It happens to every data center manager at some point. The time to say goodbye. Perhaps surprisingly, it’s not as easy as it sounds. Since data center decommissioning projects involve more complexity than many organizations are accustomed to, so much can go wrong. Surprises, delays, and errors are inevitable. So, to make the project more manageable, we’ve broken down the data center decommissioning operation into bite sized chunks. This is our data center decommissioning checklist.

Stage 1: Planning and Discovery

• Identify and engage all relevant stakeholders and decision makers
• Assign a project manager
• Develop a budget
• Establish an estimated timetable
• Create a detailed scope of work for all processes, safety procedures, stages of removal/demolition
• For live decommissions (removing production equipment from the network), schedule the decommission for off hours with a buffer for unexpected circumstances

Network Tools Discovery

• Choose the optimal network discovery tool:
  - For flat networks, agentless discovery tools are pragmatic and provide comprehensive data
  - For heavily virtualized environments, NetFlow enabled tools with comprehensive network gear recognition are more appropriate than application-centric tools
  - Don’t hesitate to trial multiple discovery tools in the discovery phase
  - Proprietary databases, or 3rd party hosted tools may not give you access to raw data. Ensure that any tool used will give you access to the raw data so that you may gain intelligence that is meaningful to your specific environment

Physical Discovery

• Conduct a physical audit and hold specialists accountable to double verify all inputs
• Compile a hardware map from the physical review and relate that map back to your applications and data

Expert Review

• Review existing CMDBs, discovery tool outputs, and physical audits with qualified experts:
  - Developers
  - Operators
  - Users
  - Management
  - For any expertise gaps rely on external decommissioning/migration experts

Updated Asset Map

• Finalize an updated environment map, including dependencies where relevant for live decommissions, for all of the following elements:
  - Software
  - Servers
  - Storage systems
  - VMs
  - Network equipment
  - Air and cooling equipment
  - Power equipment
  - Cabling

Final Planning

• Create an implementation plan which lists out all actions and responsibilities for project members
• Create a master set of go/no-go criteria with your stakeholders, to ensure that all effected parties have the ability to weigh in on initiation of the decommissioning tasks. Hold a formal go/no-go meeting on the day of the decom and abort if appropriate risks are raised
• Create contact info list for team members to facilitate communication on any issues that may arise, including vendor contacts such as ISPs
• For live decommissioning projects, contact end users about potential downtime to limit frustration in case of errors
• Commission any tools or equipment needed for physical logistics:
  - Fork lifts
  - Drive shredders/ degaussers
  - Pallets and crating for servers
  - Polyethylene foam for packing
  - Hoists
  - Trucks
  - Tip guards
• Conduct background checks on temporary personnel – you don’t want any data or hardware disappearing!
• Use asset tags to denote future location, personnel responsibilities and planned processes

Stage 2: Decommissioning Stage

• Log all servers that are to be decommissioned
• Determine final designation for all hardware assets: Repurpose, Resale, Recycling
• Identify and retain all software licenses associated with to-be-decommissioned systems
• Schedule the cancelation of any vendor maintenance contracts associated with servers or software, which are not being migrated
• For live environments, run tests/simulations on all backups or Disaster recovery to ensure functionality
• Create a comprehensive backup just prior to decommissioning
  - Complete a final verification of the new backup
• Complete data erasure inside the racks with data erasure software
• Disconnect equipment from the network. Remove from ACLs, subnets, and firewalls
• Cut power to all equipment that’s being decommissioned
• Pull all rack hardware
  - Consider tip guards if moving racks
• If drives have not been erased, shred or degauss remaining storage media
• Verify that either a certificate of erasure or destruction has been issued for data security purposes
• Fulfill security policy requirements and be sure to leave a comprehensive audit trail
• Complete rack teardown

Stage 3: Disposition

• Complete hardware palletization and packing
  - Ensure that servers and other hardware slated for reuse are crated with cut-to-fit polyethylene foam; standard freight shipping is insufficient
• Correspond with financial department and accounting department so that servers/ fixed assets are taken off the books and software licenses are accounted for and recovered
• Identify recycling vendor or downstream site for recyclable materials
• Coordinate with other departments to transition any hardware slated for repurposed use

While it’s an extensive list, the complexity of data centers today necessitates a thorough approach. So much depends on data, its real time availability, and its security that the data center has become the central nervous system for organizations. While this is generally understood, the implications of poor decommissioning are not. Given more rigorous data protection regulations, unscrupulous ‘recycling’ and the potential for business downtime, taking a detailed and pragmatic approach to every stage of decommissioning is essential. In the long term it will save time, costs, the environment, and customers.

Jeremy Schaller is a data center industry analyst and heads PR & marketing for Exit Technologies, an R2 certified, global data center IT Asset Disposition Company (ITAD)