The Undocumented Risks of Data Migrations
All organizations hate introducing risk into their production IT environments. Yet many organizations of all sizes regularly take undocumented risks whenever they migrate data from one storage target to another. The need for data migrations often occurs any time they introduce a new storage array or storage target into their environment. While the risk level varies from one organization to the next, none are completely immune from them.
IT staff, regardless of their role in the organization, generally dislike migrating data from one storage target to another. The degree to which they dislike these data migrations frequently depends upon the risks associated with completing them.
In a best-case scenario, IT staff find them an annoyance. They have the expertise, software, and processes in place to successfully execute upon them. While they do not “love” them, they do not find them overly burdensome either.
In a worst-case scenario, IT staff loathe them. They may need to use multiple different software tools to migrate data. They lack the skills and expertise to manage these tools and complete the data migration quickly. They must manage a process that involves many people and touches multiple groups.
Love or hate them, data migrations create risk. Data migrations require that IT staff perform tasks that are may be poorly documented or undocumented since they perform them infrequently.
Step 1 – Pre-data Migration Prep
- Create network routing between the servers and new storage array (New zones on FC switch)
- Assign storage capacity to the servers (LUN masking)
- Determine which software tool(s) will be used to migrate data
- Contact appropriate people and groups
- Schedule change control
- Find a date that will work for everyone
- Verify existing servers can see new storage
Step 2 – Data Migration
- Initiate the migration
- Verify all data is successfully moved
- Verify applications run as good or better on new storage
Step 3 – Post-data Migration Clean-up
- Remove unneeded network routes (Remove unneeded zones)
- Remove unneeded storage assignments on storage arrays (Remove unneeded LUN masking)
- Delete data on the old array
Aggravating the situation, everyone expects data migrations to occur without disrupting production applications. These expectations include applications experiencing minimal or no performance impact while the migrations take place.
To top it off, the responsible IT staff may have to work nights and weekends in addition to fulfilling their normal responsibilities. This combination of balancing competing priorities, working extra hours, and managing a complex, unfamiliar task introduces risks that companies may fail to fully grasp.
Data migrations represent what many IT staff view as both an unavoidable task and a necessary evil. However, as more organizations look toward next the generation of storage solutions to store their, look for solutions that transparently and non-disruptively handle this task. By selecting cloud storage, software-defined storage, or a scale-out storage solution as their next storage target, they can ideally make their next data migration their last data migration.