CCCEU 2014

The MySQL world is full of tradeoffs and choosing a High Availability (HA) solution is no exception when it comes to your cloud infrastructure. This session aims to look at all of the alternatives in an unbiased nature. Preference is of course only given to opensource solutions. How do you choose between: MySQL replication, MHA, DRBD, Tungsten, Galera Cluster? Do you integrate Pacemaker and Heartbeat like Percona Replication Manager? How do you get automated failover with MySQL? The cloud brings even more fun as geographical redundancy is rolled into this. This and more will be covered in a walkthrough of MySQL HA options and when you will apply them.

Cloudstack has a set of first class objects such as virtual machine, volume, network etc.
In order for the admin to have better control over these objects, he/she needs to have the ability to add meta information to the objects and be able to expose/hide them from the end user. In addition, admin should also be able to deploy these first class resources and have the capability to expose/hide them from the end user. Nitin Mehta is going to discuss as to how is this currently possible in Cloudstack and how this ability allows a service with higher functionality to manipulate objects managed by CS.

CloudStack has certain limitations in choice of virtual disk controller used for volumes attached to instance deployed over VMware. There's great need to bring in granular disk controller support and auto-detection of most efficient disk controller given guest operating system in vSphere. Will walk through current support for various virtual disk controllers in CloudStack over VMware, the limitations and need for enhancements in this area. Will cover detailed proposal addressing limitations by supporting granular SCSI disk controllers and auto detection of recommended disk controller used for volume(s) that suites best for a guest operating system in vSphere. Different types of SCSI disk controllers includes Paravirtual,LsiLogicSAS,LsiLogicParallel,BusLogic.Will talk about various levels of controller configuration like volume,instance,template and cloud. Will present upgrade scenarios.

Scaling CPUs and Memory in a cloud environment isn't the biggest problem we face when deploying cloud infrastructure. Storage however is. How can you deploy mass scale and affordable storage behind your CloudStack environment to keep it scalable? Storage has always been enclosed in a black box which you bought from a vendor. Ceph is fully open source and scales into the PetaBytes. Scale your storage as you need!