First Deployment

After private/personal configuration is complete, the general idea is as follows:

$ fab dbserver launch
$ fab loadbalancer launch
$ fab appnode launch

Warning

Currently untested with all possible combinations and quantities of server types.

Scaling up and down

First start with a combined dbappbalancer, then when necessary split into an appbalancer and dbserver, followed by loadbalancer, appnode, dbserver.

Once application growth warrants more power, increase/decrease the number of appnodes, add a cachenode, lognode, or tasknode.

Growth beyond this can only suggest that the application is bringing in revenue. The good news is that the deploymachine has already prepared the application for further expansion by neatly separating repsonsibility between services.

Warning

• Currently implemented are loadbalancer, appnode, dbserver and appbalancer.
• Not yet implemented are dbappbalacner, cachenode, lognode or tasknode.
• The splitting process is not yet implemented, stay tuned.

Ongoing Deployments

First use the bootmachine to launch and provision your new machine:

$ fab openstack.boot appnode2
$ fab root provision

Next to add a new app node to the mix follow this process:

$ fab appnode launch
$ fab kokki:loadbalancer

Deploymachine provides a Fabric API for maintaining your application. Commands for git, pip, supervisor, nginx, django and more are included or type fab -l. Take a peek inside the contrib folder for all the available options, or help extend the existing library. Domain specific commands can be imported from your personal in fab_local.py. An example fab_local.example.py is included.