Systems and computer program products for highly available computing systems. Techniques for verifying a replicated snapshot integrity includes steps for storing a snapshot at a first computing system where the snapshot has a corresponding first data integrity value (e.g., a checksum). Another storing operation stores a replica snapshot as two or more portions at respective two or more computing nodes of a second computing system. A second data integrity value that corresponds to the snapshot that was stored at the two or more computing nodes is calculated. Corruption is detected by comparing the first data integrity value to the second data integrity value. Corruption of a to-be-assembled snapshot can be detected before incurring the cost of assembly of the two or more portions of a replica snapshot. Alternative snapshots that can be assembled from alternative portions are considered for assembly. Only uncorrupted snapshots are restored to a designated computing system.