6-1
CHAPTER 6
Checkpoints/Snapshots
This chapter provides an insight on how checkpoints are created, maintained and
deleted.

6.1 Overview

The goal of the checkpoints is to minimize the number of copies when creating a
checkpoint. This document discusses what happens to a checkpoint from the time it
is created to when it is removed.
Checkpoint lifetime can be divided into three main stages: 1. Creation, 2. Active as a
pseudo filesystem, and 3. Deletion. These stages are described in the following
sections.

6.1.1 Volumes

In the operating system, every mounted file system is represented by an in-memory
data structures called fs_online. The fs_online of a volume is similar to a gate - all
accesses to the volume are routed from there. fs_online keeps information about a
volume, including the capacity, file-handle of the root directory and status flags. A
file-handle is the virtual identifier required for accessing any file system object in
system. A file-handle maintains information about the virtual volume, and the
corresponding object in that volume. In the case of checkpoint volumes, file-handles
also contain information about the checkpoint identifier (or cpid) that contains the
actual checkpointed object.