Performance of Snapshot File Systems

Snapshot file systems maximize the performance of the snapshot at the expense of writes to the snapped file system. Reads from a snapshot file system typically perform at nearly the throughput rates of reads from a standard VxFS file system.

The performance of reads from the snapped file system are generally not affected. However, writes to the snapped file system, typically average two to three times as long as without a snapshot. This is because the initial write to a data block requires reading the old data, writing the data to the snapshot, and then writing the new data to the snapped file system. If there are multiple snapshots of the same snapped file system, writes are even slower. Only the initial write to a block experiences this delay, so operations such as writes to the intent log or inode updates proceed at normal speed after the initial write.

Reads from the snapshot file system are impacted if the snapped file system is busy because the snapshot reads are slowed by the disk I/O associated with the snapped file system.

The overall impact of the snapshot is dependent on the read to write ratio of an application and the mixing of the I/O operations. For example, a database application running an online transaction processing (OLTP) workload on a snapped file system was measured at about 15 to 20 percent slower than a file system that was not snapped.