Operations
Database Settings
Right-size the memory the SHA metadata database uses for backup metadata, and apply the change with a safe, gated restart.
What The Database Memory Setting Does
The SHA keeps the index behind your EBA repositories - backup metadata such as recovery-point and block-tracking records - in an internal metadata database on the appliance. Settings, Database lets you right-size the memory that database uses, so metadata lookups stay fast as your backup estate grows.
A health banner summarizes the result: healthy when the current setting comfortably covers the metadata, tuning recommended when memory is below the recommended size or the metadata is approaching the current setting, and pending when a change has been saved but not yet applied by a database restart.
When To Change It
Leave the setting alone while the banner reports healthy. Plan a change when the banner recommends tuning - typically because backup metadata has grown along with your estate - or when you have deliberately resized the appliance host and want the database to use more, or less, of its memory.
The same tuning assessment appears in the appliance's comprehensive health output, which points you back to this page when action is needed.
Choosing A Memory Profile
The page offers five choices.
Match the profile to your host's RAM
The page does not stop you from choosing a profile larger than the appliance host's physical memory. Follow each profile's recommended host RAM - for example, Extra Large (32 GB) expects a host with 64 GB or more - or the database may fail to start after the restart.
Applying A Change Safely
Applying a change is a deliberate two-step process. Save DB Setting stages your chosen profile: nothing changes live, the change stays pending until the database restarts, and the pending state survives page reloads. Restart DB Safely then applies it, and is available only while a change is pending and no protected work is running.
The restart is guarded by an active-work gate: it stays blocked while backup jobs, in-flight backup pipeline stages, secondary copy operations, or validation jobs are running, and the page lists exactly what is blocking. If the gate stays blocked longer than expected, review current activity on the Settings, Process Gateway page. Even with the gate in place, plan the restart as a maintenance action - the appliance's metadata services are briefly unavailable while the database restarts.
After the restart, the appliance verifies that the new memory size actually took effect. On success the pending state clears and the page confirms the restart succeeded. If the running memory is still the old value, the change is kept pending and the page reports that it was not applied - always check the banner and the Current DB Memory figure after a restart rather than assuming the change landed. On some installation types the appliance can only record the setting as pending rather than staging it automatically, which makes this post-restart check especially important.
Viewing this page requires read access to settings. Saving a profile and restarting the database require settings write permission - users with read-only access can see the controls but receive an insufficient-permissions error when using them.
Related Docs