In this chapter, we covered a wide variety of operational and maintenance tasks that are part of a Ceph cluster's lifecycle. We explored cluster topology, a topic we'll continue in Chapter 20, Monitoring Ceph, as well as Ceph's logging system. We also delved into ceph.conf for startup configuration, injection for dynamic runtime configuration, and admin sockets for harvesting a wealth of status and configuration detail. Also touched upon was the value of a formal configuration management system such as ceph-ansible.

We also expanded on the topics of adding and replacing failed drives that were touched upon in the previous chapter and offered strategies for the oft-overlooked nuances of working effectively with remote hands.

It may seem odd that we have not devoted an entire chapter to the step by step initial deployment of production Ceph clusters. There are readily available automation tools and resources to help with this process. Duplicating their multivariate minutia here or cutting and pasting large swaths of ceph.com would leave less space to focus on the insights and wisdom the authors strive to share.

For initial cluster bootstrapping we recommend the excellent ceph-ansible project available here:

https://github.com/ceph/ceph-ansible/wiki

A useful blog post detailing one admin's experience using ceph-ansible to bootstrap a cluster can be read here:

http://egonzalez.org/ceph-ansible-baremetal-deployment/

Those wishing a more hands-on, under-the-hood bootstrap experience will enjoy the reference site:

http://docs.ceph.com/docs/master/rados/deployment/

In the next chapter, we will explore the monitoring of your Ceph clusters' well-being.