Introduction: This article introduces you how to customize a plan during data migration, including disaster recovery and project management skills.
At certain moments, due to business transformation and other reasons, people will experience multiple data migration processes. The scenario is to move the current database data from one storage system or computer to another.
Data migration is a very complex task. According to statistics, 70-90% of data migration will not reach the goal.
Migration is a difficult thing, and according to statistics, 70% to 90% of people fail to meet expectations. You must first consider all the pros and cons. In addition, you also need to be prepared for data migration that may cause downtime, inconsistencies, etc. —— Dominik Kolasa
This article provides you with some valuable migration best practices to help more students complete this process correctly.
Ten rules of data migration:
(1) Data backup
Before starting to migrate data from one system to another, the administrator is required to ensure the safety of the current data and back up the current data to prevent any potential risks from causing data loss.
It is a small probability to delete the database, and the disk loss overflows, the data is not completely readable, etc., and the integrity should be checked before the backup, and the data can be restored to the original state when there is a problem.
(2) Verify the complexity and quality of the data
The next step in data migration is to verify the complexity of the data and use it to decide which method to use is best. Including checking and verifying different forms of data, storage location and format, data format after migration, etc.
Check the cleanliness of the current data, whether it needs to be updated. Data quality should be evaluated and tested, such as data transmission under the firewall state, distinguishing between good and bad data, and eliminating duplicate data.
(3) Unified data standards
When we know the complexity of the data, we need to develop a series of standards. Why develop standards? In order to let people find problems within the knowable range, make sure to avoid exceptions during the project implementation phase.
Data is constantly changing and variable. The development of standards helps people to integrate data and ensure better use of data in the future.
(4) Determine current and subsequent business rules
It is very important to formulate the current data migration process and future business rules, including verification of multiple business rules, to ensure consistent data transmission, and to establish data migration rules.
After completing a set of data migration rules, the rationality of these rules and subsequent optimizations, such as the evaluation of data complexity, should be immediately evaluated when implementing data migration.
(5) Create a data migration strategy
接下就是进行迁移策略的的确立。有两种方法，一个是"大爆炸"式，另一个是”涓流“式迁移。
选择”大爆炸“是将整个迁移和传输在指定时间内完成，例如24小时内。当数据移到新数据库系统，迁移的实时过程即告关闭。此种方式最快，但风险也高。
”涓流“式数据迁移会将过程分成几个阶段。新旧两个系统会同时并行，中间不存在停机时间。虽然这个方法有些复杂，由于系统在迁移过程中不中断，可靠性更高。
（6）交换迁移心得
数据迁移的过程中，会有多个团队共同参与，彼此交流迁移流程是数据迁移的重要实践过程。彼此了解预期，亦能更好的分配任务与责任，列出所有任务和可交付成果，为项目分配角色，验证自己是否有正确的资源来完成任务。
你要着重考虑的事项：
1）谁对数据迁移过程拥有最终决定权
2）谁有权决定它是否成功完成
3）谁负责迁移后的数据验证
如果无法实现明确的任务和职责，就会导致组织混乱，延迟数据迁移之过程，甚至让迁移计划失败。
（7）正确的工具
在迁移中可以手动或使用脚本迁移吗？能，但不是一个好主意。在数据迁移过程中，使用正确的工具可以对数据做基本分析，数据发殃，数据质量验证以及完整测试，这让迁移过程更快捷，更高效。
根据组织和业务用例需求，选择最怡当的工具是数据迁移计划和过程的关键组成部分之一。
（8）风险管理策略
此外，在数据迁移过程中，要充分考虑风险管理，即在迁移过程中可能会发生的问题。列出这些问题，交提出解决它们的方法，特别是如何防止它们发生，包括旧数据损失，安全，用户测试，应用程序依赖项等，从而让数据迁移过程更顺利。
（9）时刻以敏捷心态对待
在数据迁移过程中，采取敏捷方法保持数据的高质量。其中包括高频的测试，发现并消除错误，在整个过程中保持透明。敏捷方法将明确任务划分，职责权属，从而使迁移成本和进度变得更加可预测。
（10）谨记测试内容
在数据迁移的每个阶段：规划，设计，实施和维护来测试数据迁移。只有如何，我们才能及时获取到所需的结果。
Data migration is a complicated process, but it is something that both companies and individuals must go through. If it is an unplanned data migration, it is a good habit to back up the data in time.
Whether it's migrating from one system to another, or migrating to the cloud or data merging, formulate risk management strategies, grasp potential problems, and propose solutions to quickly complete data migration.