T
his is the most remarkable deployment that I’ve had the pleasure to post about.
The RackN team has refreshed the original OpenCrowbar Ceph deployment to take advantage of the latest capabilities of the platform. The updated workload (APL2) requires first installing RackN Enterprise or OpenCrowbar.
The update provides five distinct capabilities:
1. Fast and Repeatable
You can go from nothing to a distributed Ceph cluster in an hour. Need to rehearse on VMs? That’s even faster. Want to test and retune your configuration? Make some changes, take a coffee break and retest. Of course, with redeploy that fast, you can iterate until you’ve got it exactly right.
2. Automatically Optimized Disc Configuration
The RackN update optimizes the Ceph installation for disk performance by finding and flagging SSDs. That means that our deploy just works(tm) without you having to reconfigure your OS provisioning scripts or vendor disk layout.
3. Cluster Building and Balancing
This update allows you to place which roles you want on which nodes before you commit to the deployment. You can decide the right monitor to OSD/MON ratio for your needs. If you expand your cluster, the system will automatically rebalance the cluster.
4. Advanced Networking Topology & IPv6
Using the network conduit abstraction, you can separate front and back end networks for the cluster. We also take advantage of native IPv6 support and even use that as the preferred addressing.
5. Both Block and Object Services
That’s a lot of advanced capabilities included out-of-the-box made possible by having a ops orchestration platform that actually understands metal.
Of course, there’s always more to improve. Before we take on further automated tuning,
we want to hear from you and learn what use-cases are most important.
Like this:
Like Loading...
Pingback: From Metal Foundation to FIVE new workloads in five weeks | Rob Hirschfeld
Pingback: RackN fills holes with Drill Release | RackN