Making a backup copy of data before rebuilding data on a node

A computer-implemented method comprising:

• determining that a secondary node is available after a period of unavailability, wherein the secondary node functions at least to replicate data from a primary node with which the secondary node is communicatively coupled, wherein replication to the second node occurs substantially synchronously during periods of normal operation for the second node;
• storing a copy of previously-replicated data on the secondary node when the secondary node is available after the period of unavailability before rebuilding of the secondary node to recover from the period of unavailability;
• causing the copy of the previously-replicated data to be used as a backup copy to the primary node during the rebuilding;
• rebuilding the data on the secondary node after the copy has been made, wherein the rebuilding comprises asynchronously writing data to the data on the secondary node;
• and
• causing the secondary node to synchronously replicate data from the primary node after the rebuilding.

A non-transitory computer-readable medium having stored thereon instructions that, when executed by one or more processors, are configurable to cause the one or more processors to:

• determine that a secondary node is available after a period of unavailability, wherein the secondary node functions at least to replicate data from a primary node with which the secondary node is communicatively coupled, wherein replication to the second node occurs substantially synchronously during periods of normal operation for the second node;
• store a copy of previously-replicated data from the secondary node when the secondary node is available after the period of unavailability before rebuilding of the secondary node to recover from the period of unavailability;
• cause the copy of the previously-replicated data to be used as a backup copy to the primary node during the rebuilding;
• rebuild the data on the secondary node after the copy has been made, wherein the rebuilding comprises asynchronously writing data to the data on the secondary node;
• and
• cause the secondary node to synchronously replicate data from the primary node after the rebuilding.