How Veritas File Replicator works

Veritas File Replicator provides replication where the replication unit is a file system. VFR replicates an entire file system, or a specified consistency group to a single target. The replicated files are sent over the network on a periodic basis with a period ranging from 15 minutes to a 180 minutes. The target is updated to match the current status of the source files as of the most recent start of replication using Storage Checkpoints. The changes since the last replication iteration are efficiently identified using the FCL (File Change Log) and Storage Checkpoints.

A replication job is directional. The job definition includes the source and destination, both host addresses, and the mount points. It also includes the interval at which incremental replication is scheduled. At each interval, a file system point in time image is taken and the file/directory changes are sent to the target as a stream of deltas where it is applied to have the target image mirror the point in time image of the source.

A consistency group consists of an include list and an exclude list. An include list specifies a combination of files and directories within a file system that needs to be replicated as a single consistent unit. An exclude list specifies files and directories that should not be replicated. Both include and exclude lists are optional, and if no path is specified in the include list, the entire file system is replicated.

The following directories and files from the root of the file system are not replicated and may not be specified in the exclude list:

At system startup, if enabled, the replication scheduler and the replication daemons are started. Once the replication job is started using the vfradmin command, the scheduler daemon enables the FCL if it is not already enabled, takes a file system Storage Checkpoint, and starts the replication process against the newly created Storage Checkpoint. Old Storage Checkpoints are removed once they are no longer useful.

The replication daemon is useful on machines that act as a replication target. It listens for incoming connections and applies the changes on the target file system. After every iteration, a file system Storage Checkpoint is taken. The old Storage Checkpoint is deleted once a new one is created.

The Storage Checkpoints used by VFR are data full checkpoints and will consume extra storage on both the source and the target file systems. Also note that the FCL is enabled by VFR and it is not advisable that you disable some FCL events or disable the FCL entirely. VFR will detect such events and will use a less efficient method to identify the changed files and directories. VFR will also enable the missing FCL events and turn FCL back on.