Are chef-repo used for store server data in a one place controlled by CVS?


#1

Are chef-repo used for store server data in a one place ?

The development of cookbooks easier to do in separate repositories?

I’am right ?

Thanks.


Best regards,

CVision Lab System Administrator
Vladmir Skubriev


#2

depends. do you use lot of community cookbooks? do you have berks/librarian
integration? we use a single repo which holds only our cookbook
customizations (often times wrapper around community cookbooks) and
roles/env/databags. all community cookbooks are version frozen and managed
by berks. We dont use version constrained environments, in fact our
deployment process ensures that we have only one version of a cookbook
(community cookbook as well as ours). Deployments are gated via CI ->
Staging environments -> Production environment. It working smooth
currently, but we are really small team, if you have large groups, or if
you want to a consume chef scripts that resides in services code base, then
independent repos might be more useful
best
ranjib

On Mon, Oct 28, 2013 at 9:47 PM, Vladimir Skubriev
skubriev@cvisionlab.comwrote:

Are chef-repo used for store server data in a one place ?

The development of cookbooks easier to do in separate repositories?

I’am right ?

Thanks.


Best regards,

CVision Lab System Administrator
Vladmir Skubriev


#3

29.10.2013 09:16, Ranjib Dey ???:

depends. do you use lot of community cookbooks?
We use community cookbooks of course for example now we use:
apt
build-essential
chef_gem
database
fail2ban
helpers-databags
line-cookbook-master
logrotate
mysql
ntp
nut
openssl
openvpn
postgresql
ruby-helper
rvm
xfs
yum

Some of them is only for dependecies, but doesn’t using in real
environment: helpers-databags, postgresql, ruby-helper, xfs, yum

Now we using this cookbooks unchanged. Also we do not need to use
Library Application Idealogy too.

do you have berks/librarian integration?
So we planned to use berkshelf with a chef-repo, and may be spiceweasel.
But at first I must integrate a chef-repo with a berkshelf, and later
spiceweasel.

I think that everything must start from one place i.e from spiceweasel
and/or git repo. Because a large number of tools and methods sometimes
knocks me up. May be not enougth practice tasks in past.

I also use my cookbooks:
cbacula
cdhcpd
cldap
cftp
cinotify
cmsmtp
cnameserver
cnfs
credmine
csamba

and etc

we use a single repo which holds only our cookbook customizations
(often times wrapper around community cookbooks) and roles/env/databags.
You do a cookbook development(cookbook customizations) in chef repo.

You follow this instruction ?:
https://wiki.opscode.com/display/chef/Working+with+Git+and+Cookbooks

Added specific information and maintained you customized cookbooks. Okey
but how about updates from community cookbooks ? I think this is not
trivial task to merge you own customizations and updates from master.

I strive to do not modify cookbooks from community. And using them as is
or write my own cookbooks with using lwrp’s from community cookbooks.

all community cookbooks are version frozen and managed by berks.

Are you save you own cookbooks in separate directory for example:

chef-repo/cookbooks
chef-repo/cookbooks-community/apt
/openvpn
/mysql
/database
?

We dont use version constrained environments, in fact our deployment
process ensures that we have only one version of a cookbook (community
cookbook as well as ours). Deployments are gated via CI -> Staging
environments -> Production environment.
You use travis-ci?

It working smooth currently, but we are really small team, if you have
large groups, or if you want to a consume chef scripts that resides in
services code base, then independent repos might be more useful
best
ranjib
I read about git sub trees mechanism and want to use them in a chef-repo.
I have been studying this issue for.
And may be ask question again i.e. about sub tress in chef-repo


Best regards,

CVision Lab System Administrator
Vladmir Skubriev