POC flocker (not meant to be merged)

[rolele]

Hi,
this is probably not something you want to merge at all but this is just to experiment with flocker.

In this POC, flocker is using loopback backend which is not meant to be used at all.
From what I understand flocker works with providers such as aws or google cloud but you have to upload your entire stack on those services to make it work well. Anyway this is just for experimenting.

Now you can have a custom influxdb config file.
This config file is copied in the influxdb_volume (flocker driver).
This meant that the influxdb container can start on any host, the config file will follow the container.

Note: I removed the vagrant provisionning.
I added a ansible.cfg to avoid the big alias.

Cheers

[jamesdmorgan]

Hey looks interesting. I'll do some more reading up on flocker. Currently investigating EFS as we are 100% amazon at work and it looks like an easy alternative to GlusterFS that we currently use.

[rolele]

Talking about that. I would like to try GlusterFS do you have any ansible script that setup the environment to work with docker swarm mode?
What do you think about those? https://github.com/CiscoCloud/mantl/tree/master/roles/glusterfs

[rolele]

another question. is it possible to setup glusterfs with vagrant? or do you have to use. a cloud provider?

[jamesdmorgan]

Hi,

I don't have any scripts I can share as the ones at work aren't publically available. There do seem to be some good guides available though. You could try http://www.jeffgeerling.com/blog/simple-glusterfs-setup-ansible. That role you shared does look good and I don't see why it couldn't be run via Vagrant. You could define the groups in the vagrant file then install the server clients similar to docker swarm / consul

James