Issue link: https://epubs.iltanet.org/i/4636
ILTA White Paper Infrastructure Technologies 8 agnostic, which allows for different hardware to be used without affecting the VMs. Organizations requiring less storage at the primary site or shared storage at the DR site should consider an IBM DS3200, which is shared storage that has up to four shelves of combined SAS or SATA drives connecting to servers with a maximum of six unique or three redundant 3GB/s SAS connections. This setup provides the performance benefits of a SAN without using iSCSI or Fibre Channel technology. ORIGINAL DR IMPLEMENTATION Since Davis LLP was an early adopter of a virtualized infrastructure, many of the virtualization- friendly backup and data replication technologies available today were unavailable at the time. The primary form of data synchronization was limited to pricey SAN-based replication of storage volumes in logical unit numbers (LUNs), which protected the data but not the servers. Listed below are the firm's challenges that limited our DR infrastructure at the time: • Immature virtualization DR products • Limited funds due to the capital cost of the consolidation and virtualization projects • Existing tape backup infrastructure not end of life • Data center consolidation and replacement of our core e-mail and document management system (DMS) infrastructure meant no baseline for storage, CPU, network and WAN utilization was available to develop a clear DR vision Working within these limitations, the firm developed a best of breed DR plan, which utilized tools included with each product, existing backup infrastructure and software from Vizioncore. This included native SQL database mirroring, Microsoft Exchange Server database dumps, tape backup, disk-to-disk backup, rsync, Microsoft's Distributed File System Replication (DFSR) and Vizioncore's vReplicator and vRanger (formerly esxReplicator and esxRanger, respectively). • SQL We used both the EVault SQL agent for disk-to-disk backup and periodic database dumps. SQL database mirroring replicated data to a standby server at the DR site. Failover was a challenge because the mirror would have to be broken and clients then redirected to the SQL instance at the DR site. • Exchange We backed up Exchange using both the EVault agent and NTBackup. Once the firm upgraded to Exchange 2007, continuous cluster replication allowed for clustering on separate datastores, which allowed for automated replication of Exchange mailboxes through log shipping to the DR site. • Disk-to-Tape and Disk Backup The firm continued to use tape-based technology, gradually migrating to disk-based as both the tapes and tape drives were retired. With both tape- and disk-based backups, the firm utilized existing licenses with agents on each guest (server running within an ESX host) operating system that required backup. The backup issues were: