Backup and protection of Exchange servers

Backups of the Exchange server can be divided into two primary destinations. The first is a backup of the entire system's data. This includes core data for the Windows operating system, Server status, application installation, and Exchange software installation directory, plus a backup of the Exchange database. Backup to the system can be done by NetBackup itself to the local operating system backup work. The main discussion below is a backup of the Exchange database.

As all user data is stored in the Exchange database, how to protect them becomes the focus of a solution. The Exchange Server provides several ways to back up its data. After combining the pros and cons of various approaches, we can simply type the methods of protecting exchange data into two categories: database backup and Mailbox backup. Database backups are mandatory. When a disaster occurs, restoring the entire database is the only way to get all the data back to the Exchange server. A mailbox backup is optional, but it has an absolute advantage when it requires fast mailbox data or public folder data recovery. For these two types of backup patterns, you can divide your Exchange server's data protection tasks into three areas:

Application Protection

This protection includes backup and recovery of application files for Exchange servers, cluster support, and disaster recovery steps to restore the entire application system. At the level of application protection, our focus is on backing up the application files and configuration of the Exchange server. Because Exchange runs on the Windows operating system, the protection of the application is also primarily reflected in the backup of the system. We need to back up all of the operating system files and the exchange's installation directory, as well as the server's system state. In implementation, the data that holds all Exchange user information is first protected by backing up the Active Directory server. This ensures that the most important user data in the Windows Active Directory is saved. Then, because of the use of MSCS in the system, we also need to build the information backup to the cluster to ensure that the recovery can still be properly added to the existing cluster system.

Database protection

Database protection includes the way that you use to back up your Exchange server data, such as backing up and recovering the disk volumes where Exchange data resides.

The Exchange server primarily has two databases that hold user information? D? D Directory and Information Store (information Store). The information store is the destination of user data storage, and it is divided into public and private databases. The data for all public files is stored in the public database. All user mailboxes are stored in the private database. To provide better support for scalability, clustering, and backup, both Exchange2000 and 2003 allow information storage to be split into different database storage groups to serve specific users. Each storage group can be protected separately, and the transaction log is shared among different groups, providing a more flexible data protection scenario. A directory is a database of users (recipients) that are stored internally in exchange. Starting with Exchange 2000, this database is merged into the Active Directory of Windows. Although data changes in the Active Directory are not as frequent as information stores, in order to ensure consistency between users and their data, we also need to include directory backups in a backup strategy. Exchange uses a shared transaction log for each database within a storage group, which allows NetBackup to enable more granular backup of the exchange's databases through the exchange agent? D? D differential incremental backup or cumulative incremental backup.

So when we make an online backup of the Exchange database. To ensure that the service is running correctly, Exchange provides a backup interface for this task. This interface can use several backup methods:

VSS Writer Backup

The new Exchange 2003 writer support can be used to take snapshots of a single storage group to achieve full backups. So in the case of disaster recovery, this feature can greatly speed up the recovery.

Full backup

Full backups will have the selected database and transaction log backed up, and then delete the log, freeing up disk space. Full backup is the foundation of all the more flexible ways to backup.

Differential incremental backup

A differential incremental backup stores the logs associated with the database, and then deletes the source log on the disk. The benefit is that only data that has changed since the last backup is backed up, meaning that incremental backups have the lowest impact on the Exchange server. Of course, if you need to use data from differential incremental backups to restore your Exchange server, the process is cumbersome, and each differential incremental backup data needs to be restored sequentially after full recovery to get the final results. Therefore, using differential incremental backups can save time while backing up, but will require more time to recover.

Cumulative Incremental backup

Cumulative incremental backups also back up the log files associated with the database, but will not be deleted after the backup is complete. Thus, a cumulative incremental backup backs up data that has changed since the last full or differential incremental backup occurred. Then, at the time of recovery, we just need to do a full recovery plus a cumulative incremental backup recovery. Of course, as time increases, the amount of data to be processed by a cumulative incremental backup becomes more and more. If you keep using cumulative incremental backups, the last amount of data can be equal to a full backup.

Copy

Will select the database and its log backups, and do not make any modifications to the source data. It can be used in conjunction with the several backup methods mentioned above, without affecting the operation of other backup modes.

Mailbox protection

The technology for mailbox protection enables administrators to make very granular backups of exchange data. You can even make a backup of a specific message in a user's mailbox. Not only does the recovery process become faster, but it also makes it easier for administrators to make policies. Protecting the Exchange server's data at the mailbox level is a way that more and more large businesses will inevitably choose. It allows administrators to fully and flexibly consider different users of the mailbox data protection requirements, so as to develop different strategies to achieve important users of important letters protection. Although mailbox backups can provide very fast backup and recovery in some cases, they are much more expensive than database backups. The reason is the following three points:

1. A mailbox backup must be implemented as a supplemental form of database backup. Because the backup to the mailbox does not include internal data and internal storage information for Exchange.

2, the overall speed of the mailbox backup is slower than the database backup. Although Exchange provides a high performance interface for backup software for its database backup, it can be used only for database backups and not for simultaneous backup of mailboxes. A backup program can only use clients such as Outlook to access mailboxes and messages through Exchange MAPI. The overall speed that it can achieve is very different from database backup. Using cumulative incremental or differential incremental backups of mailboxes can significantly reduce the time required for backups.

3. A mailbox backup actually duplicates a portion of the work done during a database backup, and because a mailbox backup requires a large number of entries, the catalog information it generates grows very fast.