Christopher Browne wrote:
> - In reflecting on it, I think I'd like for some of the derived classes
> to be derived out of data residing in external sources like LDAP/DBMS.
Yes, a native ability to import, AND export, class information would be
the bees knees...
Maybe it can already be found in the documentation(anyone have a link?),
but I'd be interested to see how else one might structure a "files:"
example like Neil offered for any three particular files. I'm not so
much interested in modularization, as in simplification. I'd like for a
service ad
Neil,
I've not made the leap yet from cf2, but so far your practical use cases
and examples and questions on cf3 have been more useful even than the
reference guide in starting to "get" cf3. Thank you.
Maybe it's just that this is only an example, but I have to wonder about
the value or econo
[EMAIL PROTECTED] wrote:
> Dear Ed,
>
> All I can say is:
>
> THANKS!
>
> I may hack it around a bit but if I do that I will post it back to the
> group.
>
> Best Regards,
>
> Brett
>
>
> -Original Message-
> From: Ed Brown [mailto:[EMAIL P
Back before the 'packages:' action had an install option (~2 years
ago), when it could only define classes, I wrote a simple[-minded?]
perl wrapper that would parse the list of defined classes and hand off
to yum. I'm still using it because it's been stable, avoids
limitations like this argume
Brad Lhotsky wrote:
I want to append a
postgresql* to exclude, because I manage those RPMs manually.
It would be nice if yum allowed for more flexibility on the
commandline, but we deal with the need for various yum configuration
file differences by defining variables and building the file en
fengine
to get the modes right on new directories). Atom notes that it does
NOT check/fix modes recursively though.
-Ed
Ed Brown wrote:
This has come up a few times in the past, it's really a surprising and
disappointing missing capability. I don't think you CAN both create a
file
For slides from a 15 minute talk at last year's LISA conference about
how to get started VERY simply, see:
http://www.usenix.org/events/lisa05/htg/brown.pdf
-Ed
Digant C Kasundra wrote:
Is there a decent step-by-step how-to on how to setup 2 systems?
Just curious. It might make the learnin
Brian E. Seppanen wrote:
I don't follow that because with
actionsequence = ( processes.pre shellcommands processes )
is appropriately executing in the order that I want.
You're right. If you use '!pre' to qualify processes that you want to
run only at the second processes actionsequence item
Here is an excerpt from an exchange last year about this:
---
On Wed, 2005-01-05 at 06:34, Christian Pearce wrote:
> cfagent.conf
> control:
> actionsequence = ( copy shellcommands shellcommands.secondrun )
In my experience, if you qualify any action, you had better
Depends on what daemon you mean. cfexecd can run as a daemon and
provides scheduling, like cron, for cfagent. Or, you can run cfexecd
from cron and not have it become a background process (-F option). No
daemon necessary there.
cfservd is not necessary on the client side either. You lose t
Michael Grubb wrote:
I understand that this mechanism works as documented, meaning that it
iterates of each server in the list. I guess I was hoping for a little
more intelligence, in that it would only go onto other items if the
first one failed.
I wondered if this is what you were rea
Michael Grubb wrote:
Here are the relevant entries in update.conf:
import:
main.cf
#... snip ...
copy:
$(cf_master_inputs)
dest=$(cf_workdir)/inputs
server=$(cf_policy_hosts)
The relevant entry in main.cf:
control:
cf_policy_hosts = ( polserv1.domain:pol
I don't know when list iteration was extended to the server field, but I
can tell you it doesn't work in 2.1.11 and does work in 2.1.18. What
version are you using?
-Ed
Michael Grubb wrote:
In the documentation for list iteration it says this:
__
e
of difficulty understanding or trusting what it's going to do or what it
is doing. I think that improved simple, concise, clear reporting and
querying abilities could really take cfengine 3 to another level of
adoption.
-Ed
Forwarded Message ----
From: Ed Brown <[EMAIL P
On Tue, 2006-03-21 at 10:10 -0500, Brian E. Seppanen wrote:
> Does someone know of an easier way of testing
> for the existence of a user that wouldn't require a call to getent?
> Just wondering if there is a better solution out there.
Don't know if this is better, but it is quieter: you can a
16 matches
Mail list logo
