This is due to an issue with slow disk consolidation
Currently there's no resolution and is under investigation by the SRM team.
There are steps in the workaround to help remediate the issue.
Increase the vrReplication.synchronizationTimeout and vrReplication.reverseReplicationTimeout values in the vSphere Replication Advanced Settings from 7200 to 14400 seconds.
The steps for increasing the timeout are described here: https://docs.vmware.com/en/Site-Recovery-Manager/8.4/com.vmware.srm.admin.doc/GUID-C151BC3A-86E7-44D9-B475-DA1D091C619E.html#GUID-C151BC3A-86E7-44D9-B475-DA1D091C619E
Raising the timeout of the vSphere Replication Advanced Settings will not impact the current replications.
It might lead to a slightly slower failover as the synchronizationTimeout and reverseReplicationTimeout are increased.