The package apt-transport-https is required to successfully add some
of the debian repos. This commit moves the package resource
installation into the rhizo_base::apt::debian class instead of the
rhizo_base:📦:debian class to break a dependency cycle that
would otherwise exist if rhizo_base::apt::debian depended on all of
rhizo_base:📦:debian.
The /etc/php/7.0/apache2 directory is created by the
libapache2-mod-php package on installation. Be sure this package is
installed before attempting to set its configuration files!
Install osmo-sip-connector from package repo
Install the sip connector config file.
Disable LCR in runit
Also adds hiera param lcls for new functionality with sip-connector
Puppet treats included classes specially, and does not assume that
they are contained in their outside class for the sake of dependency
ordering. This means that any other resources requiring the base
non-distribution specific classes were actually not capturing the
ordering dependencies on resources defined in the distribution
specific class!
The key referred to in `modules/rhizo_base/manifests/openbsc.pp`. A
default value of 0 (anything but "1") will result in a new config
being generated from a template.
rhizo::grps is referred to in
`modules/rhizo_base/manifests/init.pp`. The default value of "none" is
only a guess but allows puppet provisioning to complete and the
osmocom stack to start.
The ntp_servers key is referred to in
`modules/rhizo_base/manifests/init.pp` and was otherwise undefined. It
is in the common file since it seems similar to other parameters
already in the file, but there is no strict reason for the current
location. The assigned value is also just an initial default guess.
This update addresses the bug
https://tickets.puppetlabs.com/browse/MODULES-6545 resulting in a
spurious concat warning from the postgresql module. This commit makes
the minimum update to address this issue, although much newer versions
of the concat module are available at this time (v6.0.0).
We need a new backup strategy.
In the meantime, dramatically reduce the size by not dumping large tables.
TODO: even increase the frequency of dumping the more volatile data?
TODO: implement some kind of incremental backup.
