The move has allowed it much greater flexibility for web-based projects. It has also cut down on the need for dedicated SAN skills and reduced physical footprint by 20x.
The council – a Windows shop with 3,500 employees – had an HDS AMS SAN which had reached end of life and was experiencing multiple issues with scalability and operational problems, said IT operations manager Adrian Davies.
“Adding more disk had become difficult and costly because of the way different tiers were Raided. It wasn’t just a case of buy another disk and slot it in. We were having to buy an entire tray at £20,000, and it was difficult to anticipate and manage spend,” said Davies.
Davies also reported a string of operational issues, such as an HBA card problem that caused I/O saturation in the HDS array’s dual controllers and led to outages across council systems in December 2015.
“Management software was not intuitive. It was difficult to extract reports for capacity management out of the SAN. We could have upgraded, but licensing costs were prohibitive,” said Davies.
Davies’s team has now nearly completed a switchover to a Nutanix-based hyper-converged architecture at two datacentres – a primary site and a secondary backup site.
Hyper-converged products combine compute and storage in one box with virtualisation capability. They have emerged in recent years as competition to discrete server and storage products, with key suppliers including Nutanix, Scale Computing, Simplivity and VMware’s EVO:Rail.
Read more about hyper-converged infrastructure
Capacity on Nutanix at the primary site consists of 240TB of HDD plus 32TB of flash storage in five two-node clusters of NX8235 boxes, with eight Nutanix clusters totalling around 300TB of HDD and 18TB of flash capacity at the secondary location.
Stockport now runs 450 VMware virtual machines on Nutanix, said Davies, but it will consider switching to Nutanix’s Acropolis hypervisor at its secondary site when all data is migrated to the infrastructure in September 2016.
Currently, the council’s servers are 95% virtualised, although a few physical servers have been retained as part of a staging process and due to issues of licensing that means some workloads must remain on physical servers. However, Davies believes everything they have could eventually be virtualised.
Benefits of move to Nutanix infrastructure
Key benefits to the council of moving to Nutanix hyper-converged infrastructure are ease of management, scalability and hugely reduced times to restore lost data, said Davies.
“It’s a lot easier to manage, and that was a key pre-requisite in the procurement process. We have lost staff with SAN skills and don’t have the budget to replace them, so we wanted something that our generalist IT staff can use to allocate disk and run up servers for people,” he said.
“Operationally, if we want to recover a disk we can do it easily. With the old platform we had to recover from tape to bare metal and that could take days. Now we’re down to a recovery time of minutes or hours.”
The move has also seen physical footprint reduced from 10 racks to half a rack.
Did Davies have any concerns about moving to hyper-converged infrastructure from the tried-and-tested SAN architecture?
He said: “We needed to do something different to achieve very different outcomes and hyper-converged infrastructure seemed an obvious natural step from siloed architecture that delivered what we needed.”