Guest blog: Mike Johnson, Rocket Software.
Let’s talk about purchasing products that relate to customer backup and recovery needs.
Clients might wonder why they might need additional solutions for data protection, when they’re already mirroring their particular data to another site. After all, with the step of mirroring, they presume this ensures a full backup copy which can be restored anytime. However, the real issue concerns the restore process from a mirrored copy. Problems can arise which can be disastrous, as the examples below illustrate.
Restoring without corrupting
Although many clients are mirroring their data to another site, they ultimately desire a third transportable version of that data.
This is so they can easily go to another data centre. Here is another problem with sole reliance upon mirroring: when a person mirrors their data, data that is corrupt also gets mirrored. However, the portable third copy allows the consumer to restore from a corruption-free backup.
When the storm hits
In 2005, Hurricane Katrina swept through the southern United States with a vengeance.
Many states suffered devastating damage to their infrastructures, homes, businesses, and daily lives. Naturally, the IT world was not exempt. A particular company based their main site out of Louisiana and their mirrored site out of Alabama. Once Katrina hit both of these regions, they lost their primary data, and their backup.
It is now standard practice to place a mirrored site east of the primary site, and east of the San Andreas Fault line. The lesson here is that reliance upon mirroring alone is not enough to ensure protection.
De-indexed by Google
In 2011, Google made clear why sole reliance upon mirroring in not wise with the accidental deletion of over 100,000 accounts.
This error hit all primary and mirrored sites together. So, all of the email accounts were lost for good. The question is: what can be learned from these examples of reliance upon mirroring alone for protection of data?
Hopefully, it is now quite obvious that mirroring cannot give complete protection of data, and should only be used as one piece of the puzzle towards keeping customer data protected.
In support of mirroring
In the interests of balance, let’s discuss what mirroring is good for.
It does provide a smooth way for businesses to continue their various operations. The problem is, it does not grant historical recreation of data. Ultimately then, this task is dependent upon the programmer, with the associated risk of human error.
It is helpful to also focus on how mirroring should be used. Keep in mind that this tool is slightly expensive, and so it makes sense to use it on key storage groups instead of for everything. For example, clients could use mirroring for their most important applications and simply use the tool of backup for storage groups that are less critical.
Mirroring is not a good enough tool to offer complete protection of data.
In fact, mirroring ought to be paired with an additional backup and recovery plan so as to avoid problems when it comes to recovery. Institutions such as large banks, insurance agencies and government institutions (with mainframe environments) must pay careful attention to this information, as they hold some of the most critical data.
Below you will see what a healthy backup and recovery solution looks like:
- Point-in-time products that replicate data must be made automated.
- Many historical backup versions of data must be made to allow the flexibility of being able to restore from whatever timeframe needed and that is corruption-free.
- Portable copies of data must be created that can go anywhere and be used anywhere.
- Recovery granularity must be offered which grants the ability to restore from various places.