Adding additional storage for your repository data
A Data Center license is required to use this feature. Learn more
As your repository data grows it may become too large to hold in a single storage location. When this happens, you'll need to add additional storage space. There are two ways to do this:
- Increase the storage space allocated to your shared home directory
- Add additional data stores
Increasing the storage space allocated to your shared home directory is the most straightforward option, and we recommend it for most customers. If this is not an option for you (for example, because of a company policy limiting the size of disks or partitions), you should add a data store.
Before adding a data store, read through the considerations below.
On this page:
Data store considerations
The ability to add data stores is aimed at organizations with limitations on disk partition size. It's not intended as a way to increase disk performance. In addition:
- Data stores cannot be moved or removed, once added.
- Any nodes that don't have access to the data store won't join the cluster until it's mounted and the node is restarted.
- Repository data cannot be moved between data stores. Once a repository has been created in the shared home directory or on a data store, all of its data will stay there, including the data in its forks, any pull request attachments, and Git LFS data.
Adding a data store using Storage Management
You don't need to stop Bitbucket to add a data store.
To add a new data store:
- Create a new shared filesystem.
- Mount the shared filesystem on all nodes making sure that:
- the filesystem is writeable by the user running the Bitbucket process.
- the data store is not a subdirectory or parent directory of the shared home or another data store.
- the data store is mounted on all nodes in the same location using NFS.
- Go to
>> Enter the data store directory path, and select Check path.
- Select Add data store.
- If you get an error use the IP address included below the message to find the affected node and fix the problem.
Shared home use after adding a data store
After one or more data stores have been added to Bitbucket, the system will not use the shared home directory to store new repositories, except in a few situations.
For example, new forks of repositories stored in the shared home directory will still be created in this directory. This is because Bitbucket stores forks in the same location as their origin repository to save disk space. New attachments and new LFS objects for repository hierarchies stored on the shared home will also continue to be added to the shared home. Lastly, the shared home will continue to hold data that is not related to a single repository, such as user avatars.