Posted on Wed 23 December 2015 in hints-and-kinks • Tagged with OpenStack, Juju • 3 min read
Juju is Ubuntu’s supported and preferred means of deployment automation for an OpenStack cloud. While in Juju, a deployment unit (a Juju charm) generally expects to fully own the filesystem it is being deployed on, Juju allows you to co-deploy charms on the same physical machines, by way of …
Posted on Thu 17 December 2015 in hints-and-kinks • Tagged with Ceph, Python • 1 min read
In case you need a quick overview of how many radosgw objects live in your Ceph cluster, here‘s how you do that in one (slightly involved) line of Python.
Posted on Wed 18 November 2015 in hints-and-kinks • Tagged with Ceph • 7 min read
We’ve noticed that there are a few common misconceptions around radosgw performance, and we’re hoping that this post can clear up some of those.
radosgw is of course Ceph’s RESTful object gateway. That means that you can use any client that speaks the Amazon S3 or OpenStack …
Posted on Tue 10 November 2015 in presentations • Tagged with Conference, OpenStack, Open edX • 1 min read
Posted on Wed 28 October 2015 in presentations • Tagged with OpenStack, Conference • 1 min read
Of everything that we can build and deploy in a highly-available fashion in OpenStack, deploying highly available networking has been one of the trickiest, most complex aspects to get right.
I team up with Adam Spiers (SUSE) and Assaf Muller (Red Hat) to discuss high availability in OpenStack Neutron.