Rotating the (secondary) Mirror Drive vs. Rotating the (primary) Backup Drive
When you would like to store your data offsite, you can choose to Mirror your Backup Storage location or you can replace your Backup Storage device to store the old media off-site. Review the following pros and cons to decide which works best in your situation.
*** Lockstep highly recommends rotating the Mirror drive rather than the primary Backup Storage drive. ***
Benefits of Rotating the Mirror
Why would I decide to rotate my Mirror? When you would like to frequently take data off-site, we recommend that you follow the rotating Mirror plan. Rotating the Mirror allows you to leave your Backup Storage drive "as is" and simply move a copy of the backup data (the Mirror) to an off-site location for safekeeping. Leaving the Backup Storage drive in place ensures its normal operation and availability to users. All of the backup history is retained on both the Backup Storage drive and the Mirror drive.
Problems with Rotating the Mirror
If data stored on the (primary) Backup Storage drive for some reason becomes corrupted, the corrupted data is then replicated to the (secondary) Mirror drive, effectively placing corrupted data in both places.
Why would I decide to rotate my Data Repository? - Some users choose this option when they think in terms of Archiving. They decide to rotate the Data Repository because they are thinking, "We have an Archiving policy to comply with. I need to take all of our data off-site every quarter. It is the end of the quarter and now I would like to take everything off-site and start fresh with new media." Rotating the Data Repository on this infrequent basis allows you to comply with your Archiving policies.
If you find that you need to rotate data off-site frequently (daily or weekly), consider using the Mirror for this purpose. We recommend using the Mirror in a rotation schedule because it is not disruptive to your Backup Clients.
Benefits of Rotating the Data Repository
Each Data Repository has an independent copy of the backup data. In other words, if something affects one copy of the
Data Repository, the other copies in the rotation pool are not affected.
Problems with Rotating the Data Repository
The first instance of rotation invokes a complete re-baseline backup to occur, which is time consuming. Each time the Data Repository is rotated, the Backup Clients need to resynchronize with the new Data Repository. This resynchronization process takes a little time when the first backup occurs after the Data Repository is changed.
When using multiple Data Repositories in a rotation schedule, you may need to cycle through one or more of the Data Repositories should you need to restore a very specific version of a file. For example, if you need to restore a very old version of a file by a specific date that you have in mind, you will need to be running the Data Repository that was in use at that time. Using our example above in which you have 2 Data Repositories, in which you use each for 60 days apiece, if you want to restore a file as it existed in the first 60-day rotation cycle (or a file that is 61 to 120 days old) you will need to restore from the first Data Repository.