A recent webinar hosted by StorMagic, featured the challenges and solution approach taken by the City of Milwaukee. The challenges faced by George Knops, network coordinator for the City of Milwaukee's Water Works Department were not unusual:
- The servers in the department and the associated hardware were reaching the end of their useful life.
- Replacing servers and provisioning new servers was a monumental task.
- Storage was not shared across servers, so that one server had 0.5 terabytes of available storage, while another was running out of storage capacity.
- Although it was operating two manned data centers, the department had no application-failover capability within the two data centers and no disaster-recovery capability across the data centers.
The department defined several desired outcomes from the server-replacement initiative:
- Virtualize servers to facilitate rapid provisioning of new servers,
- Remove the one-to-one binding of servers to physical storage,
- Implement a server and application-failover capability within each of the two manned data centers,
- Implement disaster recovery capability across two data centers.
The department evaluated several options and ultimately implemented a two-server VMware ESX cluster in each of their two data centers. VMware ESX enabled rapid provisioning of new virtual servers. VMware ESX, together with VMware HA and VMotion, enabled applications to fail over within a data center and across the two data centers.
However, in order to take advantage of the application fail-over and disaster-recovery capabilities and maintain application access to data, the department needed to implement shared storage. For this, the department chose the StorMagic SvSAN. SvSAN virtualized the internal storage in the ESX servers and provided mirroring of the data between the two servers in each data center. In that way, when applications failed over from one server to another, they maintained access to their data.
In addition to achieving the desired outcomes, the department found that the StorMagic SvSAN was simple to implement and use. The SvSAN is managed from an application plug-in within VMware's vSphere. Using the management utility, new data stores can be provisioned in minutes.
Shortly after implementing the VMware ESX environment with the StorMagic SvSAN, the Department had an opportunity to experience the high-availability and fail-over capabilities first hand. The network coordinator put a VMware server in maintenance mode to apply some automatic updates. After the updates were applied, the ESX server came back online. However, the administrator failed to do a "power on." The administrator then took the second ESX server and put it in maintenance mode and started applying the patches to it, and the first server went down. The coordinator was left with an entire site down. The coordinator was able to go into the individual host machines, and because of replication between the SvSANs, was able to immediately restore application access to data.
A full replay of the webinar can be viewed on the StorMagic website.
Ensuring high-availability and disaster-recovery capabilities for applications should be considered as part of every server-replacement and server-consolidation initiative. Previously affordable for large companies with similarly large IT budgets, solutions are now available and affordable for small and medium sized businesses and companies on tight budgets. Shared storage is a prerequisite for high-availability and rapid disaster-recovery. In order to achieve affordable shared storage in virtual server environments, companies should consider leveraging storage virtual appliances, such as those offered by StorMagic.