Backup and Recovery

How long could you do without it? A day, a week, a month? Forever? "It" refers to either a service--such as messaging, file, print, or remote access-- or data, like a Word document or an email. Your answer to the time question will, along with budgetary constraints, influence the level of service your backup and recovery system will provide.

Email has become mission critical to most businesses

Data loss can occur due to the following:

• Hardware or system malfunctions (44 percent of all data loss)
• Human error (32 percent)
• Software corruption (14 percent)
• Computer viruses (7 percent)
• Natural disasters (3 percent)

The backup plan should provide the maximum possible data protection in a cost effective manner with minimal or no interference to regular business operation. The backup and recovery services implemented should:

• Provide the ability to restore lost data and bring failed system back to normal operation within reasonable time frame, or as defined in the SLA.
• Be scalable.
• Provide ability to operate within short backup and restore windows.
• Meet data retention requirements.
• Be easy to manage and operate.

The data protection methods used in this example include:

• NT Backup files.
• Disk to disk to tape (D2D2T) backup.

Data recovery methods presented here include:

• Near-line bkf file retrieval
• Archived tape retrieval
• Recovery Storage Groups
• OnTrack PowerControls

Backup and Recovery System Design

Running redundant network services such as DNS, DHCP, and Active Directory enables replication of service data from one system to another and provides continuous services in case of failure. Data corruption is another matter, since the corrupted data will be replicated.

In case of a natural/man-made disaster where the entire IT infrastructure is damaged, server redundancy may not help. In such situations the organization depends on backup solutions that include off-site duplication either on another server or tape.

Old Backup Configuration

Briefly, we backed up three servers (two file and one Exchange) to NTBackup bkf files and copied them to the backup server. Then the bkf files were spooled to tape along with SQL agent backups. The Monday through Thursday tapes were kept on-site, while the tape run on Friday was kept off-site. Each day the bkf files were overwritten. Graphic below:

 

Drawbacks to this scenario:

• After a week the daily tapes are overwritten.
• If restoring a file older than 24 hours, must go to tape.
• Increased reliance on tape hardware for restores.

New Backup Configuration

Utilizing the space and bandwidth available, I configured our backups to daily copy all backup files to two separate servers. Weekly those backup files (bkf files) are written to tape. That tape is take off site for archiving. See graphic below:

Benefits to this configuration:

• Complete backup window is increased from one day to one week.
• Backup granularity is increased dramatically.
• Reliance on tape hardware is reduced.

Utilizing scripts, batch files and robocopy, I was able to leverage our disk space and bandwidth into a home grown D2D2T solution!