The B-Virtual Distributed Storage System saves a check sum with every data block saved on disk. The storage maintenance agents perform frequent background integrity checks on the data blocks. If a data block would have been corrupted due to a unnoticed disk write error, disk bit rot or unlawful tampering with the data, the maintenance agent will detect that error. It will delete the corrupted check block and regenerate a new good check block based on the redundant data on other disks. As such data integrity issues are identified and resolved in the background, without impact tot he end-user data availability, and without requiring operator intervention.
If a corrupted check block would be read while retrieving data from the Distributed Storage System, then that check block will be ignored in the decoder. As long as sufficient check blocks can be read from the DSS storage nodes, the data will be reconstructed without impact to the data availability. In such event, maintenance tasks are automatically added to the maintenance queue, such that a maintenance agent can fix the corrupted check blocks in the background.
Concluding, DSS frequently checks the integrity of stored data. Silent data corruption is automatically detected and repaired, without any impact on data availability to end users.