Best Practices for Backup Methods and Data Retention
In the old days when people backed up to tape drives a certain strategy had to be used. The most popular was Grandfather/Father/Son tape rotation. This method required 19 to 23 tapes and ensured that there were always backups of the seven most current days, the four most current weeks, the eleven most current months, and the most current year. This strategy is not required with Online Backup.
The Online Backup Guide for Service Providers is a complete 196-page guide on starting and operating an Online Backup Service – the latest revision of Rob Cosgrove’s industry defining RBS Book originally published in 1987. The entire book is being published here, chapter by chapter.
The old method required keeping close tabs on which tape was which, and rotated them so most of the tapes received even wear from the tape machine, so they could be replaced in groups.
Restoring from tapes required finding the tape containing the backup that had the version of the file you want, and restoring it from tape. To do a complete disaster recovery required restoring first the most recent full backup, then all incremental backups since then. It was a long and arduous task. Tapes would break, or would become unusable because of heat, humidity, wear, and magnetic distortion.
When I designed the first Online Backup software, I predicted a problem with that kind of strategy, and saw an opportunity to employ a powerful function that tape didn’t have – Indexing and database-driven restore.
This gave rise to what was later called Synthetic Full Backup – full backups that are built out of pieces of previous incremental and full backups already on the Server without the need to re-send data to the Server. It saves a lot of time on restore and saves space on the Server.
In addition, it allows for creating full restores to any point in time, and the on-the-fly custom design of virtually any type of restoration of any subset or groups of subsets of full backups by date, file size, file type, backup set, drive location, customer-defined “tags,” file name or folder name, or any other file attribute. Restores became very granular, and very fast, and can be done all in a single pass without the need to restore from multiple backup sets.
As far as I know, my RBackup and Mercury are the only two Online Backup packages currently available that can process Synthetic Backups.
There are two basic types of backups, and two more special cases supported exclusively by RBackup. The first two respect the “Archive Bit,” a piece of information carried along with the file name by the operating system which is used by backup software to determine if a file has been changed since its last backup.
It works like this: Whenever you edit a file, the operating system automatically toggles the Archive Bit ON. When the file is backed up, the backup software can toggle the Archive Bit OFF. This makes it very quick for backup software to scan the drive for files that need to be backed up. It just looks for those with the Archive Bit turned ON.
Incremental Mode – Backs up files that have been changed since the last backup, and then turns the Archive Bit off.
Differential Mode – (Exclusive to RBackup) Backs up files that have been changed since the last backup, but leaves the Archive Bit ON. This special type of backup is used when an end user wants to use RBackup AND regular tape backup software at the same time. Tape backup software is not typically smart enough to have a Differential Mode, so the procedure is to run the tape backup AFTER the Online Backup. The files are backed up online, and then backed up again to tape, and their Archive Bits are turned OFF by the tape backup software.
Full Mode – Backs up all selected files regardless of the setting of the Archive Bit.
BitBackup Mode – (Exclusive to RBackup) Backs up just the parts of files that have changed since the last backup. For example, if only one paragraph changed in a document, BitBackup will back up just that paragraph. This saves time and bandwidth, with the added bonus of also maintaining a local Data Store for immediate restoration at full internal network speed. This also provides a local backup AND Online Backup for redundancy.
Some software, like RBackup and Mercury can operate without using or changing the Archive Bit at all. Rather, they use more reliable methods like cataloging the date and time stamps.
Tomorrow I’ll talk more about Best Practices for categorizing data into Critical, Important, and Archival groups.
Rob Cosgrove is the President of Remote Backup Systems, founder of the Online Backup Industry, and a vocal advocate for maintaining the highest standards in Online Backup software. His latest book, the Online Backup Guide for Service Providers: How to Start and Operate an Online Backup Service, is available online now, on Amazon.com, and at bookstores.
Remote Backup Systems provides brandable, scalable software and solutions to MSPs and VARs enabling them to offer Online Backup Services.