Rollup throws errors like "Epoch chain broken for volume"
Each incremental recovery point of a volume depends on the previous recovery point of that volume. If this chain of dependency--or epoch chain--breaks, data after the breakage can become inaccessible.
Run rollup on each agent in turn until you discover which one is throwing the warnings:
Right-click on the agent and select Properties
Go to the Retention Policy tab
Click Run Now
Divide protection groups:
Go to the agent's Summary page in the Replay Admin Console
Change the interval of each volume that's getting the error, so none of those volumes are on the same protection interval as non-problematic volumes. NOTE: If you skip this step, forcing a resync will force it on all volumes
Force a resync:
Go to the agent's Summary page
Beside the problematic volume, click Actions->Force Resync
This will take and transfer a new base image, but only store the changes since the last good recovery point, effectively quarantining problematic ones
Once the resync completes, you can move all volumes back to their original protection intervals