Disaster Recovery Planning

Disaster recovery requires considerable planning. For example, a local disaster can cause severe financial loss to the organization. To reduce this, the organization must be able to quickly execute the disaster recovery (DR) plan to bring its systems online. It requires a robust disaster recovery plan and periodic DR drills to ensure that everything will work as planned. Oftentimes, organizations have well-intended disaster recovery plans, but they have never tested these plans for readiness; in a real disaster, the plan does not go smoothly. DR planning is based on the specific organization business requirements. However, here are some general areas that will need to be addressed to put this plan into action.

Use project management software, such as Microsoft Project, where people, resources, hardware, software, and tasks and their completion can be input to provide a systematic approach to managing the tasks, resources, and critical paths. Then develop a checklist of detailed steps for recovery.

Disaster recovery solutions with Windows failover cluster are commonly used to provide hardware redundancy within the data center site and can be configured across data centers by use of a more expensive SAN solution to deploy a geographical dispersed Windows failover cluster. Moreover, log shipping and data mirroring can both be inexpensively deployed as disaster recovery solutions. Another solution is SAN replication to maintain a copy of the data in the remote location. Certain organizations have a standby disaster recovery solution available to take over all operations or, at the very least, the mission-critical operations. A few of these organizations failover to the disaster recovery site periodically to validate that their plan will work in a real disaster. Others may not have a disaster recovery plan but an agreement with another organization that offers disaster recovery.

You'll need hardware at the remote site, and if it is not already available, you need to have a plan to acquire the hardware required to bring the organization online quickly. Document the current hardware that will be needed. For computers, consider the number of CPUs and speed, hyperthreading, dual-core, 64-bit, local disk drive capacity, RAID level, and the amount of physical memory required. Preferably, try to acquire the exact hardware to minimize surprises. For the SAN or disk arrays, consider the disk space requirements, LUNs required, RAID level, and the SAN cache. For the network, acquire a similar network infrastructure to maintain the same performance. Some questions to ask:

• How quickly can these computers be made available? Who will deliver or pick up the hardware?

• Will the computers be preconfigured, or will the DR team need to configure them? Who will provide the expertise, and what is the availability for that staff resource?

• Will the SAN be preconfigured with the LUNs and RAID levels, or will the DR team need to configure? Who will provide the expertise, and what is the availability for that staff resource.

• Who will acquire the network equipment, and who will have the expertise to set it up and configure it?

• Will the DR site have Internet access, to download service packs, hotfixes, and for e-mail?

Make a detailed list of all software required, any hotfixes and service packs. Take an inventory of how each is going to be available to the DR team. Make sure that the software is at the required version level and that licensing keys are available and valid. Determine who is responsible to make available the software and the contact information that staff member. If the software is in a certain location, know who has the physical keys and what access they have. In this scenario, 247 access is required.

You need to have a current list of staff resources to be contacted in the event of a disaster that is current and periodically updated. Know who is responsible to maintain this list and where will it be found during a disaster. You also need to be sure of the chain of command and who is onsite and offsite.

Who will be onsite to execute? Create a detailed plan of the roles required and who is to fill those roles. Additionally, make sure there is a backup resource in case a staff resource is missing. How will the staff get to the DR site? Determine how many are needed in each role and who will be the overall project manager or lead to escalate any issues. Who will have the passwords? What logins are required to make the systems available for business?

The DR site must be accessible 247 and conveniently located. As larger DR deployment can take days to execute, the site should have beds for staff to take naps and easy access to food and transportation. Identify who has the key to access the remote site; if that person is not available, who is their replacement? Can resources remotely access the site if they must work from a remote location, and what is required to have remote access turned on? Are the backups for all the databases available at the DR site, or who is responsible to bring them there? If that staff resource is unavailable, who is their replacement?

To make sure that the DR plan will work during a real disaster and reduce loss, as a best practice, periodically simulate a disaster drill and put the DR planning in action to identify if there are any steps that were not taken into account, how quickly the organization can be expected to be online again, and areas that can be streamlined to speed the process. Most important is to ensure that the plan will execute as expected, smoothly and quickly. To get the most effect from this simulated scenario, everyone should approach it as if it was a real disaster and take exactly all actions as planned.