|
Grex > Oldcoop > #361: OpenBSD: Discussion of appropriateness for grex and technical merits. | |
|
| Author |
Message |
| 6 new of 80 responses total. |
dtk
|
|
response 75 of 80:
| 
Jan 7 17:11 UTC 2013 |
I thought FreeBSD (or anything that does not have the look-and-feel of
pre-SystemV SunOS/OpenBSD) was anathema here. Huzzah! Sounds like a
porting nightmare for all of the weird custom codes that are run here.
- DTK
|
cross
|
|
response 76 of 80:
|
Jan 7 18:21 UTC 2013 |
Yeah. We're moving to a model where it's kind of understand that if you want
to force policy, you have to be involved. Porting will be a pain, but at
least we'll correct the mistake of using OpenBSD in the first place
|
dtk
|
|
response 77 of 80:
|
Jan 8 13:22 UTC 2013 |
So that future ports and upgrades are less painful, would it make sense
to put at least the customization into CFEngine or the like? This would
also make the system more self-documenting. I know it makes more work
up-front, but after the first upgrade, it has paid dividends. -DTK
|
cross
|
|
response 78 of 80:
|
Jan 8 14:25 UTC 2013 |
I think if we were running across multiple machines, something like cfengine
or puppet would make a lot of sense. But given that we're just on one
computer, and likely will be forever, I'm not sure it's worth the hassle.
We have a set of home-grown scripts for doing things like that (basically,
a big Makefile) and all of the customized files either live in a dedicated
directory hierarchy, or are in a Subversion repository.
|
dtk
|
|
response 79 of 80:
|
Jan 10 19:52 UTC 2013 |
That makes sense (although CFEngine runs easily as single, stand-alone
node).
|
cross
|
|
response 80 of 80:
|
Jan 19 14:54 UTC 2013 |
Yeah. I need to give cfengine another look.
|