
Sadly...
The number of customers willing to touch Lustre or GPFS (I don't care if IBM renamed it), is still abysmally low. You have only a few super deep accounts that run this gear that keep Lustre and GPFS viable to be maintained. Storage in recent times has become stupidly easy on the enterprise side of the house(Pure / Nimble), and the performance has gone through the roof. As a consequence they are going into other areas of the data center like HPC, and It's squeezing out lower end HPC environments that still required dedicated staff and gear to maintain. DDN needs some serious front end developers to wrapper their gear, as it's still a science experiement getting even their pre-engineered and packaged gear working. Last Exagrid I did took 3 days with me and DDN onsite to get running. Verses a Pure Flashblade that was out of the box, racked, configured, and presenting NFS for an HPC workload in about an hour.