Overcoming Data Restrictions with Cloning Solutions
To kickstart the migration process, Sneed and his team of engineers needed to access the data from each application so they could transition from their existing environments to AWS. The apps that the Contegix team needed to move were Jira, Confluence, Bitbucket and Bamboo. This included the database (Postgres) and all related files. The catch? Due to the customer’s company policy, Contegix engineers couldn’t access the production instance.
“We worked around this obstacle by cloning,” explains Sneed. “All four applications were cloned over to sandbox instances. This way, our team could securely perform changes to the applications without touching the customer’s production.”
But when it came time to actually migrate each application’s data to AWS, massive file sizes threatened to stop the process in its tracks. The Bamboo instance in particular was thought to be the largest Bamboo instance in the entire Atlassian ecosystem, too large to transfer at all.
Compressing Massive Files Sizes to Drive Mobility
The team needed to reduce the application sizes to generate movement, but another challenge presented itself: the agency’s team didn’t have Atlassian admin permissions after the apps were cloned. Sneed edited the database to grant the the customer personal admin access, where they formulated scripts to run against Bitbucket and Bamboo to greatly reduce their size––reasonable enough for them both to move.
Although the cloning and compression processes were time-consuming, the pressure was on to work swiftly. If the apps weren’t successfully migrated by the end of July, the agency would have incurred hundreds of thousands of dollars in relicensing fees. Each application took a few hours to clone, reducing the size of the Bamboo instance took about twelve hours, plus an added three hours to export, then another six hours of import time—all just to move the data.
After a collaborative effort between Sneed and the agency’s team, the Bamboo instance was reduced by 80% and the database became much more manageable to migrate.
“The original estimated synchronization set was five days, or around 44 and a half total hours. We got it down to 14 hours for the production run for Bamboo,” Sneed says.
Ensuring Functionality, Making Deadlines and Saving Dollars
Any successful migration hinges on each app functioning properly once all is said and done. Due to the nature of the application, data stored in Bamboo is constantly changing as teams use it, which triggers changes across a number of areas in the file system. So when users stopped using the app for the migration, it was critical Sneed’s team completed the data transfer before teams logged back in.
“After the migration is complete, and as soon as you turn on the applications, each tool is going to try and connect all the things it thinks it should connect to. We did have to make sure that quite a few pieces of the app were functioning as expected, with new values or new routes of communication, so that when users were granted access again to the applications that all the expected functionality remained, even though various parts of the applications had changed,” explains Sneed. “That's how we ensure that the continuous infinity loop of the development cycle continues on.”
Now, the agency’s newly acquired division is completely hosted on AWS, the most compatible database that most teams utilize for Atlassian products, which helps to improve the longevity of the instance overall. Sneed’s team executed a timely, completely synchronized migration, which allowed users to pick back up right where they left off and avoided additional costs for the agency.
Learn more about how Contegix can support your team during an Atlassian migration.