Abstract: Responsibility for protecting a distributed file system is split across different client hosts. Each client host is assigned a subset of the distributed file system to backup. An index host is identified to create a common namespace under which different client file indexes generated by the different client hosts during backups of respective subsets of the file system will be maintained. During backups, different client file indexes are received from the different client hosts. The client file indexes include metadata about files in the file system included in the backups. Each of the different client file indexes are associated to the same common namespace to provide a unified view of all files across all subsets of the file system that have been backed up by the different client hosts.

Abstract: Responsibility for protecting a distributed file system is split across different client hosts. Each client host is assigned a subset of the distributed file system to backup. An index host is identified to create a common namespace under which different client file indexes generated by the different client hosts during backups of respective subsets of the file system will be maintained. During backups, different client file indexes are received from the different client hosts. The client file indexes include metadata about files in the file system included in the backups. Each of the different client file indexes are associated to the same common namespace to provide a unified view of all files across all subsets of the file system that have been backed up by the different client hosts.

Abstract: Check point recovery based on identifying used blocks for block-based backup files is described. At least one data block is identified that is used by a system since a point in time in response to receiving a request to restore the system based on the point in time. At least one data block, corresponding to the identified at least one data block, is recovered from at least one backup file for the system, without reading each data block backed up via the at least one backup file for the system. The system is restored based on the recovered at least one data block.

Abstract: Check point recovery based on identifying changed blocks for block-based backup files is described. At least one data block is identified that changed for a system since a point in time in response to receiving a request to restore the system based on the point in time. At least one data block, corresponding to the identified at least one data block, is recovered from at least one backup file for the system, without reading each data block backed up via the at least one backup file for the system. The system is restored based on the recovered at least one data block.