How Mindbody migrated 60,000 SQL Server databases to AWS (XNT317)
Summary of the Video Transcript
Introduction
The session is about how Mind Body migrated 60,000 SQL Server databases to AWS.
The presenters are:
Leor Sadan, a Solutions Architect at AWS
Tim Ford, the Manager of Database Administration at Mind Body
About Mind Body
Mind Body is a business-in-a-box solution for yoga studios, fitness centers, spas, and salons.
They have about 3 million active users globally and are the industry leader.
Reasons for Moving to the Cloud
Save money by converting sunk costs on the data platform to an operational model.
Leverage infrastructure as code and other modern methods to reduce repetitive tasks.
Improve High Availability (HA) and Disaster Recovery (DR).
Take advantage of the breadth and depth of AWS services.
Initial Assessment and Planning
Mind Body used AWS's Optimization and Licensing Assessment (OLA) to:
Right-size their infrastructure for the cloud
Optimize their Microsoft licensing stance
Their on-premises SQL Server architecture consisted of:
60,000 standalone databases on SQL Server 2016 Standard Edition
4 clusters with 6 availability groups across 3 nodes each
Challenges with the Initial Migration Approach
Planned to migrate the databases first, but encountered 20ms latency issues due to the distance between the Oregon and LA regions.
Couldn't pursue a "big bang" migration approach due to the impact on customers.
Leveraging Local Zones
AWS introduced Local Zones, which provide infrastructure closer to the end-users.
Mind Body was able to test and run their workloads in the LA Local Zone, reducing the latency issue.
However, there were some limitations with services available in the Local Zone, such as the lack of GP3 storage initially.
Addressing the Storage Challenge
The original plan to use GP3 storage was not possible in the Local Zone, leading to a significant cost increase.
Mind Body worked with AWS to architect a solution using GP2 storage in a RAID 0 configuration to meet the performance requirements.
AWS subsequently released GP3 support in the Local Zone, allowing Mind Body to convert their storage online with minimal impact.
Achieving High Availability
Without access to shared storage in the Local Zone, Mind Body leveraged a third-party tool, SIOS, to provide block-level replication and HA for their SQL Server clusters.
The Migration Approach
Mind Body used a homegrown database migration tool, "DB Mover," to migrate the 60,000 standalone databases.
For the availability groups, they used log shipping to migrate the data from the LA region to the Oregon region.
For the failover cluster instances, they created a duplicate infrastructure in Oregon and used SIOS to mirror the data.
The Cut-over Process
Mind Body was initially given a 2-hour window for the entire migration, which was later extended to 30 minutes.
They faced some challenges, including a few clusters that tried to cut over early, but were able to recover quickly.
Outcomes and Lessons Learned
Mind Body achieved:
Approximately 13% cost savings by moving to the Oregon region
8-month ROI on the migration
Improved HA and a 30% improvement in backup and recovery
Key lessons learned:
Continuous training and upskilling are crucial
Leveraging AWS partners is invaluable
Detailed planning and preparation are essential for a successful migration
Next Steps: Application Modernization
Mind Body is now exploring options to further modernize their application, including potentially moving from SQL Server to PostgreSQL to reduce licensing costs.
These cookies are used to collect information about how you interact with this website and allow us to remember you. We use this information to improve and customize your browsing experience, as well as for analytics.
If you decline, your information won’t be tracked when you visit this website. A single cookie will be used in your browser to remember your preference.