (^) works well.

core:
1 -  turn on locking in production (or think about it) -- important for rsync especially.
1 -  import over http://, ftp:// in addition to existing file and rsync:// and ssh://
      NOTE:  possibly allow import via curl (just w/o updates)?

4 -  make "cobbler sync" only change what needs to change, to be more efficient
     (though is this really a problem?  Copying doesn't take THAT long...)
4 -  make the "scrub orphans" bit a seperate command, warn by default, with an
     option for "--delete".  Seperate from "import".   This could probably
     be a part of "check" to avoid making another command.
5 -  "check" should check if services (esp httpd) are started

user interface:
5 -  "cobbler add" should be distinguished from "edit" so as to not overwrite
5 -  GTK (or Web) UI for Cobbler ?
5 -  PXE menus would be interesting for default behavior.  
     Set local boot to the timeout default, show all profiles.
     Depends on how many entries can be added...

code cleanup:
5 -  remove the cobbler_msg logic since i18n isn't likely

ideas:
5 -  make the multiple-bootloader system more plugin oriented to allow
     for alternative bootloaders (not neccessarily kickstart related)
     to be added.  (Low priority)
5 -  kickstart success tracking
     originally I implemented some of this using a CGI based system, though it really would benefit
     more from a custom webserver (python's HTTP server would be fine) and providing a few more options
     to integrate with the API.  this would allow tracking of more data w/o having to store it in files, etc,
     and a bit more opportunity for dynamic stuff.  It also prevents needing to tweak the permissions
     model used by the serializer, (and some additional SELinux problems).
5 -  a "--next-profile" option to set the profile after a machine has kickstarted, i.e. to boot locally, etc
     (useful with PXE-anytime configs, or perhaps PXE configs that just probe HW or do something weird)