As remote repositories are unsure about it quality, rely on shacache and
use it to download and install slapos.node.

This improves quality of playbook and somehow escapes chicken and egg
problem of shacache access and required tools.

This release fixes problem with filling monitor log files with
"AH03490: scoreboard is full, not at MaxRequestWorkers.Increase
ServerLimit."

/reviewed-on nexedi/slapos.package!37

/reviewed-on nexedi/slapos.package!36

Use vifib-kernel to upgrade debian kernel and distro to a recent version

Combining when and register made it impossible to have "portable" way to
find default (public) IPv4 with shell script. This is expected behaviour
of ansible, see https://github.com/ansible/ansible/issues/12779

As Ansbile provides information about a host, use it
(ansible_default_ipv4.address) to fetch this kind of address.

As the process of installing from ERP5 can be controlled with a script,
create such script instead of many steps.

This script is safe to be run many times.

By setting default slapos_ansible_offline to false and using it as
conditional for repository role it is possible to drive, from example
from host variable, if playbook shall restrain itself from running
oneline required roles.

/reviewed-on nexedi/slapos.package!33

This is ERP5 Standalone installation (as erp5-standalone script
provides) offline installation.

/reviewed-on !34

As package provided does not ensure creation of /opt/slapgrid and
/srv/slapgrid add them if not present.

slapos.node package conflicts with firewalld, thus the latter one has to
be absent.

Instead of referring to minor version of Centos, use the currently last
released one version 7.

/reviewed-on nexedi/slapos.package!30

There is no need to use absolute path to executable, as PATH environment
variable provide one and it is easy to be overriden by the user.

/reviewed-on nexedi/slapos.package!31

headers['Location'] might not exists, so get it with fallback to
nothing.

By using dict comparision with assertEqual the result will be presented
as diff between expected and found values. This allows test reader to
understand the reason of failing test, instead of trying to guess what
is really wrong with the deployed site.

"wendelin" key is used in many places in the playbook and slapos so it
is required to find infromation with this key.

Remove temporary directory after exection.

/reviewed-on !29

Instead of chosing arbitrary directory and removing, often not
accessible directories, use standard mktemp for temporary location.

/reviewed-on nexedi/slapos.package!28

Since "[playbook] Make Wendelin like ERP5" Wendelin became compatible
with ERP5 SR, so compilation detection script shall support it.

@rafael - that's last one of my pre-prepared work for playbooks. Here in order to hide "internal" variables of playbook (like `backend_url_path`) and reuse them between different playbooks.

/reviewed-on nexedi/slapos.package!27

Wrong tags were used.

There is no need to differentiate Wendelin from ERP5, as both can use
same cluster configuration, and only BT5 list is different.

The ERP5 software type changed to default, as create-erp5-site has been
implemented there.

Upgrading to tag 1.0.56 requires to update the software_type of ERP5 to default, as it happened with recent changes in slapos.git repository for ERP5.

Cc: @rafael - please feel free to push it to proper audience.

/reviewed-on nexedi/slapos.package!25